Success with Studio Room(and Speaker)Correction with Reaper/Room Equalisation Wizard

[kodebode2]

Hi,

I apologise ahead, as I am not the most concise of writers. This is a long read, but it may just have a really good solution to your challenge.

Problem : to improve the sonics in your studio/listening environment, by "tuning" the speaker and room.

I obtained really great results and thought it good to share. The results were so spectacular, this is my way of giving back to the wonderful people who have made Reaper and other top class tools for this purpose, available, many of them for free. Thank you.

Some background.

1. Many years ago, as a live mixing engineer I got into the habit and belief that I could correct speaker and in particular, the anomalies of a room, using graphic EQ on the main mix, and/or attenuating the most significant resonances via channel EQ, especially iin the low end. I also applied graphic EQ's to the foldback - stage mixes, and at the time, this did make a huge difference in clarity at the venue where I did this, over several months - tuning purely by ear and using lots of hours to listen to tracks I knew well. This habit started a few years earlier, at other venues, but the aforementioned was my most significant application. I however caution that over time, I have discovered aspects of this approach that will need modification, for effectiveness. As some of you have already mentioned on this thread - with EQ - sometimes less is more, and if I may conclude on this point, attempting to correct long reverberation reflections with EQ alone, in my opinion is counterproductive, as it causes you to create excessive EQ compensation, which is unpleasant to listeners near the speakers. So for live EQ correction, I would recommend some moderation, or an alternative thought process be adopted as IMHO, EQ alone will not correct long reverberation reflections, but could effectively resolve the speaker corrections. Ultimately audio is still art, and taste/tailoring/experimentation is required in the application of any method.

2. A while back I attempted using a similar approach, just my ears, to create the eq curves in ReaEQ, to tame only the low end and the mids.

Of course this was after putting in a bit of acoustic treatment. My studio is approx 4 meters by 4 meters - height 2.5 meters, with some soft absorbent furnishing, so it's already fairly well damped. ie RT60 decay of reverberations is fairly fast as I confirmed, through measuring (much later).

I was not satisfied with the results as they did not give me equivalent results on all manner of music and audio,is so I took off the EQ - Fortunately in Reaper all of this is on the Monitoring FX Chain, oh that is such a God send, one of the most usefull things about Reaper - which means at mixdown, if you mix - you have nothing to do, turn on or turn off, as your monitoring FX are not used in the mix.

3. Recently I revisited this idea of using EQ, and ran into some startling limitations, related to some strangeness in the results which I could hear, further to which I started a thread on the Mastering Forum at Gearslutz.com for help. at the link below, which I suggest you kindly read through, in some detail, as there are a few thoughts of others which led me to my current solution, along with quite a bit of research and God's good grace to enable me work through the issues.

https://www.gearslutz.com/board/mast...orrection.html.

OR

https://goo.gl/sl3F7m

4. I have used the Sonarworks demo on my headphones and AKG K702, using the stock correction available for the AKG K701, which was the only one close enough available on their site - a few years back. At the time I was monitoring on headphones only, and it really improved the results, but I have since moved on from using headphones to monitor as I have reservations about their accuracy, and am skeptical about the abilities of a single driver headphone, no matter how well designed, which is what most of them are really, to equivalently cover the full frequency spectrum - my own layman reservations.

5. Though I've posted my approach on the aforementioned thread, I'll summarise here.

While this is specific to monitors, you may be able to use some of this in your approach to correcting your headphone response.

Tools :

a) SPL Meter
b) Measurement Microphone
c) Room Equalisation Wizard software (REW)
d) Reaper DAW - version 4 or 5 (works same on both)
e) Reaper ReaVerb - used as a convolution plugin
f) Liquidsonics Reverberate LE - convolution plugin
g) Voxengo SPAN - to help you check levels and compare.
h) Audio interface - EMU 0404 USB- (or any other decent audio interface with as flat a frequency response as possible in inputs and outputs)
i) Operating System - Windows 10 (at the time I wrote this, REW and Reaper are also available for MAC OS, and this should work in some earlier versions of Windows)

I have deliberately omitted mention of my monitors, as I think the point of this post is that the approach can be used on any monitor.

Link here to an example of the principle applied to a really basic single horn speaker. Not sure what tech they used but the overall concepts and outcome, must be similar.

https://youtu.be/bWLnYmuTGS4

Following the documentation in the REW help document, and also some other links I had read on the internet :

1. Measured Left and Right Speaker individually, from the listening position, where my head would be.

2. Generated EQ filters to correct - to align with a target curve, using the Generic EQ option. exported these to text files, for left and right.

3. Now here's where things take a departure from the frequency correction approach discussed on this thread.

http://forum.cockos.com/showthread.php?t=183340


4. Instead of applying the filters in an EQ, simply out of laziness, I thought, why don't I try to generate correction impulse responses instead, these are somewhat similar to the EQ correction filters, but seem to also address time based anomalies, such as aligning the phase of frequencies, so that the arrive at the same time. It really came from my laziness, the thought of copying 15+ eq's for each of Left and Right was a bit too much effort.

I generated a stereo correction impulse file, from the measurements of Left and Right.

5. Initially I imported the impulse into Liquidsonics Reverberate LE, and was astounded - at the results, note I turned down the dry signal to - infinity i.e zero, and set the wet signal to 0 db or as near as possible - which is one of the challenges of this tool - difficult to set this to zero !

I did have to compensate level with another trim plugin, to match by ear the volume differences so as I toggled the trim plugin and Reverberate LE - the subjective volume would remain the same. Due to the change in the frequency spectrum, achieving this is another whole thread of discussion, but I state here that the ability to compare at equal loudness is key, so you avoid the pitfall of considering the louder option is better. No matter how you try to compensate the loudness, it will still be subjective cos of the relative change in frequencies across quite a bit of the spectrum. My trim plugin of choice is Sleepy Time Records - Stereo Tool - from an outstanding software developer of some of the most useful tools that have been on literally every channel in my DAW. awesome tools.

The results were clearly better than the uncorrected version, as I discovered unknown to me that my - "accurate monitors" had deceived me for at least 18 months, and only measurement revealed to extent of their deceit.

I've put all my gushing comments in the other thread on GS, so will not repeat here. Suffice to say, I was in audio heaven, but it was about to get even better.

6. Waking up with fresh ears the next morning, listening again, I could not shrug off a certain discomfort I had at the back of my mind with the end result, so I thought - could an alternative convolution tool improve the audio, and I imported the same stereo impulse into ReaVerb. Cut a long story short, ReaVerb did a much better job.

Thanks to all who developed the technologies, ideas, tools, and especially the freeware and shareware. The least I can do is post my good fortune to thank them, and others could benefit from my "discovery".

The results are unprecedented. At the listening point - I currently enjoy - pin sharp focusing on the audio, enhanced clarity, great stereo, even at low listening levels, over a wide listening area - it has completely improved my monitoring, and my listening pleasure. I am more confident that my mixes will translate better to other listening environments.

[Borodog]

Interesting. How did you create the stereo impulse response file from them left and right measurements?

[Tod]

Originally Posted by Borodog

Interesting. How did you create the stereo impulse response file from them left and right measurements?

Heh heh, exactly my response in the other thread.

[kodebode2]

Originally Posted by Borodog

Interesting. How did you create the stereo impulse response file from them left and right measurements?

Originally Posted by Tod

Heh heh, exactly my response in the other thread.

I could not also imagine how simple, quick, yet effective using correction impulses was.

I've attached screen shots of the menu options I used

1. To get to the Export Menu for the impulses.

2. The form to select which measurements would be left and right, sample rate etc., and my settings for this export.

It exports as a stereo wave file, which you can import into your convolver tool, whichever tool you choose.

I typically run my DAW at 96K, I recommend you generate for your DAW sample rate.

I really hope this helps others.

Do let me know how this approach compares with the results of the frequency filters using EQ.

