|
|
|
06-22-2017, 04:28 PM
|
#1
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
ReaEQ (Hi & Lo pass filters) ?
Hi all, from what I understand, the ReaEQ Hi & Lo pass filters have a slope of 6dB per octave.
The question is, if I combine either 3 Hi or 3 Lo pass filters, all at the same exact freq, would that give me a slope of 18dB per octave, and if so, what quality would it be in comparison to a good EQ that has 18dB per octave Hi/Lo pass filters.
I think I've seen this question come up before, but I don't remember the outcome.
|
|
|
06-22-2017, 04:41 PM
|
#2
|
Human being with feelings
Join Date: Mar 2010
Location: Adelaide, South Australia (originally from Geelong)
Posts: 5,598
|
Hi Tod,
Yep. Stacking filters at the same frequency does increase the slope exactly as you say. I do that regularly with ReaEQ and have not had any issues with sound quality. ReaEQ is pretty transparent and is my go-to EQ when I don't need or want character.
I would love to see Cockos add a feature to adjust the slope of a single band though and that has been requested very frequently.
|
|
|
06-22-2017, 04:58 PM
|
#3
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
Quote:
Originally Posted by ReaDave
Hi Tod,
Yep. Stacking filters at the same frequency does increase the slope exactly as you say. I do that regularly with ReaEQ and have not had any issues with sound quality. ReaEQ is pretty transparent and is my go-to EQ when I don't need or want character.
I would love to see Cockos add a feature to adjust the slope of a single band though and that has been requested very frequently.
|
Thanks a lot Dave, I really didn't know, so it's a sigh of relief because ReaEQ is also my main EQ, there's literally nothing you cant do with it. I've tried many and even got some fairly expensive ones, but they just don't have any where near the flexibility that ReaEQ does.
|
|
|
06-22-2017, 05:59 PM
|
#4
|
Human being with feelings
Join Date: Apr 2015
Location: Tucson, AZ
Posts: 373
|
If stacking software filters like this work like the real thing then it also means the phase shift will be three times larger.
|
|
|
06-22-2017, 06:24 PM
|
#5
|
Human being with feelings
Join Date: Dec 2012
Posts: 7,272
|
I sure wish somebody would explain to me how changing the bandwidth parameter doesn't do about the same thing. Sure seems to change the slope from everything I've seen. How to equate bandwidth to filter poles and/or Q so that you could dial it in to match a specific slope is a bit more questionable, but...
Edit - But I think stacking filters like the OP suggests is going to change the actual cutoff of the filter, because it will be a total of 9db down at whatever frequency you've chosen, rather than -3. In an actual 3-pole filter, the cutoff would still be the -3db point.
|
|
|
06-22-2017, 06:35 PM
|
#6
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
Quote:
Originally Posted by C-H
If stacking software filters like this work like the real thing then it also means the phase shift will be three times larger.
|
Thanks C-H, so are you saying that stacking the 3 filters will cause a phase shift that the real thing does not have, or that it will just end up the same as the real thing?
|
|
|
06-22-2017, 06:42 PM
|
#7
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
Quote:
Originally Posted by ashcat_lt
I sure wish somebody would explain to me how changing the bandwidth parameter doesn't do about the same thing. Sure seems to change the slope from everything I've seen. How to equate bandwidth to filter poles and/or Q so that you could dial it in to match a specific slope is a bit more questionable, but...
Edit - But I think stacking filters like the OP suggests is going to change the actual cutoff of the filter, because it will be a total of 9db down at whatever frequency you've chosen, rather than -3. In an actual 3-pole filter, the cutoff would still be the -3db point.
|
Hi ashcat, I've tried the bandwidth before, but it causes a rather skewed visual line in ReaEQ, stacking them does not do that. But I don't know the real significance of that. Like you, I'd be curious too, to know how that works.
|
|
|
06-22-2017, 07:59 PM
|
#8
|
Human being with feelings
Join Date: Apr 2015
Location: Tucson, AZ
Posts: 373
|
Quote:
Originally Posted by Tod
Thanks C-H, so are you saying that stacking the 3 filters will cause a phase shift that the real thing does not have, or that it will just end up the same as the real thing?
|
A 'real' 2-pole filter have twice the phase shift of a 1-pole filter.
