Voxengo Premium Membership - All Voxengo Plugins For a Fixed One-Time Fee
Forums     Plugins     Soniformer Sonic Finalizer Detector Type ??

This topic was created before release of the latest product version, and it may contain details irrelevant to this version.  Replying is disabled for this topic.

Last

Next

Previous

First



Hi Aleksey,

Does Sonic Finalizer use a peak detector or rms, or even average ?  I'm in to rms or rms average lately to solve a problem I have and am wondering which plugs of yours use an rms detector in the compressor.

Thanks,

Kylen


It uses peak detector.

I will think about an RMS detector, but it will make plug-in consume much more CPU.  It will require to calculate square root for all 32 bands, and that's not what current CPUs are capable of.  Running average looks like a better solution.  I'll check it in the near future and if it works good, I'll incorporate it into SF.


Thanks for thinking about this Aleksey.

I have a full mix where the variation of the rms levels is too great meaning it was mixed badly.  The apparant loudness of the song changes during listening and you want to reach for the volume knob.  I'd rather not try to deal with this using a tool with a peak detector even though potentially there are ways opening up the attack.  Still the detector kicks in the compressor every time a peak crosses the threshold.

In other words if the problem I'm trying to solve is an apparant loudness issue then I'd like to use a tool with a detector that is at least listening to the loudness (rms or non-rms average would work I think) instead of the peaking of the dynamics.

So if you get to updating Sonic Finalizer with an averaging detector (and selector switch I assume) then I would be among the first to try it out to see if it can handle this type of problem that is normally fixed by re-mixing.

A tool with 32 bands listening to the average level would be quite nice and very smooth and transparant.  Now the Sonic Finalizer could manage dynamic headroom (average rms db to peak db) as well as apparant loudness (normally average rms to max rms - but non-rms average may work).

Thanks,

kylen


Sorry, but why don't you just automate the volume, how fast can these volume changes be?

Hi Andrew,

Thanks for thinking about this.  I left out the part about the volume changes only occur on a few one-third octave bands - maybe a half dozen or so.  It's when the bass guitar, vocal, and lead guitar poke out from the 'average' or rms level too much in a few spots to upset the dynamic balance.  I could turn down the volume of the entire wave but they would still be sticking out.

That's the kind of thing a multiband compressor usually can deal with, all of mine use Peak detectors though so the compressor would only turn on when the sound passes the threshold of the peaks.  I could leave the compressor on all the time by setting the threshold lower and use attack/release to try and compress the stuff I want (which I'm trying now) but that wouldn't be as accurate I don't think...

Now if Aleksey were to include an averaging detector on the 32 band Sonic Finalizer that might just do the trick.  I couldn't say till I tried it out though.

kylen


I see your problem now :-) Perhaps you could automate an eq.  I notice that CurveEQ is not automatable in SONAR with VSTAdaptor -- perhaps this would be impossible because there are a potentially unlimited (?) number of control points, and how would they be identified?  Is there any way round this, Aleksey?

No, CurveEQ was not designed to support automation.  The fact you can define many control points is one of the reasons automation was not implemented.

Actually, since all plug-ins I develop are VST effects, I never try to implement any strong automation features.


Hi Andrew,

The problem of controlling the dynamics of certain frequency ranges is one that multiband dynamic processors like Sonic Finalizer were designed to deal with.

I'm not saying that your idea of automating EQ won't work but you can imagine trying to move a few frequency ranges around with a ramp time in the milliseconds for "x" number of milliseconds every "n" interval - in addition with no gain reduction lights on an EQ.  No Thanks !

The great power of Sonic Finalizer to control the dynamic range of 32 bands of frequencies using its' peak detection is already evident.  And by opening up the attack time > 50ms along with dropping the threshold and playing with the release time I can get it to see below the peak 'envelope' to get the compressor to stay on more of the time to adjust loudness somewhat.  But it is not following the loudness envelope.

If the detector were listening to the rms envelope then it would be listening to approximately the same thing I'm listening to in this case - the apparant loudness.  That is a valid compressor technique, in fact that's why there are rms switches on compressors for when you want to adjust apparant loudness dynamics.  Adjusting peak dynamics and dynamics range is stightly different and requires the peak detector.  But the 2 techniques are commonly used together to accomplish a more smooth natural sound.

Aleksey mentioned that he may investigate what the rms or average detector would do for Sonic Finalizer.  If that was added (and it was useful) then it would put Sonic Finalizer way ahead of other multibands.  My research shows that you won't find multiband processors with rms/average detectors until you get out into some of the expensive hardware types.

Whenever you want a tester Aleksey I'm here !

Thank you,

kylen


Thanks for the explanataion -- your wish sounds interesting :-)

OK, kylen, I'll post some info when the newer version of SF will be available.
This topic was created before release of the latest product version, and it may contain details irrelevant to this version.  Replying is disabled for this topic.

Last

Next

Previous

First