Listening to the changes, I am shocked that I had been so deceived by my monitors/room. The dynamics in so much music I am listening to is back! I never knew it had been missing, thinking I was enjoying the benefits of studio monitors and the "flat frequency"- explaining the somewhat bland nature of what I was hearing as part of the studio monitor sound, but that ear smacking thump was missing, even though my monitors have the ability to do bass easily down to 45 hz.

I did not know that the whole bass and mid-range was out of balance at specific frequencies and had been out of sync, Until yesterday, I never knew anything was missing, cos it was still better than what I had been my prior monitoring system. With the basic correction, especially voices and drums (actually everything) sounds so much more natural, more thump, and smack (and I do not mean just bass - but the right proportion of frequencies for every sound is back - ) like I'm listening to a mini concert of live musicians - even on recordings that I am certain was not done 100% live - i.e definitely a studio recording one with overdubs.

Huge change, now guitar picking, drums, anything dynamic, literally pops out of the speakers, and i think the filters also correct the timing/phase so that the audio arrives at your ear as close as possible in phase - so the audio sounds even more cohesive than using frequency filters alone, which actually changes the phase, i.e. alignment of timing of frequencies.

Now drum beats sound like almost like someone is actually hitting the a real drum somewhere within the cabinet of the speakers, it is really awesome - some tracks become so "lifelike". especially voices, tracks, and albums that I had previously dismissed as not good enough, not realising that my monitoring was not telling me the truth. I feel so absolutely embarrassed to have gone on for so many years in ignorance.

Actually its anger at how I have read so much disinformation in highly respected music and audio magazines and online sources. I've been an avid audio technophobe for about 20+ years, but with my current audio interface, some room treatment- this room correction ranks as the most significant 3 changes ever to my studio, the room correction, being definitely the greatest - really like night and day... Thumping dynamic music on anything you play which has dynamics, - with ease....

The speaker manufacturers need to do a much better job to help us derive value from our investments - I've been sitting on a gold mine and never knew it.

[DVDdoug]

Good to hear you got good results! But, no acoustic treatment?

"Traditionally", you start with acoustic treatment. There are a lot of things that can't be fixed electronically/digitally and if you start-out with good monitors in a good room you are preventing rather than compensating for problems.

1. Many years ago, as a live mixing engineer I got into the habit and belief that I could correct speaker and in particular, the anomalies of a room, using graphic EQ on the main mix, and/or attenuating the most significant resonances via channel EQ

...- tuning purely by ear

I assume the "big acts" are now using some automatic EQ, possibly making some final tweaks by ear.

Live sound reinforcement is similar but with a different philosophy. You're just trying to get good sound in the room. In a mixing/mastering situation you're trying to get accurate sound, so that it sounds good everywhere else.

but I have since moved on from using headphones to monitor as I have reservations about their accuracy, and am skeptical about the abilities of a single driver headphone, no matter how well designed, which is what most of them are really, to equivalently cover the full frequency spectrum - my own layman reservations.

As an overly-broad general statement, good headphones probably have better frequency response than speakers/monitors. And, you can get very-good headphones for a couple hundred dollars. In a blind test, you might choose a pair of $300 headphones as "the best in the world".

The physics that require multiple drivers in speakers/monitors don't really apply to headphones because with the driver "interacting" directly with the ear, you don't need a large woofer.

Headphones are difficult to measure because although you don't have to deal with room acoustics, you have to deal with the effects of the ear, and unlike room effects, you don't want to remove the effects of the ear.

And, bass is experienced differently with headphones. With speakers (or live sound) you feel the deep bass in your body.

And, the sense of space is experienced differently.

...Headphones are just "a different experience". And, that's why pros don't use them as their primary tools.

[kodebode2]

Originally Posted by DVDdoug

Good to hear you got good results! But, no acoustic treatment?

"Traditionally", you start with acoustic treatment. There are a lot of things that can't be fixed electronically/digitally and if you start-out with good monitors in a good room you are preventing rather than compensating for problems.

I assume the "big acts" are now using some automatic EQ, possibly making some final tweaks by ear.

Live sound reinforcement is similar but with a different philosophy. You're just trying to get good sound in the room. In a mixing/mastering situation you're trying to get accurate sound, so that it sounds good everywhere else.

As an overly-broad general statement, good headphones probably have better frequency response than speakers/monitors. And, you can get very-good headphones for a couple hundred dollars. In a blind test, you might choose a pair of $300 headphones as "the best in the world".

The physics that require multiple drivers in speakers/monitors don't really apply to headphones because with the driver "interacting" directly with the ear, you don't need a large woofer.

Headphones are difficult to measure because although you don't have to deal with room acoustics, you have to deal with the effects of the ear, and unlike room effects, you don't want to remove the effects of the ear.

And, bass is experienced differently with headphones. With speakers (or live sound) you feel the deep bass in your body.

And, the sense of space is experienced differently.

...Headphones are just "a different experience". And, that's why pros don't use them as their primary tools.

It's this kind of overly critical response that sometimes makes me wish I did not bother posting. We all come from different levels of experience, ability and knowledge, and not all of us are audio professionals who do this for a living.

I simply discovered something really important, that I was really excited about and thought that others, who may not know about this, would benefit from my own learnings.

Thanks. Your response was not a positive experience for me. I'll probably stay off the blogs, if this is the kind of response I get to my openness and good intentions.

[karbomusic]

Originally Posted by kodebode2

Thanks. Your response was not a positive experience for me. I'll probably stay off the blogs, if this is the kind of response I get to my openness and good intentions.

First off, welcome to the forum (though you've been here as long as I), no need to leave.

I can't really find anything wrong with his reply, he's providing additional information which falls into that same learning category. Doug is a smart guy and as far as I can tell is both educated and works in the field professionally, I wouldn't just abandon his posts like that - he's just providing additional helpful input.

There are lots of very smart members here that can provide very helpful info and/or augment what others post so that everyone learns together.

[kodebode2]

Originally Posted by DVDdoug

Good to hear you got good results! But, no acoustic treatment?

"Traditionally", you start with acoustic treatment. There are a lot of things that can't be fixed electronically/digitally and if you start-out with good monitors in a good room you are preventing rather than compensating for problems.

I assume the "big acts" are now using some automatic EQ, possibly making some final tweaks by ear.

Live sound reinforcement is similar but with a different philosophy. You're just trying to get good sound in the room. In a mixing/mastering situation you're trying to get accurate sound, so that it sounds good everywhere else.

As an overly-broad general statement, good headphones probably have better frequency response than speakers/monitors. And, you can get very-good headphones for a couple hundred dollars. In a blind test, you might choose a pair of $300 headphones as "the best in the world".

The physics that require multiple drivers in speakers/monitors don't really apply to headphones because with the driver "interacting" directly with the ear, you don't need a large woofer.

Headphones are difficult to measure because although you don't have to deal with room acoustics, you have to deal with the effects of the ear, and unlike room effects, you don't want to remove the effects of the ear.

And, bass is experienced differently with headphones. With speakers (or live sound) you feel the deep bass in your body.

And, the sense of space is experienced differently.

...Headphones are just "a different experience". And, that's why pros don't use them as their primary tools.

I think I better respond to online bullies like you, in a more firm manner and direct manner. The anonymity of the online world, has the down side of a certain lack of respect, and decorum, which in your case needs a definite redress.

You obviously did not read my initial post, where I clearly stated that I did have some minimal acoustic treatment. You were so erroneously hasty to come to your own exalted contradictions, which was the primary purpose of your post, which added no value to the thread.

I also think it would be better if you gave the opinions of others more respect rather than attempt to enter into pointless arguments, which I will not be drawn into. It's my opinion that headphones are not the best for what I do, which led to my preference for studio monitors, that's my opinion, I need not justify this opinion, and neither does it need the kind of retort you responded with, which did not add any value to the thread.

May I add - if you have nothing better to do than be, for no reason, except maybe an inverted glorification of self by putting down others, critical of the efforts of others to move the world forward, and share knowledge, please stay off the blogs, or just read, and should you have something to add, say it nicely with respect for your fellow human being with feelings, or just zip it if you have nothing positive to contribute.

