VST Request - Crest Factor Decreaser

DSP, Plug-in and Host development discussion.
User avatar
Vospi
KVRist
210 posts since 1 Jun, 2011

Post Thu Jan 19, 2017 5:54 pm

(0) I would appreciate a PM with your results with Stereo Tool, before and after.
(1) There's possibly no reliable correlation between the process you've tried and RMS/Crest.
(2) There's no processor that will achieve what you're asking for without severe damage to a signal, especially in regard of recreating peaks back. There's a whole industry in place trying to achieve any kinds of automatic results in that regard, and no serious sound guy uses them automatically or relies on them at all, ever.

- If you're trying to find a solution for broadcast levelling purposes, there are other, much more elaborate ways.
- If you want to process your own tracks, you should really mix for the purpose, and no processor would ever be able to replace it, short of using machine learning like Prisma in somewhat distant future.
- If the interest is hypothetical, then you should start with understanding that RMS doesn't equal loundess, peak doesn't equal punchiness; these factors in real life are all greatly frequency-dependent and time dependent; I can create a sound that will have a great contrast in terms of volume perception and close to none in terms of crest and all that. And any kind of "expanding" would probably just ruin it, really, even if tuned manually and very carefully. I can as well create a sound that will have big RMS, nice crest factor and no punchiness at all!

RMS and Crest are used solely for human reference and are not nearly enough for these kinds of automatic decisions.

Also, note: you're using metrics that accumulate over time, which is a super-unreliable idea for judging microdynamics, which is what you're after, ultimately.

P.S.: With all that said, "Crest Factor Decreaser" is just a not gained limiter, really. :)
Last edited by Vospi on Thu Jan 19, 2017 7:08 pm, edited 9 times in total.
http://vospi.com/music
http://soundcloud.com/vospi
more than 1000 followers, licensed on 3 continents.
русский подкаст о звуке: https://vk.com/stereopodcaster

JCJR
KVRAF
2360 posts since 17 Apr, 2005 from S.E. TN

Re: VST Request - Crest Factor Decreaser

Post Thu Jan 19, 2017 6:59 pm

Hi MannyAudio85

Apologies if you already know all about this, but here are a couple of explanations by Bob Orban and others about phase rotation and other topics related to orban radio processors.

http://www.orban.com/orban/support/orba ... th_1.3.pdf

http://www.261.gr/orbanfoti.html

So far as I can tell, the phase rotation is primarily to reduce the crest factor of radio speech, especially male announcers, which according to Orban tend to be rather unsymmetrical waveforms with a large crest factor. The broadcast processor's low-midrange phase rotation makes speech more symmetrical, allowing more compression, bigger gain without bad side-effects or clipping.

However, phase rotation will not reduce the crest factor of all waveforms. For instance phase rotation drastically increase the crest factor of a square wave. The RMS power stays the same, and the square wave signal in most cases sounds the same (if un-clipped and no post-processing gain adjustment) but the peaks get louder compared to the RMS.

mannyaudio85
KVRer
10 posts since 19 Jan, 2017

Re: VST Request - Crest Factor Decreaser

Post Thu Jan 19, 2017 8:10 pm

JCJR wrote:Hi MannyAudio85

Apologies if you already know all about this, but here are a couple of explanations by Bob Orban and others about phase rotation and other topics related to orban radio processors.

http://www.orban.com/orban/support/orba ... th_1.3.pdf

http://www.261.gr/orbanfoti.html

So far as I can tell, the phase rotation is primarily to reduce the crest factor of radio speech, especially male announcers, which according to Orban tend to be rather unsymmetrical waveforms with a large crest factor. The broadcast processor's low-midrange phase rotation makes speech more symmetrical, allowing more compression, bigger gain without bad side-effects or clipping.

However, phase rotation will not reduce the crest factor of all waveforms. For instance phase rotation drastically increase the crest factor of a square wave. The RMS power stays the same, and the square wave signal in most cases sounds the same (if un-clipped and no post-processing gain adjustment) but the peaks get louder compared to the RMS.
So, going along with this information - can a "rotation" filter not be created that would "drastically increase" (or decrease - closer to 0 not further away from it) the crest factor of a sine wave? Aren't most sub-bass tones just sine waves?

Thanks.