Almost sure about this... it does make sense.
It is also true that stacking 3 x filter each with a cutoff freq at Fc with not produce a 3-pole filter with the same Fc. It will be lower.
Edit:
This applies to hardware filters. I don't know how ReaEQ is programmed to handle this but I would assume it is supposed to react like a real filter unless some Voodoo is made with zero-phase shift filters.
|
|
|
06-23-2017, 12:34 AM
|
#9
|
Human being with feelings
Join Date: Mar 2010
Location: Adelaide, South Australia (originally from Geelong)
Posts: 5,598
|
Stacking filters does indeed cause a greater phase shift which is also true of the hardware equivalent. I've never had a problem with that audibly when using this to roll off low or high end on an instrument or mic.
It is something to be aware of if using the rolloff in a parallel processing situation though. But the same is true when using hardware EQ.
You can actually view the effect this has on phase by clicking the option to show phase in ReaEQ.
Adjusting the bandwidth control for a steeper filter in ReaEQ is different to stacking filters because as you make the cutoff steeper, you also introduce a resonance peak that emphasizes the frequencies around the cutoff.
This can sometimes be useful and I often use it on bass instruments. It enables you to roll off the very low end whilst also emphasizing the audible bass range.
Don't go too far with it though. The peak becomes very narrow and will create nasty note resonances.
|
|
|
06-23-2017, 12:44 AM
|
#10
|
Human being with feelings
Join Date: Nov 2010
Location: Mullet
Posts: 829
|
ReaEQ's filters are 12db/oct
__________________
I like turtles
|
|
|
06-23-2017, 12:57 AM
|
#11
|
Human being with feelings
Join Date: Mar 2010
Location: Adelaide, South Australia (originally from Geelong)
Posts: 5,598
|
Quote:
Originally Posted by bezusheist
ReaEQ's filters are 12db/oct
|
Yep. Missed that point in Tods original post.
This is all the more reason for the often requested feature to be able to select the slope in ReaEQ HP/LP filters.
6/12/18/24/48dB per octave would be very useful.
|
|
|
06-23-2017, 01:03 AM
|
#12
|
Human being with feelings
Join Date: Mar 2007
Location: UK
Posts: 472
|
I'd love to see steeper slopes on the lo and hi pass filters.
Can I just ask a dumb question? Stacking ReaEQ, is that just adding two or three instances to a track and setting them all as either lo or hi pass as required?
__________________
Bought REAPER V1.5 and still going strong today with V5.
Thanks Justin & Co !
|
|
|
06-23-2017, 01:07 AM
|
#13
|
Human being with feelings
Join Date: Mar 2010
Location: Adelaide, South Australia (originally from Geelong)
Posts: 5,598
|
Quote:
Originally Posted by Bassman1
I'd love to see steeper slopes on the lo and hi pass filters.
Can I just ask a dumb question? Stacking ReaEQ, is that just adding two or three instances to a track and setting them all as either lo or hi pass as required?
|
Not a dumb question.
What we're talking about here isn't stacking multiple instances of ReaEQ but rather, stacking filters within the same instance.
This can be done two ways. You can create a new tab and select the filter type and values to match the filter you've already set or you can use an existing unused tab for the same purpose.
|
|
|
06-23-2017, 01:40 AM
|
#14
|
Human being with feelings
Join Date: Mar 2007
Location: UK
Posts: 472
|
Ah with TABs, not instances, that makes loads of sense now.
Thanks you Dave
__________________
Bought REAPER V1.5 and still going strong today with V5.