[kodebode2]

Originally Posted by karbomusic

First off, welcome to the forum (though you've been here as long as I), no need to leave.

I can't really find anything wrong with his reply, he's providing additional information which falls into that same learning category. Doug is a smart guy and as far as I can tell is both educated and works in the field professionally, I wouldn't just abandon his posts like that - he's just providing additional helpful input.

There are lots of very smart members here that can provide very helpful info and/or augment what others post so that everyone learns together.

I have provided a more pertinent response, which makes it a bit more clear what I object to. Please read my further response.

[Tod]

Originally Posted by karbomusic

I can't really find anything wrong with his reply, he's providing additional information which falls into that same learning category. Doug is a smart guy and as far as I can tell is both educated and works in the field professionally, I wouldn't just abandon his posts like that - he's just providing additional helpful input.

There are lots of very smart members here that can provide very helpful info and/or augment what others post so that everyone learns together.

Hi again kodebode2, I echo what karbo is saying, I didn't read anything in Doug's post the was negative to what you're talking about.

I think you've sparked some real interest here, and speaking for myself, I'm very interested and even excited about what you're doing.

I think all forums like this will have the "nay sayers", but just ignore them, or make them fess up to what they mean. And I really don't think Doug was "nay saying".

I want to proceed with your method here, but I'm also in the middle of some other projects that demand my attention, so I probably can't get to this until next Monday.

But please, do continue.

[karbomusic]

Originally Posted by kodebode2

I have provided a more pertinent response, which makes it a bit more clear what I object to. Please read my further response.

I don't think your edit changes anything as he is only providing helpful **technical** information. At some point the decision to participate socially and learn without feeling slighted is going to be up to you. It would be far more productive and academic if you didn't take the path you are taking now but again, only you can make that choice. To be honest, if you aren't happy with any poster's reply, it is much more powerful to just not reply to it.

We'd all like you to continue, it's interesting, but you can't let such benign posts upset you like that Sincerely trying to help here.

[kodebode2]

Originally Posted by karbomusic

First off, welcome to the forum (though you've been here as long as I), no need to leave.

I can't really find anything wrong with his reply, he's providing additional information which falls into that same learning category. Doug is a smart guy and as far as I can tell is both educated and works in the field professionally, I wouldn't just abandon his posts like that - he's just providing additional helpful input.

There are lots of very smart members here that can provide very helpful info and/or augment what others post so that everyone learns together.

As you observed - I rarely post, even though I've been on the blogs for a while, unless I have something worth asking or contributing to.

The other reason I posted here today, was to say thanks to the Reaper team and community, for an outstanding product that has really moved my audio aspirations, forward. Job done.

If I may add, well over 80% of my use of Reaper, is as, probably the worlds most undervalued media player, audio monitor, which is a major purpose it fulfils for me, and being able to highlight how it has helped me achieve this, the more, especially with taking my audio listening environment to another level, was my intention. This made the unexpected recent improvement in my listening environment especially noteworthy.

I had absolutely no intention to argue or debate, only to present my experience and the benefits of Reaper and REW, to others.

Regrettably this will be my last post to this thread, as it has not been a positive experience.

For the benefit of others who may read this post and gain from it, I have posted links to 3 youtube videos which provide some insights to elements of the process I described earlier, but please note, hopefully, the process I described earlier is probably much simpler, and uses only REW(room equalisation wizard) as the only software application for the measurement, generation of correction eq filters, and deconvolution (generation of correction impuses of which deconvolution is only a part). There are many other sites on the web - too many to reference here, from which I gleaned an understanding of the different aspects, with some trial and error on my part, mostly getting familiar with REW, to get things to work.

Due to limited time, I cannot publish a detailed process post(maybe some other time), of the summary of the process, I provided in the initial post, but if you also read through the REW help guide and follow step by step, it should not be too difficult, to achieve.

https://www.youtube.com/watch?v=Fa9qlB6LK4c&t=3s

https://www.youtube.com/watch?v=-X2cGUa_Mik

https://www.youtube.com/watch?v=lM99RB4gt8o

[hamish]

He did say that he was using acoustic treatment before he made his IR's. Why should he not get huffy?

I have REW and a behringer measurement condensor and I'm looking forward to doing some measurements and monitor tuning with them in the near future.

I have experience with creating room IR's from sweeps in ReaVerb, and I agree that it is a great tool.

Thanks for sharing.

[ivansc]

what a shame that the OP has chosen to interpret/mis-interpret the comments in the way he has.
I for one was looking forward to an informed debate and hopefully an overall improvement in what is possible with regard to monitoring.

Maybe he will have second thoughts and get over his aversion to discussion.
I for one certainly hope so.

[seymour22]

Originally Posted by kodebode2

Hi,

I apologise ahead, as I am not the most concise of writers. This is a long read, but it may just have a really good solution to your challenge.

Hi, does this movie encompass the same?

https://www.youtube.com/watch?v=lM99RB4gt8o

Thank you for your time and effort!

kind regards
tom.kw

[Irvin]

Originally Posted by kodebode2

It's this kind of overly critical response that sometimes makes me wish I did not bother posting.

Your response was not a positive experience for me. I'll probably stay off the blogs, if this is the kind of response I get to my openness and good intentions.

Stop the drama.

If your 'findings' can't be questioned, yes, stay off the freaking blogs. *YOU* are the bully, with your diva attitude.

[bobtom1]

Originally Posted by ivansc

what a shame that the OP has chosen to interpret/mis-interpret the comments in the way he has.
I for one was looking forward to an informed debate and hopefully an overall improvement in what is possible with regard to monitoring.

Maybe he will have second thoughts and get over his aversion to discussion.
I for one certainly hope so.

Exactly as you said. When I read the first post in this thread I was really looking forward to an interesting and civilised discussion/debate. It's a real shame it's degenerated into something rather less. Hopefully it will come back to where it deserves to be and soon.

[slipstick]

Unfortunately I don't think the OP wants any discussion more detailed than "Wow that's great, you're a genius".

Which is a shame because he has some interesting ideas but there's not much that I'm willing to simply accept without any "peer review" permitted. There are loads of people round here who understand this stuff better than I do but I've not decided yet if he's one of them.

Steve

[cyrano]

Those who understand, know at least how little they know.

I've rediscovered EQ this week, thanks to insub, who decided to go measure his cans and describe his method.

What kodebode2 describes, is another new thing for me. And a proof that great discoveries are usually made by lazy people

[DrFrankencopter]

This is interesting, and encouraging... but I'd be curious if you could post some before and after frequency responses (and waterfall plots) from REW. I'd be even more interested if you could provide plots from different locations in the room.

The main issue with EQ'ing the monitors is that the solution is usually only valid for one location...and what works to fix the balance in that one spot can really screw it up in another. The other issues are that EQ can do nothing about 'nulls' in your room's response, and cannot do anything to dampen the ring time of the peaks.

One of the advantages of using EQ over impulse response correction is that you can select which bands to control, and which ones to ignore. But, the IR approach looks really easy to just try out. I might just have to give it a shot.

Cheers

Kris

[Tod]

Originally Posted by DrFrankencopter

This is interesting, and encouraging... but I'd be curious if you could post some before and after frequency responses (and waterfall plots) from REW. I'd be even more interested if you could provide plots from different locations in the room.

The main issue with EQ'ing the monitors is that the solution is usually only valid for one location...and what works to fix the balance in that one spot can really screw it up in another. The other issues are that EQ can do nothing about 'nulls' in your room's response, and cannot do anything to dampen the ring time of the peaks.

One of the advantages of using EQ over impulse response correction is that you can select which bands to control, and which ones to ignore. But, the IR approach looks really easy to just try out. I might just have to give it a shot.

Great Kris, let us know what you come up with.

EDUT: I was going to give this a shot today, but I need to get a SPL meter first. I've got everything else.

[DrFrankencopter]

Originally Posted by Tod

Great Kris, let us know what you come up with.