JCJR
KVRAF
2360 posts since 17 Apr, 2005 from S.E. TN

Re: VST Request - Crest Factor Decreaser

Post Thu Jan 19, 2017 8:48 pm

mannyaudio85 wrote: So, going along with this information - can a "rotation" filter not be created that would "drastically increase" (or decrease - closer to 0 not further away from it) the crest factor of a sine wave? Aren't most sub-bass tones just sine waves?
Hi Manny

So far as I know, a static steady sine wave test tone will always have the same crest factor because it is a "pure tone" with only one harmonic. If we "static" phase shift a sine wave, regardless how much, we still have the same sine wave and under most circumstances will always sound the same and have the same peak-to-RMS ratio.

If we dynamically phase shift the steady sine wave test tone, as with a guitar phase shifter effect which is always changing the phase, a "moving phase shift", the dynamic changing phase would affect the pitch of the sine wave, adding pitch bend or vibrato effects, but I think the crest factor would stay the same.

There is a psycho-acoustic effect (at least to my ear) that fast pitch vibrato can sound louder than a no-vibrato steady tone, but a good level meter should read the same amplitude even if the ear is fooled into thinking the vibrato made it sound louder.

The reason a square wave (and some other simple geometric steady waves) can have bigger peak to RMS ratio when phase-shifted, is that they contain a certain mix of many harmonics, and the square wave is the "lowest crest factor" most compact way to mix that ratio of harmonics together. The lowest peak-to-RMS ratio. Similarly I think the certain mix of harmonics that makes a triangle wave or sawtooth wave-- Any other scrambled-phase mix of the same harmonics and levels would have a bigger crest factor than the original triangle or sawtooth.

If we phase shift the square wave, some of the harmonics are moved out of this "most compact" phase relationship, so it generally sounds the same to the ear, but it doesn't look like a square wave anymore, and the peaks of the wave get bigger, though the RMS of the entire wave cycle stays the same. Just that some places in a single-cycle of the square wave don't get as loud after the phase shift, and other places in the single-cycle get louder after the phase shift, but the RMS or "average" of all the many harmonics stays the same.

The crest factor in actual music or speech is measured the same as steady test tones, but it "means something else". For instance you could make a song with nothing but many tracks of pure sine waves playing bass, chords, melody and even percussion. You could add VCA amplitude envelopes to make the different sine wave tracks sound a little different. Maybe fast attack medium release amplitude envelope on the sinewave bass. Maybe fast attack fast release "sine wave blips" for the drums. Maybe medium attack, medium release amplitude envelope on the sinewave chord track notes. Whatever.

Adding the several tracks together, they would peak loud at certain times, and they would also add together for a long-term average RMS level. So the crest factor would be the loudest transient peaks, maybe when the bass sinewave, chords sinewave, and drums sinewaves add together real loud for an instant. The ratio between the occasional loudest transient peaks versus the long-term RMS.

I'm just saying, a compressor or expander can't change the crest factor of a steady test tone (except possibly temporarily when the tone starts or the tone stops). But a compressor or expander could change the crest factor of the above described song, even if all the notes in the song are pure sine waves. Because the song is not steady and un-changing as is the case with a test tone.

Maybe phase shifting with allpass filters would also affect the crest factor of our song made out of many pure sine waves. Because the phase shift would change the "micro-alignment" of the different sine waves in the different tracks, and so they would add together to make a different set of "instantaneous peaks" at certain times in the song.

Hopefully the above is fairly accurate and understandable. :)

mannyaudio85
KVRer
10 posts since 19 Jan, 2017

Re: VST Request - Crest Factor Decreaser

Post Thu Jan 19, 2017 9:20 pm

JCJR wrote:
mannyaudio85 wrote: So, going along with this information - can a "rotation" filter not be created that would "drastically increase" (or decrease - closer to 0 not further away from it) the crest factor of a sine wave? Aren't most sub-bass tones just sine waves?
Hi Manny

So far as I know, a static steady sine wave test tone will always have the same crest factor because it is a "pure tone" with only one harmonic. If we "static" phase shift a sine wave, regardless how much, we still have the same sine wave and under most circumstances will always sound the same and have the same peak-to-RMS ratio.

If we dynamically phase shift the steady sine wave test tone, as with a guitar phase shifter effect which is always changing the phase, a "moving phase shift", the dynamic changing phase would affect the pitch of the sine wave, adding pitch bend or vibrato effects, but I think the crest factor would stay the same.