Thanks Justin & Co !
|
|
|
06-23-2017, 08:12 AM
|
#15
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
Quote:
Originally Posted by bezusheist
ReaEQ's filters are 12db/oct
|
Humm, I seem to remember a discussion on this forum some years back that concluded they were 6dB/oct. That's why I thought they were 6dB/oct, but that was quite a while back.
Heh heh, so at 12dB/oct there's no way to get 18.
Oh well, 24ddB/oct will work fine, I'm just using it to experiment with.
|
|
|
06-23-2017, 10:19 AM
|
#16
|
Human being with feelings
Join Date: Mar 2017
Posts: 33
|
Quote:
Originally Posted by Bassman1
Can I just ask a dumb question? Stacking ReaEQ, is that just adding two or three instances to a track and setting them all as either lo or hi pass as required?
|
Thank you for asking this. I was assuming they were talking about multiple instances. Multiple tabs makes more sense.
__________________
Remember, there is no right way to eat a Rhesus.
|
|
|
06-23-2017, 10:24 AM
|
#17
|
Human being with feelings
Join Date: Mar 2007
Location: UK
Posts: 472
|
Quote:
Originally Posted by mibo1022
Thank you for asking this. I was assuming they were talking about multiple instances. Multiple tabs makes more sense.
|
I had to ask as I was not getting the results required. I knew I was doing something wrong
__________________
Bought REAPER V1.5 and still going strong today with V5.
Thanks Justin & Co !
|
|
|
06-23-2017, 10:43 AM
|
#18
|
Human being with feelings
Join Date: Dec 2012
Posts: 7,272
|
In relatively quick tests, with frequency at 100Hz, bandwidth at 2, I got exactly 9db difference between 100Hz and 200Hz.
With bandwidth at 4, the -3 point shifted by a little more than 2 octaves. When I set frequency to 400, the -3 point was 113, and 200Hz was almost exactly 3db down from that.
With freq back at 100 and bandwidth at 1, 100Hz was 3db up, and 200Hz was 15db down from there, but the curve starts to flatten out above that. It's definitely not 15db all the way up.
Didnt try stacking. Probably works better.
|
|
|
06-23-2017, 11:55 AM
|
#19
|
Human being with feelings
Join Date: Sep 2010
Posts: 12,562
|
Quote:
Originally Posted by C-H
If stacking software filters like this work like the real thing then it also means the phase shift will be three times larger.
|
This.
Often not an issue. ie. The phase artifacts would pale compared to leaving the frequency content unbalanced.
But if you're parallel processing something, ReaEQ would likely be no go. You'd end up comb filtering content you wished to preserve.
|
|
|
06-23-2017, 01:07 PM
|
#20
|
Human being with feelings
Join Date: Mar 2010
Location: Adelaide, South Australia (originally from Geelong)
Posts: 5,598
|
Quote:
Originally Posted by serr
This.
Often not an issue. ie. The phase artifacts would pale compared to leaving the frequency content unbalanced.
But if you're parallel processing something, ReaEQ would likely be no go. You'd end up comb filtering content you wished to preserve.
|
We're on the same page. That's pretty much exactly what I was getting at a few posts up.....
http://forum.cockos.com/showthread.p...89#post1858289
|
|
|
06-23-2017, 01:18 PM
|
#21
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
Okay, I just tested with pink noise with a 3dB slope, and this is what I got.
Input signal to Span is -27dB, without any EQ, ReaEQ was bypassed.
With one tab set to 200hz Hi Pass, this is what I got.
Input = -27dB
200Hz = -30.2dB (Down 3.2dB from the input)
100Hz = -39.1dB (Down 9.1dB from 200Hz)
Total = 12.3dB
With two tabs set to 200hz Hi Pass, this is what I got.
Input = -27dB
200Hz = -33.8dB (Down 6.8dB from the input)
100Hz = -51.1dB (Down 17.3dB from 200Hz)
Total = 24.1dB
With margin for error I think we can safely say ReaEQ's hi/lo pass filters are 12dB per octave.