It's going to have to be later this week. I need to borrow an SPL meter from work...

Cheers

Kris

[kodebode2]

For those willing to spend a bit on a microphone and software, please use Sonarworks Reference 4 or later, to achieve similar results to mine which were achieved using free tools. (see post #1).

Sonarworks is a major commercial business, and will have much more time to respond to your enquiries, the principles are similar, and show you proof of concept and results.

Measure, Decide how much correction you need(like salt you add to taste), Calculate correction, apply. And they have a nice manual/guide, which explains it all in detail.

The advantage with Sonarworks is you can use the same tools for correcting Speakers and Headphones, and now also apply this correction to all the audio playing through your computer, not just the audio in your DAW.

I have not compared the methods so can't say which one is better sound wise, my custom approach described above or Sonarworks, what I am certain of is my described approach above does not work for headphones, as this would require specialist(expensive) microphones to measure headphone frequency response!!.


https://www.sonarworks.com/

[kodebode2]

To add, in today's world - there are different scenarios for a "studio".

1. The larger studio with a control room in which you have a prominent mixing desk, and room for at least 5 people in the control room, who need to hear equivalently from multiple positions - as much as possible the same thing, and typically you have different types of speakers - mains, mids, nearfields, mono auratones, and tiny radio speakers for listening at different distances from the speakers.

2. Project studio, especially in homes and bedrooms, where increasingly not more than one or two people, sometimes only one person works alone, and you do not need such a wide listening area.

3. Those mixing on headphones.

4. If I may add traditional PA's and these include point source and line array.

All of these reproduction methods, in addition to good practice like positioning and acoustic correction, lend themselves to equalisation, to improve the clarity (read flat frequency response) of speakers.

For those who were seeking proof of the efficacy of the method I described earlier - note that

1. In the larger studios, mains are EQ'd. and level balanced via crossovers to achieve the final sound - see Augspurgers as an example....this "tuning" comes as a standard part of the acquisition process/purchase/installation.

2. PA's use similar methods increasingly and already it has been prevalent for a while to incorporate DSP of some sort in their audio signal flow. More recently even smaller PA - point source comes with DSP as standard, at least the preset kind, with limited end user flexibility to tweak it further.

3. Sonarworks is a similar EQ correction process, using your CPU as its DSP engine.

4. Increasingly studio monitors now have this capability to use DSP of one form or another - See Eve Audio, Adam + Sonarworks - expect a product announcement anytime soon, Genelec has their own version, Neumann KH80 now uses DSP and we can expect much more from them. And the Kii threes also use a lot of DSP!!!, so if all the best new speakers (lest I forget the PMC Two twos also use DSP), my point earlier is - this approach has benefits, and with a bit of work, we can enhance using DSP from CPU's also....Even better - where the manufacturer has done their best with DSP, one can take it further, and still improve it outside the speaker, in the audio path - between the DAW and the speaker/amp.

The aforementioned is a very strong case for using computational tools to further enhance what the best physics (including acoustic treatment) cannot achieve by itself.

That was the purpose of this thread, to highlight to others, who may not be aware, that there are methods to achieve similar results(improvements in your listening environment), in a studio environment, with minimal expense and a bit of effort.

May I add that there have been some improvements in measurement software since I started this thread, which could add value to the process, such as better documentation, less bugs, better displays.

So in conclusion - Sonarworks is a lot easier - and until recently with their version 4 - it introduced latency, which has been taken care of. So you can have zero latency as an option in Sonarworks 4. This is payware and well worth it - you decide.

With REW and the related tools(or similar like fuzzmeasure) you get the tools to get under the hood, but this is not for everyone...., it has a steep never ending learning curve., and unlike Sonarworks which has only caught up on latency in version 4, the methods described earlier - incurred no latency.

Some like to buy a new car and get it serviced at the dealer.

Others like to tinker, and rebuild the engine, or do other mods to new or secondhand card.

You choose....

[Tod]

Originally Posted by kodebode2

The advantage with Sonarworks is you can use the same tools for correcting Speakers and Headphones, and now also apply this correction to all the audio playing through your computer, not just the audio in your DAW

Hey kodebode2, could your name be Kaspar by chance?

I do have Sonarworks Reference 4 and am very happy with it. I was able to remove the two 27 band graphic EQs from my studio that I"ve had for close to 30 years.

[kodebode2]

Originally Posted by hamish

He did say that he was using acoustic treatment before he made his IR's. Why should he not get huffy?

I have REW and a behringer measurement condensor and I'm looking forward to doing some measurements and monitor tuning with them in the near future.

I have experience with creating room IR's from sweeps in ReaVerb, and I agree that it is a great tool.

Thanks for sharing.

Spent a while listening to your soundcloud ELectricLAdyLAN.... today...great ideas..

A whole spectrum from acoustic - live to synthetic....(synthesizers)...

I especially enjoyed the Masked Man- where's my little girl - sounds so lifelike, on well set up speakers, and the front to back image is superb. The main vocal sits right in front of you and the other tracks are in layers of positioning behind it in the distance.

Lots of great material - Mr Ethereal et al...

What speakers do you use?

The only caveat I had about the Behringer measurement microphones, was their lack of a correction file. Typically the microphones which provide a correction file (i.e the microphone was tested at the factory and an adjustment file is created which you can download - specific to your own microphone) allows the measurement software to interpret the results more accurately.


Nevertheless even without an individual correction (aka calibration file) for your specific microphone, there is a measure of improvement that is possible, but any discrepancies from flat frequency - which all measurement microphones will have, are not recognised by the measurement software, and adjusted for.

Sonarworks makes a decent microphone which sells for as low as £40+ on amazon.co.uk, i.e it can be bought cheaper than what Sonarworks sells it for.

Dayton Audio also has the EMM-6.

Both of these relatively inexpensive measurement microphones provide calibration files which you can download.

Sonarworks microphone. the XREF20. provides a 0 degree (perfectly on axis) calibration file like the Dayton Audio, but also has a 30 degree calibration file, for use when you position your speakers in an equilateral triangle (60 degrees at each angle of the triangle) and the angle between the listening position and each speaker is therefore 30 degrees. However this additional custom 30 degree calibration file - ideally intended for pointing the microphone straight ahead dead center between the left and right speakers, is in a proprietary format, which only works with Sonarworks software...!

Either of these 2 mics, because they come with calibration files are recommended. Shop around for best pricing....

While I own an EMM-6, If I was buying now, and probably will still buy it nevertheless, the Sonarworks XREF is more accurate in the most critical part of the audio spectrum in the 1st 7 octaves, which contains most of what we pay attention to, based on my research of their calibration files (deviation from flat measurement) - assume we start at 30hz as the lowest frequency that most people can hear easily,

60hz - 2nd octave, 120hz 3rd octave, 240 - 4th octave, 480hz 5th octave, 960 6th octave, 1920 (we approximate to 2Khz for simplicity) 7th octave, 4Khz - start of 8th octave. From their measurement calibration files. pretty accurate in this region, requiring almost no correction, up to about 5Khz which is a good sign of a good measurement microphone, and this character is consistent across different microphones from the same supplier..... XREF is a definite good buy...

[kodebode2]

Originally Posted by Tod

Hey kodebode2, could your name be Kaspar by chance?

I do have Sonarworks Reference 4 and am very happy with it. I was able to remove the two 27 band graphic EQs from my studio that I"ve had for close to 30 years.

So absolutely delighted to hear of the improvement in your audio...with Sonarworks.

It took me a while (3 recent years) to go through the hoops and come to the conclusion that while graphic EQ's are better than nothing if that's all you've got, but parametric EQ's, which is what Sonarworks uses internally - lots of them - parametric eq's, do a better job. I've used both... para and graphic. para is better. The beauty or challenge with Sonarworks is the exact makeup of the correction how many eq's(this is the real nut - how many eq's are needed), and the settings of each parametric eq, is not divulged/disclosed, which can be a blessing to avoid users tinkering with this. Thankfully you do still have some broad tweaking options for adjusting to taste..

