|
|
|
05-24-2020, 07:48 PM
|
#1
|
Human being with feelings
Join Date: May 2009
Location: Japan
Posts: 1,162
|
killing bad reflections in a room with an IR response.. Would this be possible?
Hey all, I'm stuck doing online lessons for my students, and I have to do it from a classroom. However the room has awful reflections, and my students have a hard time hearing me.
So the cogwheels started turning. In theory, if I make a "room IR" with my mic and a decent speaker placed where I usually speak, load that IR in an impulse loader and flip the polarity... Would that awful room sound null?
Anyone tried this? Is there a reason this wouldn't work? Is there something I should be aware of before I try?
Please feel free to post any comment, as I'm not that knowledgeable in this field.. I've only recorded a few cab IRs back in the day, and thought this would be a neat solution to an annoying problem.
|
|
|
05-24-2020, 09:58 PM
|
#2
|
Human being with feelings
Join Date: Sep 2010
Posts: 603
|
First, I thought no way, then I thought wait, maybe that's brilliant! But, not sure, but will be following.
__________________
John
|
|
|
05-24-2020, 10:45 PM
|
#3
|
Human being with feelings
Join Date: May 2009
Location: Japan
Posts: 1,162
|
Quote:
Originally Posted by Serenitynow
First, I thought no way, then I thought wait, maybe that's brilliant! But, not sure, but will be following.
|
Yeah, this probably won't work the first time off, but I just might be onto something. I'll keep you posted.
Anyone willing to chip in and share their results are most welcome
|
|
|
05-24-2020, 10:57 PM
|
#4
|
Human being with feelings
Join Date: Dec 2017
Posts: 2
|
I would think you'd be better off using a different mic?
What type of mic are you using currently?
|
|
|
05-24-2020, 11:29 PM
|
#5
|
Human being with feelings
Join Date: May 2009
Location: Japan
Posts: 1,162
|
Quote:
Originally Posted by elborracho
I would think you'd be better off using a different mic?
What type of mic are you using currently?
|
I'm using an AT4040, and I also have a few handheld dynamics of various brands.
Oh, I know that using a cardioid stuck in my face will drastically reduce the room reflections... Or even using a LAV mic... We ordered a Lav mic to avoid having to have a handheld next to our pieholes all the time.
I could also use shotgun mic, if I had a couple grand to spare
I just thought that having that kind of tool would reduce the need for all that gear..
It would be neat to just flick an IR on and kill the room sound.
And yeah, I'm also a bit of a nerd, and I like the science bit behind it. I like that kind of stuff.
|
|
|
05-25-2020, 12:40 AM
|
#6
|
Human being with feelings
Join Date: Mar 2018
Posts: 111
|
Quote:
Originally Posted by g4greg
Hey all, I'm stuck doing online lessons for my students, and I have to do it from a classroom. However the room has awful reflections, and my students have a hard time hearing me.
So the cogwheels started turning. In theory, if I make a "room IR" with my mic and a decent speaker placed where I usually speak, load that IR in an impulse loader and flip the polarity... Would that awful room sound null?
Anyone tried this? Is there a reason this wouldn't work? Is there something I should be aware of before I try?
Please feel free to post any comment, as I'm not that knowledgeable in this field.. I've only recorded a few cab IRs back in the day, and thought this would be a neat solution to an annoying problem.
|
Hi Greg, there are two conditions you need to fully cancel a wave; one is phase, so not exactly polarity, and the other is amplitude. It’s not easy to do, because if it was, all FOH sound man would have perfect live concert mixes.
Note I said “wave”, but in actuality it is “complex material”, so a combination of many waves, with different arrival times from reflections. Not possible in 2020, unless you have predictable waves, that can be compensated and nullified by a sound system with no latency. There are some technologies like Bose headphones that can partially cancel some predictable noises, like jet engines, so not impossible...but just not yet.
|
|
|
05-25-2020, 12:43 AM
|
#7
|
Human being with feelings
Join Date: Aug 2009
Location: NL
Posts: 1,453
|
The process you're looking for is deconvolution, but I'm afraid that it's ill-posed, meaning that even small amounts of noise can ruin the result quite dramatically, especially if you don't apply some form of regularization.
There's techniques for it in microscopy, and they most often try to model what in your example would be the room, but in the case of microscopy is the point spread function (a blurring of the image by optics). That way, you have an estimated "noise free" model of your impulse response.