Regarding the phase, I turned it on but I don't understand what I'm seeing so I can't comment on that.
|
|
|
06-23-2017, 01:26 PM
|
#22
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
Quote:
Originally Posted by ReaDave
|
Yeah, Parallel processing could be a problem but I can't forsee any time I would do that with ReaEQ, or even if it was involved. Now that may not be true for others but it is for me.
|
|
|
06-23-2017, 03:08 PM
|
#23
|
Human being with feelings
Join Date: Apr 2015
Location: Tucson, AZ
Posts: 373
|
Quote:
Originally Posted by Tod
Okay, I just tested with pink noise with a 3dB slope, and this is what I got.
Input signal to Span is -27dB, without any EQ, ReaEQ was bypassed.
With one tab set to 200hz Hi Pass, this is what I got.
Input = -27dB
200Hz = -30.2dB (Down 3.2dB from the input)
100Hz = -39.1dB (Down 9.1dB from 200Hz)
Total = 12.3dB
With two tabs set to 200hz Hi Pass, this is what I got.
Input = -27dB
200Hz = -33.8dB (Down 6.8dB from the input)
100Hz = -51.1dB (Down 17.3dB from 200Hz)
Total = 24.1dB
With margin for error I think we can safely say ReaEQ's hi/lo pass filters are 12dB per octave.
Regarding the phase, I turned it on but I don't understand what I'm seeing so I can't comment on that.
|
As for the phase.
What is shown is the phase shift between input and output signal. For a 1-pole (6dB) filter the phase shift should be 45 deg at the cutoff freq (Fc). For a 2-pole (12dB filter) it should be 90 deg.
The sign depends on if it is a LP or HP filter.
Edit
Try white noise instead. It's easier to understand with a signal without any slope built in to it.
Last edited by C-H; 06-23-2017 at 03:19 PM.
|
|
|
06-23-2017, 04:33 PM
|
#24
|
Human being with feelings
Join Date: Dec 2012
Posts: 7,272
|
Quote:
Originally Posted by Tod
With margin for error I think we can safely say ReaEQ's hi/lo pass filters are 12dB per octave.
|
???
You said (and essentially agreed with me) that it's 9db drop over the first octave. Neither of us actually measured the second octave to see if it's actually consistent from octave to octave, but it sure does look like 9db/octave to me.
When you stacked them, the -3db point moved to the right, so that 200 was at -6 (which I predicted), but it drops 18db over the first octave down from there. 18 being 9 x 2....
???
|
|
|
06-23-2017, 04:41 PM
|
#25
|
Human being with feelings
Join Date: May 2009
Posts: 29,260
|
I wonder what would happen if someone just slapped one on and used it and listened.
__________________
Music is what feelings sound like.
|
|
|
06-23-2017, 05:04 PM
|
#26
|
Human being with feelings
Join Date: Jul 2009
Posts: 7,571
|
for steeper filters up to 72dB/oct you can use the included JSFX: Apple 12-Pole Filter (Liteon/applefilter72db)
|
|
|
06-23-2017, 05:15 PM
|
#27
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
Quote:
Originally Posted by karbomusic
I wonder what would happen if someone just slapped one on and used it and listened.
|
Heh heh, yeah, exactly. I use HP all the time because I like the way it can clean up the mud, I don't stop to think about what the phase is doing, if it was really bad, I would hear it even if I didn't know that it was a factor.
Back in the old days, phase was not really much of an issue at all, it was rarely mentioned except for Graphic EQs, then it was a much bigger issue, and it was one you had to be aware of.
For me, the tests I made works for me, and now I have a better understanding of what's happening, thanks to all that responded, thank you.
|
|
|
06-23-2017, 05:36 PM
|
#28
|
Human being with feelings
Join Date: Sep 2009
Location: Oz - Blue Mountains NSW, formerly Geelong
Posts: 943
|
If you google Sallen Key filters you'll see how to do this with real filters. There are standard techniques for obtaining an 18 dB/8ve filter with the required -3dB frequency. Or any other slope. I'm sure these could be adapted for ReaEq.