There is a psycho-acoustic effect (at least to my ear) that fast pitch vibrato can sound louder than a no-vibrato steady tone, but a good level meter should read the same amplitude even if the ear is fooled into thinking the vibrato made it sound louder.

The reason a square wave (and some other simple geometric steady waves) can have bigger peak to RMS ratio when phase-shifted, is that they contain a certain mix of many harmonics, and the square wave is the "lowest crest factor" most compact way to mix that ratio of harmonics together. The lowest peak-to-RMS ratio. Similarly I think the certain mix of harmonics that makes a triangle wave or sawtooth wave-- Any other scrambled-phase mix of the same harmonics and levels would have a bigger crest factor than the original triangle or sawtooth.

If we phase shift the square wave, some of the harmonics are moved out of this "most compact" phase relationship, so it generally sounds the same to the ear, but it doesn't look like a square wave anymore, and the peaks of the wave get bigger, though the RMS of the entire wave cycle stays the same. Just that some places in a single-cycle of the square wave don't get as loud after the phase shift, and other places in the single-cycle get louder after the phase shift, but the RMS or "average" of all the many harmonics stays the same.

The crest factor in actual music or speech is measured the same as steady test tones, but it "means something else". For instance you could make a song with nothing but many tracks of pure sine waves playing bass, chords, melody and even percussion. You could add VCA amplitude envelopes to make the different sine wave tracks sound a little different. Maybe fast attack medium release amplitude envelope on the sinewave bass. Maybe fast attack fast release "sine wave blips" for the drums. Maybe medium attack, medium release amplitude envelope on the sinewave chord track notes. Whatever.

Adding the several tracks together, they would peak loud at certain times, and they would also add together for a long-term average RMS level. So the crest factor would be the loudest transient peaks, maybe when the bass sinewave, chords sinewave, and drums sinewaves add together real loud for an instant. The ratio between the occasional loudest transient peaks versus the long-term RMS.

I'm just saying, a compressor or expander can't change the crest factor of a steady test tone (except possibly temporarily when the tone starts or the tone stops). But a compressor or expander could change the crest factor of the above described song, even if all the notes in the song are pure sine waves. Because the song is not steady and un-changing as is the case with a test tone.

Maybe phase shifting with allpass filters would also affect the crest factor of our song made out of many pure sine waves. Because the phase shift would change the "micro-alignment" of the different sine waves in the different tracks, and so they would add together to make a different set of "instantaneous peaks" at certain times in the song.

Hopefully the above is fairly accurate and understandable. :)
Very good post, thank you.

So, creating a filter that manipulates the audio, in whatever way it needs to, to lower the crest factor to a value closer to 0 while not modifying the RMS, is an impossible task?

Sure, using Fabfilter Pr-G I can definitely do some upward expansion. This does help in some cases more than in others but it has it's limitations...A ratio beyond 1:5 will be problematic and expansion in general has it's problems. This is the reason that I proposed the idea of modifying the crest factor as opposed to the RMS value through expansion to achieve a larger difference between both values.

If we know which sounds in a song will lower the crest factor to a value closer to 0--why can a filter not scan the audio that it is presented, increase those sounds accordingly, and present an output waveform with the desired changes and RMS value in tact?

Again, I apologize for my ignorance here and/or improper use of audio terminology. I have a desired goal in mind and am more than willing to pay someone much smarter than myself to create this for me. I hope the general conclusion is that it cannot be done!

Thanks.
Last edited by mannyaudio85 on Thu Jan 19, 2017 9:35 pm, edited 1 time in total.

camsr
KVRAF
6890 posts since 17 Feb, 2005

Re: VST Request - Crest Factor Decreaser

Post Thu Jan 19, 2017 9:30 pm

Ichad.c wrote:You can not reduce the Dynamic Range(Crest) without increasing the RMS while normalizing to 0dbfs, the math and physics just don't line up. Crest in a nut-shell is Peak minus RMS. When you change the one, you change the other.
Image

mannyaudio85
KVRer
10 posts since 19 Jan, 2017

Re: VST Request - Crest Factor Decreaser

Post Fri Jan 20, 2017 4:37 am

camsr wrote:
Ichad.c wrote:You can not reduce the Dynamic Range(Crest) without increasing the RMS while normalizing to 0dbfs, the math and physics just don't line up. Crest in a nut-shell is Peak minus RMS. When you change the one, you change the other.
Ok so I gave this more thought overnight and did a quick test this morning to confirm exactly what it is I am after.