I am not aware of those being applied to audio, other than impulse reconstruction when the input is perfectly known (noise free). In your case, you neither know the source, nor the impulse without noise.
A second option is linear prediction, which is a fairly hairy topic.
Googling a bit, I noticed that there seem to be some plugins that achieve some interesting results:
https://acondigital.com/products/deverberate/
But they are still somewhat artefact-rich, so it depends on what you're willing to accept, I suppose.
I don't think subtraction will work. You'll likely just end up with tons of comb filtering effects. Deconvolution in the time domain corresponds to element-wise division in the Frequency Domain.
(Edit: Comb filtering stuff)
Last edited by sai'ke; 05-25-2020 at 12:55 AM.
|
|
|
05-25-2020, 02:20 AM
|
#8
|
Human being with feelings
Join Date: Oct 2007
Location: Lincoln, UK
Posts: 7,924
|
Make yourself a dry vocal booth...
Some musical recordings can need balanced ambience which can be quite involved, but voiceover and narrative stuff sound good even quite dry. Drying out a small area and rolling off the bottom-end you can’t control is quite easy.
Room dividers, cupboards, cardboard boxes, hat stands, lighting stands, rope between anything table on the wall -then drape moving blankets, duvets, etc over them.
Room dividers (especially the fabric-covered ones you can stick pins in) make good video backgrounds and you can quickly customise them to look relevant to your message.
You would need to think about lighting though, as you’d likely block natural lighting or overhead utility lighting. More than one source and diffuse it if you can (LED video lights are great, and cool enough to diffuse with tissue paper or thin carrier bags.
Any EQ method will only work for a very narrow window in the room and might sound unnatural if you varied from this. I’d rather drag a pre-prepared divider and a blanket to sling over a pole than try new sweeps and create convolution files if I had to change rooms at short notice.
>
|
|
|
05-25-2020, 06:26 AM
|
#9
|
Human being with feelings
Join Date: May 2008
Location: Essex, England.
Posts: 561
|
hi, have a look at pressure zone microphones, I home made one of these by putting an electret condenser mike capsule pointing at the flat surface of a plate or piece of shiny covered wood... the distance is 1-2mm, yes mm..used a credit card a spacer, it works by sensing pressure difference between plate and mike. performance is supposed to be twice as sensitive as mike on its own....anyway importantly, when compared with a desk mike the echo and reverberation is way less down. you got to try it to hear.
edit.. also called Boundary Microphones
quote:
"Advantages of PZM Microphones
The main advantage of this technology is the elimination of interference from reflected sound waves. A normal microphone will pick up sound waves from the primary source and also any reverberations, which can result in unnatural sound reproduction. In the pressure zone microphone, sound waves are always in phase and there is no interference.
PZM mics also tend to have a smooth frequency response, good off-axis consistency and strong output levels"
dave
__________________
'Retired technician - not a musician' and registered Reaper user since July 2008
'Excellence is not a skill, It is an attitude' Ralph Marston quotes.
Music at http://soundcloud.com/fixerdave
Last edited by fixerdave; 05-25-2020 at 06:30 AM.
Reason: extra info
|
|
|
05-25-2020, 08:02 AM
|
#10
|
Human being with feelings
Join Date: Jun 2011
Location: Belgium
Posts: 5,246
|
__________________
In a time of deceit telling the truth is a revolutionary act.
George Orwell
|
|
|
05-25-2020, 11:03 AM
|
#11
|
Human being with feelings
Join Date: Feb 2019
Posts: 586
|
Quote:
Originally Posted by g4greg
I'm using an AT4040, and I also have a few handheld dynamics of various brands.
Oh, I know that using a cardioid stuck in my face will drastically reduce the room reflections... Or even using a LAV mic... We ordered a Lav mic to avoid having to have a handheld next to our pieholes all the time.
I could also use shotgun mic, if I had a couple grand to spare
|
A lav mic that isn't any closer than a cardiod by your face is going to sound worse, not better.
You don't need a couple grand for a good working shotgun. The Deity S-Mic 2 is ~350 and pretty comparable to a Sennheiser 416, which is 999 new and still sounds great. There's also a ton of T-powered 416s that go for a similar price to the Deity, with a 40 dollar barrel adapter it will work with any regular 48v preamp you have.