Using some visual analysis tools I was able to see how imprecise Graphic EQ;s were...kinda like broad strokes....compared to parametrics

If Sonarworks divulged their correction details, the cat would be out of the bag, and you could use any other parametric eq or impulse thereof, to achieve the correction !!..., and like those in the Hi-fi world spend hours optimising this manually to personal taste....more art less science...

A few days ago I was wondering how elated the 1st person who saw micro organisms under a microscope must have felt, a whole new world.

More than ever before, having tried different methods, the impact of DSP(in hard or software) correction on what we hear in reproduced audio remains huge.

What I appreciate most is we can listen at lower volumes and yet still hear so much better, and preserve our hearing.

Let#s turn it on its head, and one day add microphone correction, to give you an even more accurate version of a recording if you want this !!!. with calibrated versions of the typical recording microphones- SM58, U87, etc... and you can adjust how much realism you want the correction to have - a whole new world...

Thanks for your kind response. I appreciate this.

EDIT - 24th Feb 2019 - I am no longer sure that Sonarworks uses parametric EQ's to implement their correction, as they divulge very little about their internals......... Spoke too soon in the paragraph above - an assumption on my part.

[Tod]

Originally Posted by kodebode2

Using some visual analysis tools I was able to see how imprecise Graphic EQ;s were...kinda like broad strokes....compared to parametrics

I did a lot experimenting with Reaper's ReaEQ a while back. I actually setup ReaEQ for 27 bands of 3rd octave EQ, of course they
were parametric bands made to mimic graphic. I never could find anything that was good enough replace my Graphic hardware EQs.

If Sonarworks divulged their correction details, the cat would be out of the bag, and you could use any other parametric eq or impulse thereof, to achieve the correction !!..., and like those in the Hi-fi world spend hours optimising this manually to personal taste....more art less science...

I built my first recording studio in 1970 and room EQ, especially the control room, has always been a major factor. Back then all
I had was a 10 band half octave stereo EQ.

Finally in 1979/1980 I built my last control room. Then in the early mid 80s I got my first computer and learned Quick Basic, one of the
higher level computer languages. I also acquired my 27 band graphic EQs at that time and one of them had a built in Real Time Analyzer to
help adjust each of the 27 bands to a flat room curve. The flat curve didn't sound too bad but but lacked, especially on the bottom end.

So I designed a program in Quick Basic where I could add percentages of the fletcher munsin curve to it. What I did was get voltage readings
of each band on the 2 EQs with their flat room response. Then in my QB program I converted the voltages to dB levels so I could add the
percentages of the flecther munsin curve in dB levels, which were then converted back to voltages. I then used those voltages to adjust
each of the 27 bands. It took a lot of experimenting, but once I got the EQs set right, that's where they are to this day.

So yes, ha ha, I wouldn't mind knowing how your correction EQs work.

[Philbo King]

Interesting stuff. I'll do some testing and post back in a few days. I'm due for another REW test anyway.

[Tod]

Originally Posted by Philbo King

Interesting stuff. I'll do some testing and post back in a few days. I'm due for another REW test anyway.

Hi Philbo, so you're using REW, I assume you have a decent acoustic mic. Are you testing from one position or several positions.

What Sonarworks does is test several positions in and around the main listening postion. This is all done in a period of 1 or 2 minutes, depending on whether you pause the measurement process or not. This took me a few tries before I found a short mic boom that I could hold the mic at just the right height, then it went quickly.

When it was all said and done, I compared Sonarworks to my old graphic EQs. At first I wasn't sure because there wasn't a lot of difference, but as I tested and retested I decided that Sonarworks sounded a little better, more clearer. Consequently I turned my old graphics off and haven't turned them on since.

[kodebode2]

Originally Posted by Tod

I did a lot experimenting with Reaper's ReaEQ a while back. I actually setup ReaEQ for 27 bands of 3rd octave EQ, of course they
were parametric bands made to mimic graphic. I never could find anything that was good enough replace my Graphic hardware EQs.



I built my first recording studio in 1970 and room EQ, especially the control room, has always been a major factor. Back then all
I had was a 10 band half octave stereo EQ.

Finally in 1979/1980 I built my last control room. Then in the early mid 80s I got my first computer and learned Quick Basic, one of the
higher level computer languages. I also acquired my 27 band graphic EQs at that time and one of them had a built in Real Time Analyzer to
help adjust each of the 27 bands to a flat room curve. The flat curve didn't sound too bad but but lacked, especially on the bottom end.

So I designed a program in Quick Basic where I could add percentages of the fletcher munsin curve to it. What I did was get voltage readings
of each band on the 2 EQs with their flat room response. Then in my QB program I converted the voltages to dB levels so I could add the
percentages of the flecther munsin curve in dB levels, which were then converted back to voltages. I then used those voltages to adjust
each of the 27 bands. It took a lot of experimenting, but once I got the EQs set right, that's where they are to this day.

So yes, ha ha, I wouldn't mind knowing how your correction EQs work.

Really interesting the ideas and approaches you have used. For me also it's been a progression.

1. Tried EQ by ear - parametric and graphic to start of with.

2. Blocked ports on my ported speakers.But somehow I'm no longer assured that this is necessary, cos I have a ported speaker, whose ports are rather large and difficult to completely block - which gives me IMHO the best results so far, without having to block its ports to improve bass response. The academic thinking behind port blocking is to reduce group delay - which otherwise leads to a perception of bass being sluggish and less distinct. More modern speakers seem to suffer less in practical terms from this phenomenon. I also discovered that non-ported speakers aso have a small amount of group delay so the benefit of port blocking may be perceptively subjective or insignificant, especially further to EQ correction.

3. Some room treatment.

4. Then took the plunge - measuring the response of speakers with REW. The key is learning which capture method works best, how long the sweep of the test signal should be, and how loud., and where to place the microphone, and angle. Only trial and error provides your own preferred combination of tweaks to the measurement approach.

5. Over time what one does with this information is subject to experience, as there are a few different ways to attempt the correction.

a) Use REW to compute the correction, automatically based on parameters you set, how many bands of parametric EQ, any additional adjustment you want in the final frequency response, probably includes things like the Fletcher Munsen curve you indicated - but I will need to check this to be absolutely sure about it, - it has a whole section for doing this, which also helps you determine how extensive you want the correction to be. In my experience there is a fine balance between the uncorrected sound of a speaker and an overcorrected speaker - both ends of the spectrum sound distorted - somewhere in the middle and only critical listening helps you determine this, is the ideal balance subject to personal preferences, where the most prominent anomalies of the room/speaker have been reversed, but not taken to extremes which becomes another kind of distortion on its own.

Hope you are not offended by this meaty example.

- Rare, medium, well done, burnt. Different strengths of heat.

I may compute a range of severity of corrections which I will try out and switch between, and by ear I know which of them gives me the utmost clarity, the one which is the most transparent, especially listening to the front to back image of the audio, can I perceive easily the phantom front to back placement of instruments in the mix - that's the one that's the best, the one that sounds more like the instruments are right there in the virtual room - this is the one which yields the most accurate flattest frequency. IN the best settings a few things stand out - horns, the attacks of instruments, pianos, guitars, voices, literally jump out of the speaker - virtually I mean... cymbals are crystal clear and every instrument is so easy to hear, with ease, at low volumes, when done right.

The shock when you get it right is having to listen to a lot of your music references to hear what they truly sound like. Many revered tracks will shock you as not being as well recorded sonically or mixed sonically as we recall. And some tracks which you never gave any import to, which were probably not major hits, suddenly sound stunning, pin sharp. I had to run several hours of auditions, and come to accept what the truth was - but all the well recorded and mixed tracks, regardless of public accolades or sales effortlessly reveal themselves.


Sometimes it may make more sense to focus on only one speaker at a time - end to end, to determine the best settings/general approach for correction. e.g use left speaker as the "prototype" to determine optimal approach and repeat on right speaker.