Note that like ReaEq, a basic Sallen Key filter is 12 dB/8ve.
(I'm thinking back 30 to 40 years so my memories of the techniques are slightly faded!)
__________________
It's "its" except when it's "it is".
alanofoz, aka Alan of Australia
|
|
|
06-24-2017, 12:52 AM
|
#29
|
Human being with feelings
Join Date: Nov 2010
Location: Mullet
Posts: 829
|
Quote:
Originally Posted by ashcat_lt
???
You said (and essentially agreed with me) that it's 9db drop over the first octave. Neither of us actually measured the second octave to see if it's actually consistent from octave to octave, but it sure does look like 9db/octave to me.
When you stacked them, the -3db point moved to the right, so that 200 was at -6 (which I predicted), but it drops 18db over the first octave down from there. 18 being 9 x 2....
???
|
the "12db/oct" figure is measured from "0", not the cutoff's -3dB.
for example...ReaEQ with a HPF set @ 100 Hz...one octave down (50 Hz) is -12 dB...
__________________
I like turtles
Last edited by bezusheist; 06-24-2017 at 01:04 AM.
|
|
|
06-24-2017, 08:41 AM
|
#30
|
Human being with feelings
Join Date: Mar 2007
Location: UK
Posts: 472
|
A slight thread hijack: Are there any free EQs that have a 'slectatable' high dB slope?
__________________
Bought REAPER V1.5 and still going strong today with V5.
Thanks Justin & Co !
|
|
|
06-24-2017, 09:11 AM
|
#31
|
Human being with feelings
Join Date: Nov 2015
Posts: 642
|
The EQs from Tokyo Dawn Labs have various slopes to select from. I can recommend those for most tasks. Beside that the the auto gain feature is well implemented and pure gold once you get used to that. Honestly it's hard to not stick with it after a while. No more ear trickery in that regard. They replaced Equilibrium in all of my sessions.
|
|
|
06-24-2017, 12:49 PM
|
#32
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
Quote:
Originally Posted by bezusheist
the "12db/oct" figure is measured from "0", not the cutoff's -3dB.
for example...ReaEQ with a HPF set @ 100 Hz...one octave down (50 Hz) is -12 dB...
|
Yes, I didn't pinpoint the position where it starts down, that might have been a little useful.
It's a very simple test and I might do it over again and save some images of it.
EDIT: Actually as I look at your picture, it appears that the point where it starts to drop down is an octave above the target frequency.
|
|
|
06-24-2017, 02:02 PM
|
#33
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
Okay, I went ahead and tested again, this time with white noise. The first picture is mainly calibrating with no EQ invoved.
The 2nd picture is with the high pass set to 200hz. It does appear that the high pass actually starts 1 octave (400hz) above the target frequency.
The 3rd picture is with 2 tabs of high pass set at 200hz. It's good to see the starting point is the same octave above the target.
However, something of significance, look at how the volume has increased with each high pass tab engaged. I find that a little alarming, however, I do remember a thread where we discussed this before. I wish I could remember where it is.
|
|
|
06-24-2017, 06:08 PM
|
#34
|
Human being with feelings
Join Date: Sep 2007
Location: The Reflection Free Zone
Posts: 3,026
|
Quote:
Originally Posted by Tod
Okay, I went ahead and tested again, this time with white noise. The first picture is mainly calibrating with no EQ invoved.
..
However, something of significance, look at how the volume has increased with each high pass tab engaged. I find that a little alarming, however, I do remember a thread where we discussed this before. I wish I could remember where it is.
|
Hi Tod. Sometimes filtering can produce peaks higher than the input signal, but using white noise it shouldn't. I just tried using the same ReaEQ filters as you, but didn't get any higher peak in my filtered signal.