You can also get a 416t cheap and send it to 416T Upgrade in lieu of the barrel adaptor. It's ~100 for the upgrade and it works just as a new 48v version, altogether you save at least 50% of the price.
Anyway, you have options, but really the only solve is getting whatever mic you have closer. Even in a terrible room an AT4040 should work totally fine over Zoom.
|
|
|
05-25-2020, 01:54 PM
|
#12
|
Human being with feelings
Join Date: Jan 2020
Posts: 1,927
|
I can highly recommend these for killing unwanted reflections in an untreated space:
https://www.audimute.com/sound-absorption-sheets
I built a vocal booth just by hanging them from the ceiling. They have grommets so you can hang them on hooks then take them down when you aren't using them.
Probably more than you are trying to spend though...this was for recording vocal tracks that need to sound halfway decent in a mix.
|
|
|
05-25-2020, 04:39 PM
|
#13
|
Human being with feelings
Join Date: May 2017
Posts: 3,202
|
The PZM or lavalier mic has the best chances of working out of all the responses. If you are walking back and forth in front of a chalkboard or blackboard, the lavalier mic is best.
An IR can improve, but not fix, one specific sweet spot in the room.
|
|
|
05-25-2020, 05:22 PM
|
#14
|
Human being with feelings
Join Date: Oct 2007
Location: home is where the heart is
Posts: 12,096
|
Quote:
Originally Posted by sai'ke
Deconvolution in the time domain corresponds to element-wise division in the Frequency Domain.
|
I'll try to digest this one lol. (seriously though, hats off to your knowledge)
Last edited by nofish; 05-25-2020 at 05:42 PM.
|
|
|
05-25-2020, 07:02 PM
|
#15
|
Human being with feelings
Join Date: May 2009
Location: Japan
Posts: 1,162
|
Guys, guys guys!
I do have a singing booth at home, with insulating foam. My home studio is quite OK reflection wise.
This is for lessons in a classroom where I can't modify the room in any way. Thanks for all the physical tips (IE microphones, room treatment and all), but I'm aware of all that.
For now, our solution is to talk with a handheld, and it works. It's just annoying when you have to write on the board or move. Hence the Lavs.
I just like nerding out on software solutions.
|
|
|
05-25-2020, 07:04 PM
|
#16
|
Human being with feelings
Join Date: May 2009
Location: Japan
Posts: 1,162
|
Quote:
Originally Posted by sai'ke
The process you're looking for is deconvolution, but I'm afraid that it's ill-posed, meaning that even small amounts of noise can ruin the result quite dramatically, especially if you don't apply some form of regularization.
There's techniques for it in microscopy, and they most often try to model what in your example would be the room, but in the case of microscopy is the point spread function (a blurring of the image by optics). That way, you have an estimated "noise free" model of your impulse response.
I am not aware of those being applied to audio, other than impulse reconstruction when the input is perfectly known (noise free). In your case, you neither know the source, nor the impulse without noise.
A second option is linear prediction, which is a fairly hairy topic.
Googling a bit, I noticed that there seem to be some plugins that achieve some interesting results:
https://acondigital.com/products/deverberate/
But they are still somewhat artefact-rich, so it depends on what you're willing to accept, I suppose.
I don't think subtraction will work. You'll likely just end up with tons of comb filtering effects. Deconvolution in the time domain corresponds to element-wise division in the Frequency Domain.
(Edit: Comb filtering stuff)
|
Now THAT is the kind of answer I was looking for. Thanks for the input.. Even though it means I don't think I can do it with the available tech right now.
|
|
|
05-25-2020, 07:27 PM
|
#17
|
Human being with feelings
Join Date: Mar 2019
Posts: 4
|
I know some echo cancellation is possible through EQing. In classrooms at the college I work at we use Crestron DSPs that have built in AEC (Automatic Echo Cancellation) and they help considerably. It is my understanding that most of the processing is done in the EQ section.
|
|
|
05-25-2020, 08:00 PM
|
#18
|
Human being with feelings
Join Date: May 2009
Location: Japan
Posts: 1,162
|
I was actually messing with reaFIR, and I got pretty decent results.
so I isolated a spot right after I said something loudly, and looped reaper on the classroom tail, not my voice. I then kicked in reafir in substractive mode, and made it analyse the profile.