When you can get a single speaker to sound 3 dimensional - which was how it used to be in the very old days of mono, yet the front/back image is still absolutely prominent- main vocal in front, backing voices behind, and you can clearly discern the reverb on the distant piano, even at low volumes, on one speaker - left or right. Mono (i.e only one side of the stereo) during the setup avoids the subjective coloring of our stereo hearing, and once you get the settings right for one speaker - as per the clarity, you can repeat for the next speaker - the whole process using similar options, of course each speaker is measured separately.

This implies that the correction needed on the left and right speaker may not be identical, due to manufacturing tolerances in the components, no physical two speakers are 100% identical.

Typically when deriving the initial curves, I try to avoid any extensive personal coloring of the auto computed correction curves. Why? REW has features to store these filters, or I can save the entire set of measurements from REW as separate files, and in each of these make slight adjustments, to include different combinations of auto and custom adjustments.

The typical outcome of the REW computation is a set of parametric EQ recommendations which you can apply to any multiband parametric EQ plugin.

You could also convert this REW computation to an impulse response which is imported into your convolution plugin, making your convolution plugin behave like a corrective EQ.

b) Using 4 plugins, in sequence. A manual process.

i. A pink noise signal generator
ii. A convolution plugin running the exported impulse of the room/speaker measurement from REW,
iii. Ideally a multi band parametric Q, to manually define correction curves.
iv. A signal display plugin to show you the effect of your correction in the EQ above, on the pink noise modified by the captured room impulse.

You manually define correction curves in the parametric EQ. Visually - no audio is passed to you speakers, ideally you mute all outputs.

You can then transfer the manually tweaked EQ settings and test with real audio on your speakers to see if you got this right.

c) Exporting the measured frequency response to an impulse which I bring into a tool known as DRC designer, which has various options for auto correction, probably more extensive than REW, but also easy to do a lot of harm. Powerful tools need guidance, a lot of guidance, and trial and error.....This has a much better tool for defining the kind of Custom curves such as the Fletcher Munsen....to further modify the auto correction. The outcome of DRC Designer is a final set of impulse responses which you import into a convolution plugin in your DAW - and your convolution plugin now behaves like a corrective EQ.

All of the above methods also allow you in REW and or DRC Designer, to apply the microphone calibration curve to improve the measured frequency response of your speaker/room, which should also make the corresponding correction approaches more accurate. REW and DRC Designer are freeware...

Note impulse referred to above = a generated audio format file, typically a .wav on Windows.

Then over time you decide which approaches results you prefer...I generally avoid method b) the manual one - pretty similar to your aforementioned...prone to human error.

I will find the time to write a mini book on all this with diagrams and full step by step how to's, and keep you posted when this is completed. But this will take time, I think including a comparison with Sonarworks is in order in the book.....that will be most interesting.

[Tod]

Originally Posted by kodebode2

Really interesting the ideas and approaches you have used. For me also it's been a progression.

Yes kodebode2, and after reading your post here I can see you have a real dedication to getting it right, and you will go to any length to do that.

2. Blocked ports on my ported speakers.But somehow I'm no longer assured that this is necessary, cos I have a ported speaker, whose ports are rather large and difficult to completely block - which gives me IMHO the best results so far, without having to block its ports to improve bass response. The academic thinking behind port blocking is to reduce group delay - which otherwise leads to a perception of bass being sluggish and less distinct. More modern speakers seem to suffer less in practical terms from this phenomenon. I also discovered that non-ported speakers aso have a small amount of group delay so the benefit of port blocking may be perceptively subjective or insignificant, especially further to EQ correction.

My speakers have left and right ports too, and there's a plate just above the horn that I can use to block them. I tried blocking the ports early on when I got the speakers, but it didn't seem to make much difference.

My control room is an old LEDE design that was popular back in the 70s & 80s. There's no parallel surfaces but it is symmetrical.

I've got my monitors mounted in the wall about 6 feet off the floor. They're also floating to avoid any structural resonance, which is beefed up a lot.

Thanks for your detailed explanation of what you've done, it's extremely interesting.

I will find the time to write a mini book on all this with diagrams and full step by step how to's, and keep you posted when this is completed. But this will take time, I think including a comparison with Sonarworks is in order in the book.....that will be most interesting.

That would be great kodebode2, the virtual tools we have today are far beyond what we had back in the 80s, even into the 90s. We did have some pretty nice hardware Spectrum and Real Time analyzers, but they were way to expensive. All I had was an oscilloscope, signal generator, and a volt meter.

[Geoff Waddington]

Originally Posted by kodebode2

2. Blocked ports on my ported speakers.But somehow I'm no longer assured that this is necessary, cos I have a ported speaker, whose ports are rather large and difficult to completely block - which gives me IMHO the best results so far, without having to block its ports to improve bass response. The academic thinking behind port blocking is to reduce group delay - which otherwise leads to a perception of bass being sluggish and less distinct. More modern speakers seem to suffer less in practical terms from this phenomenon. I also discovered that non-ported speakers aso have a small amount of group delay so the benefit of port blocking may be perceptively subjective or insignificant, especially further to EQ correction.

Yeah, one of the things you do by blocking the port (besides completely messing up the original Thiele-Small engineered design of that driver in that box ) is change the system from 4th order to 2nd order.

Because the 2nd order is less steep it inherently has lower group delay, although some is inevitable, it's built into the math.

If you were to add a sub things get trickier still, discussion best left for another day

[Philbo King]

Originally Posted by Philbo King

Interesting stuff. I'll do some testing and post back in a few days. I'm due for another REW test anyway.

Well, I gave this a try.

I captured 2 mono IR files using REW, then put them in a temporary project where I converted and saved them as a stereo IR.

Then I used ReaVerb to apply the IR to the Reaper monitoring chain, and measured freq resp using JS Sine Sweep as a source, a JS Vol plug to lower that to a reasonable level, and fed that to my monitors (through the monitoring chain).

I used a measurement mic to capture the result for each speaker, and recorded the freq response using ReaFIR in subtract + track mode for each sweep.

It did some horrible things to the measured frequency response of my monitors. In fact, instead of correcting them, it exaggerated existing peaks and dips in the measured unprocessed response.

I'm fairly sure my methodology was flawed; in theory this idea has merit. Perhaps I needed to invert the correction curve in REW before exporting it. Not sure...

Thanks anyway; I had some fun playing with it.

[woogish]

Originally Posted by kodebode2

It's this kind of overly critical response that sometimes makes me wish I did not bother posting. We all come from different levels of experience, ability and knowledge, and not all of us are audio professionals who do this for a living.

I simply discovered something really important, that I was really excited about and thought that others, who may not know about this, would benefit from my own learnings.

Thanks. Your response was not a positive experience for me. I'll probably stay off the blogs, if this is the kind of response I get to my openness and good intentions.

I think we need a "safe space" for the OP........bring yer own Kleenex....

[kodebode2]

Originally Posted by woogish

I think we need a "safe space" for the OP........bring yer own Kleenex....

Once upon a time, I would not bother to dignify your unfortunate remarks with a response, but I accept - that's the world we live in.

Under cover of anonymity, you can say anything rude, disrespectful, on online forums, and hide, like the coward you must be.

Get some self respect, improve yourself, contribute to making the world a better place, and stop trawling the internet, making stupid remarks.

And learn how to say some positive things for a change, it will do you good. A lot of good.

The web is an impartial platform, an opportunity, we decide what to do with it - good or bad.

You decide. Do some good with your life for a change. Make someone happy - truly happy. It will lead to the joy that's obviously missing which causes you to go to great lengths to attempt to humiliate others.

In this case - it did not work. Your bullying and intimidation did not succeed.

[kodebode2]

Originally Posted by Philbo King

Well, I gave this a try.

I captured 2 mono IR files using REW, then put them in a temporary project where I converted and saved them as a stereo IR.

Then I used ReaVerb to apply the IR to the Reaper monitoring chain, and measured freq resp using JS Sine Sweep as a source, a JS Vol plug to lower that to a reasonable level, and fed that to my monitors (through the monitoring chain).