I considered this sentence from JCJR: "For instance phase rotation drastically increase the crest factor of a square wave."

I took a 1 minute, raw drum loop that I had on the computer. This has file has a DR19 and -20.02 RMS reading--a 1.02db difference.

I opened that file in my DAW and generated a 1 second, 60hz sine wave. I normalized the sine wave to 0dbfs and mix-pasted it into the drum loop at the 26 second mark. I scanned the drum loop with the added sine wave again, and now I have a DR9 with -15.34db RMS value--a 6.34db difference! But when I mixed that same sine wave into the drum loop without normalizing the sine wave (so it did not hit 0dbfs, or even come close) the crest factor and RMS values were unchanged.

So, what I want is clear now: If a typical phase rotator increases square wave peaks, and decreases sine wave peaks (among other things), I want a 32-bit VST2 plugin that will push all sine waves and sawtooth waves below 100hz to full-scale 0dbfs. If there is a 100ms 60hz sine wave with a 100ms 60hz sawtooth wave 1db softer, I want the sine to be at 0dbfs for that 100ms and the sawtooth to be at -1dbfs for the 100ms. So if a song is meant to have a 1db difference between the sawtooth and sine for that 100ms, it remains 1db difference--only the louder of the two is pushed to 0dbfs full scale and the other is pushed enough so the intended difference remains.

This should be able to be done, since we can easily do the opposite with a normal phase rotator without adding unpleasing artifacts. Could sine/saw waves be pushed to 0dbfs byphase rotation/manipulation?

I apologize for not properly describing or knowing how to describe what it I needed from the beginning.

If anyone is willing to make this for me, please shoot me a PM

Thanks.

JCJR
KVRAF
2360 posts since 17 Apr, 2005 from S.E. TN

Re: VST Request - Crest Factor Decreaser

Post Fri Jan 20, 2017 12:25 pm

mannyaudio85 wrote: So, creating a filter that manipulates the audio, in whatever way it needs to, to lower the crest factor to a value closer to 0 while not modifying the RMS, is an impossible task?
Hi Manny

I don't recall and would have to look-up "standard definitions" of crest factor. Of course it is the ratio (measured in percent amplitude or dB difference) between RMS and Peak. But both RMS and Peak can be measured over different time scales.

You could measure RMS over very short time intervals, which would always be rapidly changing, but each measurement would be accurate for that short time window. For instance if you measure the RMS with a time window of 1 ms, then you would get 10 different RMS measurements per wave cycle, measuring a 100 Hz audio wave. Well actually if you calculate the 1 ms RMS windows overlapped, you could get any number of smoothly varying but different RMS measurements in each 100 Hz wave cycle. In other words, the RMS is measuring accurately according to its definition, but it is riding up and down on waves lower-frequency than the RMS time-window.

Similarly you could measure RMS over a 1 second interval, and smoothly get as many measurements per second as you like by "overlapping" the measurements. For instance at a samplerate of 44.1 kHz, on every new sample you could add the newest sample and discard the oldest sample from 1 second ago and recalc the RMS. So you could get 44100 slightly different one-second-time-window RMS measurements per second.

The longest-term RMS for a song, would be the entire song's duration calculated as the RMS window. If the song is 5 minutes long, we only have one measurement for the average power level of the entire 5 minute song.

It is the same way with peaks. We could remember the biggest peak that happened in the last 1 millisecond, or remember the biggest peak that happened in the last 1 second, or remember the biggest peak that happened in the entire song. Or remember some time-smoothed average of the peaks that have occurred, either short-term or long-term peak averages.

Am fairly certain that there are standard smoothing time scales used to make "standard" crest factor numbers, possibly many different standards, but would need to look it up again. You would need to measure RMS to the "standard time scale" and measure peak to the standard time scale, then calc the percent or dB ratio of the two to get the crest factor (according to whatever standard you want to use).

Whatever task we are interested in, determines the time scale we would care about.

JCJR
KVRAF
2360 posts since 17 Apr, 2005 from S.E. TN

Re: VST Request - Crest Factor Decreaser

Post Fri Jan 20, 2017 12:32 pm

Long Time Intervals Crest Factor

For instance if we want to digitally normalize a song as loud as it can get without clipping, we find the loudest sample in the entire song or audio file, then we multiply all the samples by whatever factor would raise the loudest sample to full-scale (which is typically +/- 1.0 in floating point digital audio).