Dude, it's not perfect, but the voice are a hell of a lot more intelligible. I might just do that to all my videos. It's quick and dirty, and it works better than I thought.
|
|
|
05-25-2020, 09:11 PM
|
#19
|
Human being with feelings
Join Date: Jan 2020
Posts: 1,927
|
Hmm, I'd probably just record my voice with AirPods. Lol
|
|
|
05-26-2020, 02:54 AM
|
#20
|
Human being with feelings
Join Date: Mar 2009
Location: Dartmouth, Nova Scotia
Posts: 11,184
|
Quote:
Originally Posted by g4greg
I was actually messing with reaFIR, and I got pretty decent results.
so I isolated a spot right after I said something loudly, and looped reaper on the classroom tail, not my voice. I then kicked in reafir in substractive mode, and made it analyse the profile.
Dude, it's not perfect, but the voice are a hell of a lot more intelligible. I might just do that to all my videos. It's quick and dirty, and it works better than I thought.
|
Yup, you'll notice on the networks lately, with everyone working from home -- at first the room sound was horrendous, but most of them now incorporate some kind of correction -- doesn't eliminate but does ameliorate considerably.
__________________
To install you need the CSI Software and Support Files
For installation instructions and documentation see the Wiki
Donate -- via PayPal to waddingtongeoff@gmail.com
|
|
|
05-26-2020, 03:33 AM
|
#21
|
Human being with feelings
Join Date: Mar 2008
Location: Northants,UK
Posts: 195
|
another way round it maybe... Deverb after recording - I 've had good results using a super cardoid mic. https://youtu.be/9c57YnS7ppA
|
|
|
05-29-2020, 01:30 PM
|
#22
|
Human being with feelings
Join Date: Aug 2009
Location: NL
Posts: 1,453
|
Do you have some recorded clips + a recorded impulse? Could give some very rudimentary methods a shot.
|
|
|
05-29-2020, 07:56 PM
|
#23
|
Human being with feelings
Join Date: Dec 2011
Posts: 2,161
|
Have you ever what a Boxing Match ... and the Ring Announcer is holding, what looks to be 2 Microphones in hand ?
Well, it is 2 microphones ... strapped together.
Notice, He'll intentional speak into only One of the mics. For reason.
The 2 mics are usually the same brand. You could use something like an SM57.
The 'trick' ... one of the Mics is 180 wired out of phase [Reversed Polarity].
Any sound hitting both mics [the room sound, or if your class lecture draws a noisy crowd], that sound will be NULL'd.
Speaking into only one of the Mics will allow the signal to 'pass', be heard.
Maybe something that would work.
|
|
|
05-29-2020, 09:20 PM
|
#24
|
Human being with feelings
Join Date: Nov 2011
Posts: 3,409
|
Quote:
Originally Posted by RJHollins
Any sound hitting both mics [the room sound, or if your class lecture draws a noisy crowd], that sound will be NULL'd.
|
Hah, never heard that one before. Great stuff. I bet the frequency response would be awful (no bass?) but the rejection would be fantastic...
|
|
|
05-30-2020, 02:04 AM
|
#25
|
Human being with feelings
Join Date: Dec 2011
Posts: 2,161
|
Quote:
Originally Posted by clepsydrae
Hah, never heard that one before. Great stuff. I bet the frequency response would be awful (no bass?) but the rejection would be fantastic...
|
Stay close on the bottom mic to reduce the cancellation.
|
|
|
05-30-2020, 08:41 AM
|
#26
|
Human being with feelings
Join Date: Dec 2012
Posts: 7,272
|
Quote:
Originally Posted by clepsydrae
Hah, never heard that one before. Great stuff. I bet the frequency response would be awful (no bass?) but the rejection would be fantastic...
|
You don’t need all that bass in a voice anyway, especially in a lecture type situation. I can never understand why people just crank the bass on voices so it’s all boomy and ringing and plosives hitting like kick drums and everything’s all muddy and unintelligible. Oh yeah and it wastes headroom so your woefully underpowered PA is clipping but still nobody can actually hear what they’re saying. Cut that shit. We don’t need Voice Of God, we need clear and understandable human speech.