I used a measurement mic to capture the result for each speaker, and recorded the freq response using ReaFIR in subtract + track mode for each sweep.

It did some horrible things to the measured frequency response of my monitors. In fact, instead of correcting them, it exaggerated existing peaks and dips in the measured unprocessed response.

I'm fairly sure my methodology was flawed; in theory this idea has merit. Perhaps I needed to invert the correction curve in REW before exporting it. Not sure...

Thanks anyway; I had some fun playing with it.

Hi Philbo,

May I start with my heartiest appreciation for John Mulcahy the author of REW, who has made this amazing tool free for all of us to use. I cannot say enough about the generosity of this highly accomplished person, who united so many worlds, DSP, acoustics, software development and more with this herculean effort.

This is partly what also inspires me to share even more of my benefits from using his product. Earlier today he responded to one of my questions about the product and this really helped me close the loop on my long journey with incremental improvement, in using DSP and related tools to augment speaker performance.

Some quick theory(more common sense really)...

PREP WORK AND PROBLEMS

Let's look at the imperfections that DSP attempts to correct or cannot fix.

1. In a physical world, it's almost impossible to manufacture two components to be exactly identical. This natural imperfection in itself is a minor but important cause of the deviations, we experience in speakers.

2. It's impossible to have your head permanently in the same place listening position all the time, in theory if you can get a speaker to be perfectly accurate from one listening position, the likelihood that it will be close to perfect in the region around the ideal listening position is high. All speakers have directivity, that causes a change however slight, whenever listening takes place in any off axis position, so they usual argument against using DSP, because of an ideal use case of a limited listening sweetspot, no longer holds water, especially for smaller studio spaces, where more often then not we use nearfields anyway and the listening sweetspot is relatively small. i.e DSP is the ideal approach for creating perfection in that ideal spot.

3. As good as any manufacturer attempts to deliver the ideal speaker in an anechoic chamber, that measures ruler flat - frequency wise, there is no way to predict its performance in any random room, with indeterminate assurance of placement position, and not knowing the personal preferences of the listener, and the furniture in the room/how much acoustic treatment exists.

4. A caveat which seems obvious, but we must restate is, some acoustic treatment to help reduce reflections and improve diffusion is a great help. DSP will not do much to improve audio quality if we place speakers in an aircraft hanger or the Taj Mahal (reflective surfaces with long echo times).

5. Experimentation with placement and toe in angles of the speakers helps. The usual 60, 60, 60 degrees is a rule, but sometimes you may need to narrow (toe in) or toe out (increase the angle) of stereo speakers.

6. Before and after the measurement process, the volume of both speakers in a stereo environment must be identical to deliver a centered phantom image, for any sources that are centered in the audio mix.

PROCESS. High level only. This is my revised process and much simpler than the earlier comments, I made, as I have taken a decision to exclude some options (clarification is included below)

1. MEASUREMENT.

Using REW, (assuming you have a microphone with a calibration file, and audio interface) and cables, and ideally an SPL Meter with C weighting. These are what I would prefer to use.

a) Measure the frequency response of each speaker. The REW manual provides a description of this. Too many details to attempt to cover here. Furthermore REW has a dedicated forum, where the members are extremely supportive.

Rather than take multiple measurements at various positions, I suggest you take a measurement from only one position, so you only have two measurements to take.

By experimentation you decide if you want to point your measurement microphone up, forward to the center point between the speakers, or directed at the center of each speakers tweeter. There are various schools of thought, to start of with, point your microphone straight ahead between the two speakers, or point at each speaker so the measurement is on axis to each speaker - this latter method is what I would prefer.

The beauty about these measurements is that REW allows you to take multiple and store them in the same file, so I suggest you take a good number of the aforementioned measurements for each speaker, and you may never need to repeat these unless you change your speaker position or room, or listening position, or your speakers.

Link to REW forum :

https://www.avnirvana.com/forums/off...port-forum.10/

GENERATE CORRECTION

a) Method 1 - Transfer the measurement done in REW to another tool such as DRC Designer ( I no longer use this method), and this other tool generates the correction.

b) Method 2 - Using REW - via the EQ feature, based on the measurements captured, generate a set of filters to correct each of the speakers. Try to use a similar target curve and target level. This is absolutely my current recommendation, pretty much almost everything done in one tool, and all the design work is saved in one file, both multiple measurements for speakers and corresponding correction filters for each of these generated by REW. All stored in one file using one tool.

REW's feature for this attempts to generate up to 20 parametric EQ bands to correct each measurement. You need to go to each measurement captured and initiate the generation of the corrective EQ's. Let's call these filters, which is the term used in REW. Ideally use the Generic EQ as the target device for generation.


As I had explained in previous posts, there are two possible methods to correct speaker/room (I say the room part with a bit of caution cos while there is proof that DSP can reduce some of the room issues, it's contribution to this is limited - it cannot completely reverse issues with room reflections)

Method 1 requires transfer of the EQ filter recommendations made by REW, to an appropriate EQualiser (hardware or software) which is in the audio chain. Especially for those who wish to implement this in software, we have a challenge - which plugin do we know that gives us 20 bands of parametric EQ. None that I can think of immediately - they may exist but off the top of my head - have not come across one. This then forces most of us into a shortcut - getting REW to generate no more than 10 EQ filters in its correction.

In my experience parametric EQ filters are insufficient to solve the kind of correction that REW auto generates for us, especially if we are limited to only 10 of these.

The other challenge with using EQ filters, is that in my experience, not all EQ's are the same, so you do not have any assurance of a deterministic correction. Each EQ plugin will have its own slant, which IMHO is nowhere near as accurate as the second method.

c) Method 3 - is a continuation of Method 2 - so you must 1st generate the E filters in Method 2, but instead of transferring the filter settings to a hardware or software equaliser, we execute a filter export in REW - which creates an impulse response of the generated EQ filters into an impulse response file - typically in .wav format, you can generate one file per speaker/measurement correction i.e separate mono files, or a stereo impulse file, containing correction for two speakers - Left and Right. All the 10 (or up to 20 correction filter) are used to create a single impulse per channel. Am exported stereo impulse file has two filter correction "impulses", stored in one file.

IMPLEMENT FILTERS

Method 1 - transferring EQ recommendations to your hardware or software EQ (not my preferred method)

Method 2 - Import the mono or stereo impulse file (containing one or two correction impulses into a hardware or software convolution tool e,g a plugin, placed in the audio playback path.

Waves IR-1 is an example of such a convolution plugin. Impulser2 from Freeverb3 bundle, is also a good one.

This second method is my utmost recommendation - the impulse files are more accurate than trying to use parametric EQ's to correct the problem.

In my experience there is by far more uniformity in the result of using impulses as a correction method, across different convolution plugins, the results are almost identical, from one convolution plugin to another. This consistency makes this approach far more predictable, in results, far less subject to an "independent" version of the truth.

That' it in a nutshell. Just three steps. Measure, Generate Correction Impulse file, Implement by loading impulse file into a convolution plugin, which is in the audio path.

Of course there are lots of minor optimisations of settings, but this post is already long enough. The forum link above has lots of great minds ready to help, if one asks nicely.

With practice, I'd expect each speaker to be measured, corrected and implemented in about 15 minutes.

[Tod]

Originally Posted by kodebode2

Once upon a time, I would not bother to dignify your unfortunate remarks with a response, but I accept - that's the world we live in.

Hi kodebode2, I'm not sure what the problem is here, personally I'm very interested in what you're doing. Heh heh, maybe the adversarial
comments are over my head, I don't know.

I'm fascinated by the things you're talking about and doing, and at one point in my life, I was obsessed with it myself. If I didn't have
my own projects that demand my attention, I'd probably be right in the middle of this, and testing things right along with you.

There are a multitude of things in the audio world today that I have not had the opportunity to indulge in, especially many technical things,
so I'm rather ignorant about many things. If I was younger, I'd be searching for the answers.