So if we scan a song and the loudest sample happens to have an absolute value of 0.58, then we can multiply all the samples in the song by 1.0 / 0.58 = 1.724. That would be applying gain of 20 * log10(1.724) = 4.73 dB. The loudest sample will now be 1.0 (or -1.0 if the sample was negative), and of course all the other samples will have a smaller absolute value than +/- 1.0. That is the loudest we can "turn up" the song without clipping at least one sample in the song.

If a song's very loudest peak happens to be 1.0, 0 dB full scale, then if the full-song RMS is -12 dB, then we have a 12 dB crest factor for the entire song BY THAT MEASUREMENT DEFINITION-- The one loudest peak compared to the full-song RMS. But there could be other measurement definitions in various meters. Probably some more-useful, better thought out definitions.

This very-long timescale crest factor could matter for instance if the entire song is recorded rather quiet-- A low full-song RMS, but there is only one loud peak, or only a few loud peaks in the entire song. We can't make the song loud without clipping just a small number of extraordinarily loud samples amongst the quiet ones. Things like peak limiters or smooth clippers can "transparently" deal with rare really loud peaks, tame the rare peaks one way or the other hopefully without the ear noticing damage. In that case, just knocking out a handful of peaks could vastly reduce the crest factor.

Peak Limiters, compressors and expanders are nonlinear devices, changing gain over time. They basically do their job by distorting the audio, but are designed to do the distortion in a gradual, slow way that hopefully does not offend the ear.

mannyaudio85
KVRer
10 posts since 19 Jan, 2017

Re: VST Request - Crest Factor Decreaser

Post Fri Jan 20, 2017 12:55 pm

@JCRC

I appreciate all of that information but I am already well aware of those things.

What I am requesting is a VST plugin that will amplify all sine and saw waves 100hz and below that the pligin encounters to 0dbfs...Nothing more, nothing less.

The form of measurement I use works wrll for me.

Can you create the plugin I am requesting? I am only interested in hiring a developer to create the filter I described 2 paragraphs up.

Can anyone do the job?

Thanks.

JCJR
KVRAF
2360 posts since 17 Apr, 2005 from S.E. TN

Re: VST Request - Crest Factor Decreaser

Post Fri Jan 20, 2017 1:53 pm

mannyaudio85 wrote:
camsr wrote:
Ichad.c wrote: So, what I want is clear now: If a typical phase rotator increases square wave peaks, and decreases sine wave peaks (among other things), I want a 32-bit VST2 plugin that will push all sine waves and sawtooth waves below 100hz to full-scale 0dbfs. If there is a 100ms 60hz sine wave with a 100ms 60hz sawtooth wave 1db softer, I want the sine to be at 0dbfs for that 100ms and the sawtooth to be at -1dbfs for the 100ms. So if a song is meant to have a 1db difference between the sawtooth and sine for that 100ms, it remains 1db difference--only the louder of the two is pushed to 0dbfs full scale and the other is pushed enough so the intended difference remains.
Hi Manny

Are the 60 Hz sine wave and 60 Hz sawtooth waves on different tracks, or all glommed together in a mix?

If they are on their own tracks then probably a plugin could do frequency-selective auto-gain on the "control track" which needs to stay at 0 dB fs, and also "side chain" to adjust the level of the "slave track" which needs to stay at -1 dB less than the "control track".

If they are all together in a mix then I doubt if a program could figure out what to do with them. Mixed-together, if the sine and sawtooth are phase-locked in perfect sync with each other, a program (or ear) would hear a single sawtooth with a 6 dB boosted fundamental, rather than hearing a 60 Hz sawtooth and also a 60 Hz sine.

If the two waves are mixed-together and not phase-locked in perfect sync with each other, then there would be beating/cancellations between the sine and the first harmonic of the sawtooth, making the job even tougher on a program.

Even in the "more feasible" case of each sound on its own track, you need to specify timescales. If a sinewave note has a 10 ms attack and 100 ms release, should the program force the quiet attack to 0 dB, and raise gain on the release to force the release to 0 dB as well, until the note finally ceases?