I'll upload and link my white noise sample for you if you'd like to compare https://drive.google.com/open?id=0Bx...TNuVm5aQXR2cmM
You might want to look at my REAPER project too (attached below)
I wonder if the difference is the super loud custom buttons in your theme?
Joking aside, perhaps try with SPAN out of the chain? Or another possibility, you are using a 'live' generator (I don't see any media file in your track) and there is a +/- 3 dB variation in the run. Ok I see the track output is a consistent -1 dBFS.
Last edited by hamish; 06-24-2017 at 06:25 PM.
|
|
|
06-24-2017, 06:19 PM
|
#35
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
Quote:
Originally Posted by hamish
Hi Tod. Sometimes filtering can produce peaks higher than the input signal, but using white noise it shouldn't. I just tried using the same ReaEQ filters as you, but didn't get any higher peak in my filtered signal.
I'll upload and link my white noise sample for you if you'd like to compare https://drive.google.com/open?id=0Bx...TNuVm5aQXR2cmM
|
Okay hamish, I just used the JS White Noise plugin, but I'll be glad to check your sample out.
Okay I got and will check.
|
|
|
06-24-2017, 06:49 PM
|
#36
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
Yeah, the frequency response outcome of the filters is pretty much the same with both tests, but like you said, your white noise doesn't create the higher outputs. I'm not sure why that is, heh heh, a little voodoo maybe.
I think the next step would be to test it with some orchestra samples. I think I can find some time tomorrow to do that.
Thanks hamish, and do you know why the difference in output of the two different signals. Obviously there's some hidden factors involved here, and I'm not sure what they are.
|
|
|
06-24-2017, 07:38 PM
|
#37
|
Human being with feelings
Join Date: Sep 2007
Location: The Reflection Free Zone
Posts: 3,026
|
Quote:
Originally Posted by Tod
Thanks hamish, and do you know why the difference in output of the two different signals. Obviously there's some hidden factors involved here, and I'm not sure what they are.
|
Tod, no I really don't know why you're seeing the higher peak when the filters are in. Most times when something like this happens in a mix it's due to a compressor detector not pulling down as much when you remove bass, or a resonance at the cutoff, or allegedly (I've never proved this myself yet) because of phase shift caused by the filter, but as I understand that should only happen in harmonic material, and not in white noise.
Have you tried using my white noise sample in your project?
|
|
|
06-24-2017, 08:00 PM
|
#38
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
Quote:
Originally Posted by hamish
Have you tried using my white noise sample in your project?
|
Yes, that's what I was saying, your sample didn't cause a rise in the output.
What ever it is that causes it, it appears the spectrum analyzer doesn't see it either.
I'll check it out with orchestra samples tomorrow.
|
|
|
06-24-2017, 09:57 PM
|
#39
|
Human being with feelings
Join Date: Nov 2010
Location: Mullet
Posts: 829
|
It is common to get a change in level when using (min. phase) HPF or LPF on a "complex" waveform due to phase shift.
Quote:
Originally Posted by Tod
Yes, that's what I was saying, your sample didn't cause a rise in the output.
|
try a HPF around 50 Hz +/-...you should get a slight rise in level.
the sample hamish posted is Gaussian white noise, which is different than white noise. It is a lot more "stable" when filtered (ie not such drastic level changes (if any) from phase shift).
__________________
I like turtles
Last edited by bezusheist; 06-25-2017 at 02:27 AM.
|
|
|
06-25-2017, 09:27 AM
|
#40
|
Human being with feelings
Join Date: Jan 2010
Location: Kalispell
Posts: 14,745
|
Quote:
Originally Posted by bezusheist
It is common to get a change in level when using (min. phase) HPF or LPF on a "complex" waveform due to phase shift.
try a HPF around 50 Hz +/-...you should get a slight rise in level.
the sample hamish posted is Gaussian white noise, which is different than white noise. It is a lot more "stable" when filtered (ie not such drastic level changes (if any) from phase shift).
|
Okay, thanks bezusheist.
|
|
|
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 04:24 AM.
|