At any rate kodebode2, I'm enjoying your posts.

[kodebode2]

My last post was a revised summary, a bit long, trying to pack it all into one post but that is impossible, however it retains the nuggets, and the key points for those who would like pursue this on their own, by reading the manuals and doing some of their own research as well as experimentation.

It represents the current state, of easily available and daresay also free tools, in early 2019. Note I work on Windows, so for those on the Mac, certain tools may not be equivalently available, so alternatives will be needed for Mac.

The results in 2019, are so much more ahead of the results I obtained when I started this thread in 2016. More than ever, the outcome is listening to music of any kind and audio of any kind with pin sharp - laser like focus on everything you hear. Anything.

Like all human knowledge all of this is of course subject to change, so it may be a work in progress, but recent listening has reached a plateau that I cannot imagine what else I could do to improve the sound without changing the speakers for far more expensive ones.

One way to look at DRC(Digital Room Correction) is to compare it to turbocharging in cars.

Two ways to deliver higher engine power -

Method 1 - Get a bigger engine - Move from a V4 to a V6 or a V8, fmore cylinders provides an opportunity to increase the engine capacity e.g.from 1.8 liters to 5 liters. This is akin in speaker terms to upgrading your speakers to one with a much better design, more accurate power amplifiers, with better power supply and better drivers, or adding a sub-woofer - simply more power - this will cost you more. And with the principle of diminishing returns, incremental improvements in audio quality from changing components to better ones, gets significantly more expensive, for each improvement. Nothing wrong with this. Sometimes it's the only logical next step to improvement. Out needs to come the wallet or payment card.

Method 2 - Add a turbocharger - so you still use the same engine but through the science of how combustion works, you augment the same engine to deliver a better result. DRC is very much like this - empowering your current speakers to go even further than their default potential. Like turbocharging, different methods reflect the skill and experience of the turbocharging designer and their implementation.

The beauty of today's speaker turbocharging, is that the end user has a good balance of tools that do some of the work for you, while leaving you with ample opportunity to add your own personal preferences and to experiment with the features of the turbocharger and discover ways to tweak it for even more performance. Car enthusiasts will be familiar with such customisation.

The easiest route is a product like Sonarworks Reference 4 which will, under the hood take many of the decisions for you, giving you some control. I have no problem recommending it, and in particular I think their measurement mic the XREF20 is a decent measurement mic which you can get for as low as $60(shop around online)... if bought separately, or via a bundle which includes the mic and software for 299 euros.

The distinct advantage of Sonarworks is you get everything from one manufacturer, microphone, measurement, correction, and implementation - one stop shop. With a key advantage - this is the only solution, that has a large number of users, that also helps improve your headphones via DRC. The basic method for headphones uses an average calibration for your headphone model..kind of an estimate based on Sonarworks measurement of a good number of headphones of each model.

To improve your headphone results even further, you will either need to get your headphone calibrated (measured) by Sonarworks - i.e you send it to them and get it back - this is a paid service, or you buy one of the headphones which has already been calibrated (measured) by them.

The only caveat with headphones is - if your headphones change over time, due to wear and tear, physical change, there may be a deviation between the current measurement of your headphone and what it was sometime in the past. Unfortunately you cannot measure headphones by yourself as this requires some expensive kit.

Sonarworks gives you a great all in one with many of the potential pitfalls taken care of. But like many products, it does not give you extensive control of the correction mechanism.

To be continued.....

[kodebode2]

DRC

If you like tweaking, and a lot of tweaking, the other extreme is the DRC toolset by Dennis Sbragion

http://drc-fir.sourceforge.net/

This toolkit has measurement, correction filter generation, but you need a convolution tool as it does not come with this.

It's free, for geeks, very command line oriented, lots of typing.

DRC Designer is a kind of intermediate tool, which uses DRC from Dennis Sbragion, and gives you a user interface to control the underlying tool. I have had some success with this, and this provided what were remarkable results, at the time, but I think because the tool is not updated frequently, some of the learnings that could improve the product has not been fed back into its development cycle, and there does not seem to be a community of users, helping to achieve this.

Another issue with DRC Designer is that because it is based on DRC by Dennis Sbragion, (which it bundles within it), the current version may become out of step with the latest version of DRC, i.e DRC may have a revised version, which DRC Designer has not included in its distribution. Technically it is possible for you to revise the version of DRC in an already installed DRC Designer by copying over some folders with new content from DRC revised versions(I think - in theory this should suffice, not that I have done this myself).

DRC Designer enables all of the functionality of DRC - measurement, and generation of correction filter impulses, which you can load into any convolution plugin.. When I used DRC Designer, I skilled using it for measurement, and simply provided DRC Designer with measures I obtained via REW.

http://www.alanjordan.org/DRCDesigne...ignerHelp.html

There are other similar tools to DRC Designer, which also use DRC or some other DRC library.

Finally there's REW - Room Equalisation Wizard, which is great for measurement, and gives you two options for correction, EQ suggestions for use in an equaliser, as well as the generation of correction filter impulses, which you can load into a convolution plugin.

The correction filters or EQ settings generated by all these tools, simply attempt to reverse anomalies in your speaker/room.

DRC is like a turbocharger - it can only do so much. And like a turbocharger, there is a fine balance between too much turbo which will destroy your engine (the sound quality) and just enough - to give your engine an inexpensive performance boost.

Generally turbochargers are cheaper to make than using a bigger engine which takes up more space in the engine bay..

All of the above is the science.

Now comes the art.

In general, the 1st and 3rd phase of the DRC process - Measurement, and Implementing convolution filters (which is in my opinion much better than using EQ - from my own testing) - provide pretty much identical results.

i.e Most measurement tools will provide pretty much the same result, and any differences between the outcome, IMHO, will be negligible. There is already enough random minutiae in physical properties like the differences in air temperature, density, and minute changes in the physical properties of your measurement microphone to contend with. If you take two measurements with the microphone in exactly the same place, they will not be exactly identical but will be more than close enough, for all intents and purposes to be identical, so the real issue here is tolerance. The tolerances between different measures by different tools should be small.

Of course there are methods to improve the quality of measurement, but in general most tools include this, and I will attempt to cover such details later on.

Similarly, there is not too much difference subjectively when you take the corrected filter impulse and use it in a variety of different convolution plugins - Reaverb is pretty good, and a free example of one, included with Reaper and available as a standalone plugin for use in any other DAW. Note, on critical listening, there is a difference between convolution engines, but the difference is not huge, the result sounds the same on 1st listen once you ensure that the internal gains used by different convolution plugins is equivalent.

So where does the rubber hit the road?

The generation of correction filters.....

While there is science behind this, and tons of research, ultimately this is like plastic surgery - a bit of science and a bit of art. The fundamental tools of computational maths are similar, but like identical ingredients given to different chefs, the dish tastes different.

You are welcome to try out different tools, and this is obviously a subjective area of evaluation, measurements of the end result do not tell the whole story. In my experience I got the best result in the generation of correction filter impulses using REW.

Caveat, all these correction impulse generators have parameters which you can set, so it's not all down to the tool, the end user has to also develop skill in their use, over time, try things out, see which options sound better, or measure better.

Like having a great musical instrument, but you've got to learn how to play it - which is the art - the human being making a judgement and responding to their observations, to improve the end result.

The makers of all these tools, including Sonarworks are scientists, tool makers, you are the artist, who can take what has been provided and learn how to deliver the best result with what has been provided.

It's similar to any other software tool. You install an EQ, but you have to learn how to use it to get what you want, this takes time, there is no automatic EQ that will deliver a guaranteed Grammy winning end result.

In the next set of posts I will break down each of the three major steps, Measurement, Generation of Correction Filter Impulses, and Implementation of Correction Filters via Convolution, in as much detail as time permits, providing hints of how to optimise each step, because it's impossible to cover it all in any detail in one post. While I use REW for the 1st two steps, most of the hints I share should be transferable to most other tools.