Alternately, if you only want "the loudest instant" of each note at 0 dB, then unless the notes always have fast attack with the "loudest instant" very near the beginning of a note, then the plugin would need a very long lookahead time (or nonrealtime execution) in order to know what the "loudest instant" of each note is going to be.

stratum
KVRAF
2114 posts since 29 May, 2012

Re: VST Request - Crest Factor Decreaser

Post Fri Jan 20, 2017 2:01 pm

Maybe one could divide the signal into overlapping windows using STFT and find the phase relationship in each window that gives the minimum crest factor and synthesize the whole waveform again, I don't know if it would be possible, just another half-baked idea. it might introduce unwanted artifacts even if it works.
Assuming that the kind of artifacts it introduces would be acceptable, for each window an optimization problem needs to be solved, and one could ask if what we have is a suitable multidimensional surface using which an efficient solution can be found, as the number of parameters to play with is quite high.
~stratum~

mannyaudio85
KVRer
10 posts since 19 Jan, 2017

Re: VST Request - Crest Factor Decreaser

Post Fri Jan 20, 2017 10:19 pm

@JCJR

I am only talking about mixes already put together. I have sent you a PM with 2 short samples to demonstrate what the Stereo Tool phase rotation process does. The phase rotation in the samples sent to you were done to the all mixed together waveform.

@stratum

That's an interesting idea and discussions/ideas like this are exactly what I was looking for. Clearly, artifacts will be presented when any changes are made to the waveform--but if they would be unpleasant is a different story. I've sent you a PM as well with the aforementioned samples. What the phase rotator in the samples has done is essentially the opposite of what I wan to be done to the sine/sawtooth wave peaks. This leas me to believe the opposite should also be possible.

Any other thoughts/ideas guys?

User avatar
Vospi
KVRist
210 posts since 1 Jun, 2011

Re: VST Request - Crest Factor Decreaser

Post Fri Jan 20, 2017 11:26 pm

Either I'm very uneducated or this thread needs to be shaken.
This all just doesn't make any sense.

To mannyaudio who talks about:
If a typical phase rotator increases square wave peaks, and decreases sine wave peaks
Do you understand that a square wave is a lot of sinewaves that sound together? That ANY song and any sound could be considered "consisting of sinewaves"?
How do you "amplify waves below 100Hz to 0dbFS"... to 0dbFS of what, of a loudest peak? Grab a multiband limiter and clip away. Or to boom and hum all the time at the loudest possible value?.. What is 0dbFS for a part of spectrum, anyway? Your other frequencies still need space to fit somewhere.
Drop all the figures aside, what is the sonic result that you want to achieve?

To people quoting each other about "the math and physics just don't line up":
I've wrote for three times now that if you limit and don't normalize, it is perfectly possible. Am I wrong?

Also, this:
So, creating a filter that manipulates the audio, in whatever way it needs to, to lower the crest factor to a value closer to 0 while not modifying the RMS, is an impossible task?
IT'S A CLIPPER. OR A LIMITER. Anything that kills peaks. ANYTHING. NOT NORMALIZED BACK. What is unclear here?
..."Closer to zero", though? Not even to one, but to zero?..

Image

So, what do you want to have a value of zero, your peaks?..
White noise has a creast factor of about 1.8, white. Noise.
People decrease their DR without horrible artifacts by very craftful mixing, very careful sound choices, multiband compression, lots of automation. Which includes not having your "sinewaves" and whatnot occasionaly align momentarily in the first place.

What you're saying is almost like "I saw a ball falling down, it only makes sense that it would go up, too. Could it?". We say — theoretically, it could — but for how long and for what purpose and what's the size of the ball? And that's the place where all the conversation just breaks down... Your answer is something like "all balls. All balls should go up, that would be awesome."

It seems to me like this thread tries to invent a wheel.
Am I missing something vital here?
Last edited by Vospi on Sat Jan 21, 2017 12:20 am, edited 2 times in total.
http://vospi.com/music
http://soundcloud.com/vospi
more than 1000 followers, licensed on 3 continents.
русский подкаст о звуке: https://vk.com/stereopodcaster

User avatar
Vospi
KVRist
210 posts since 1 Jun, 2011

Re: VST Request - Crest Factor Decreaser

Post Fri Jan 20, 2017 11:59 pm

http://vospi.com/music
http://soundcloud.com/vospi
more than 1000 followers, licensed on 3 continents.
русский подкаст о звуке: https://vk.com/stereopodcaster

Return to “DSP and Plug-in Development”