They used that same technique sometimes back in the early days of live sound as bands were getting louder and louder and they were struggling to get the voices heard. The almost-coincident microphones cancel out a lot of the bleed so you’re not amplifying so much of the rest of the band when you’re trying to get the vocals louder, and it also helps fight feedback.
|
|
|
05-31-2020, 12:10 AM
|
#27
|
Human being with feelings
Join Date: May 2009
Location: Japan
Posts: 1,162
|
Quote:
Originally Posted by RJHollins
Have you ever what a Boxing Match ... and the Ring Announcer is holding, what looks to be 2 Microphones in hand ?
Well, it is 2 microphones ... strapped together.
Notice, He'll intentional speak into only One of the mics. For reason.
The 2 mics are usually the same brand. You could use something like an SM57.
The 'trick' ... one of the Mics is 180 wired out of phase [Reversed Polarity].
Any sound hitting both mics [the room sound, or if your class lecture draws a noisy crowd], that sound will be NULL'd.
Speaking into only one of the Mics will allow the signal to 'pass', be heard.
Maybe something that would work.
|
That's actually a really neat trick!
|
|
|
07-14-2020, 05:05 AM
|
#28
|
Human being with feelings
Join Date: May 2017
Location: Leipzig
Posts: 6,621
|
Quote:
Originally Posted by RJHollins
Have you ever what a Boxing Match ... and the Ring Announcer is holding, what looks to be 2 Microphones in hand ?
Well, it is 2 microphones ... strapped together.
Notice, He'll intentional speak into only One of the mics. For reason.
The 2 mics are usually the same brand. You could use something like an SM57.
The 'trick' ... one of the Mics is 180 wired out of phase [Reversed Polarity].
Any sound hitting both mics [the room sound, or if your class lecture draws a noisy crowd], that sound will be NULL'd.
Speaking into only one of the Mics will allow the signal to 'pass', be heard.
Maybe something that would work.
|
That's fantastic. Thanx
|
|
|
07-14-2020, 06:16 AM
|
#29
|
Human being with feelings
Join Date: Aug 2019
Posts: 42
|
Ear hook mic, the mic should be closer to your mouth, like wireless stage mic.
|
|
|
07-14-2020, 06:26 AM
|
#30
|
Human being with feelings
Join Date: Dec 2011
Location: Portsmouth, UK
Posts: 4,376
|
No, wait.... two ear hook mics ... strapped together....
And we can start a whole nother thread on what to do when the strap breaks.
|
|
|
07-14-2020, 09:17 AM
|
#31
|
Human being with feelings
Join Date: Mar 2008
Location: Oakland, CA
Posts: 10,478
|
Quote:
Originally Posted by g4greg
Hey all, I'm stuck doing online lessons for my students, and I have to do it from a classroom. However the room has awful reflections, and my students have a hard time hearing me.
So the cogwheels started turning. In theory, if I make a "room IR" with my mic and a decent speaker placed where I usually speak, load that IR in an impulse loader and flip the polarity... Would that awful room sound null?
Anyone tried this? Is there a reason this wouldn't work? Is there something I should be aware of before I try?
Please feel free to post any comment, as I'm not that knowledgeable in this field.. I've only recorded a few cab IRs back in the day, and thought this would be a neat solution to an annoying problem.
|
If you can't alter the classroom, an easy solution might be one of those small portable "booths" that will block all reflections except the sound from the front (i.e., your voice).
https://www.soundonsound.com/reviews...e-vocal-booths
|
|
|
07-14-2020, 10:36 AM
|
#32
|
Human being with feelings
Join Date: Aug 2017
Location: Dallas, TX
Posts: 348
|
Another tool would be using dynamics, thinking gentle expansion and some limiting...
Allows for some gain, and drops the reverb tails.
__________________
It helps if the hitter thinks you're a little crazy
- Nolan Ryan
|
|
|
07-15-2020, 04:17 PM
|
#34
|
Human being with feelings
Join Date: Jan 2011
Location: Ottawa, Canada
Posts: 293
|
Also, use a mic intended for close up VO work. Something with reduced proximity effect, and good pattern control. RE-20 might be the classic choice. Work it close, and improve the ratio of direct to reflected sound.
De-convolution just isn't likely to work...
|
|
|
Thread Tools |
|
Display Modes |
Linear Mode
|
Posting Rules
|
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
HTML code is Off
|
|
|
All times are GMT -7. The time now is 04:46 PM.
|