KVR Audio

sweet_trip · Post by **sweet_trip** » Fri Jul 07, 2006 12:20 am

Man, Voxengo has so many different EQs and compressors and limiter, it's a bit confusing choosing a winning combo. I will be trying out the demos tonight, but I was wondering what are the best combos some of you guys use, just to get an idea.

Aleksey Vaneev · Post by **Aleksey Vaneev** » Fri Jul 07, 2006 2:50 am

There are basically only three EQs: CurveEQ, GlissEQ and HarmoniEQ. CurveEQ is a mastering equalizer mainly. GlissEQ is a track equalizer mainly. HarmoniEQ is mastering equalizer as well, and probably strictly mastering EQ due to its heavy CPU load. They all have very different character and behaviour.

I suggest you to check Product Groups on the Voxengo website. The way these groups were organized is revealing as to in what situation plug-ins should be used (mixing or mastering, and which features they posses if we are talking about multi-functional plug-in).

defjamm · Post by **defjamm** » Fri Jul 07, 2006 3:06 am

hi aleksey, i have also some questions. won't you implement an even higher oversampling-mode(just like in elephant) in your eqs or do you think the benefit wouldn't be that big? i mean just as an option.
why is there no oversampling in curve-eq? and question 3, when is marquis going to have an oversampling-mode?

Aleksey Vaneev · Post by **Aleksey Vaneev** » Fri Jul 07, 2006 3:49 am

CurveEQ does not require oversampling, because it is linear-phase EQ. Oversampling may be useful with 'bread and butter' filters which may have phase distortions when they are tuned to frequency close to Nyquist. If such filter is 'perfect', even oversampling is unnecessary.

Just to clear things up a bit, oversampling by itself does not increase quality. It may adjust sound character (if polyphase filter is used) and can provide spectral space for saturation plug-ins.

I do not plan to implement oversampling in Marquis right now, because it is already a pretty CPU-heavy plug-in for channel use (its 2 times more CPU heavy than, for example Sonitus Compressor). Moreover, I do not see this will increase its quality a lot (I've already took care about possible aliasing). Oversampling may change the overall coloration, though. But this may not be desirable in all cases.

defjamm · Post by **defjamm** » Fri Jul 07, 2006 7:17 am

Aleksey Vaneev wrote:If such filter is 'perfect', even oversampling is unnecessary.

i would stress the IF, a lot are not.

Aleksey Vaneev wrote: Just to clear things up a bit, oversampling by itself does not increase quality. It may adjust sound character (if polyphase filter is used) and can provide spectral space for saturation plug-ins.

i know that. so it has its uses with certain applications and actually increases quality with them(saturaion, distortion).

Aleksey Vaneev wrote: I do not plan to implement oversampling in Marquis right now, because it is already a pretty CPU-heavy plug-in for channel use (its 2 times more CPU heavy than, for example Sonitus Compressor). Moreover, I do not see this will increase its quality a lot (I've already took care about possible aliasing). Oversampling may change the overall coloration, though. But this may not be desirable in all cases.

if you have took care of possible aliasing than great.

one thing, this goes to all developers: i don't want sound sacrificed for cpu. there are 1000 plugins which do that. if you can make it sound great without much cpu-consumption than ok, but if not than i take quality over quantity any day.

Aleksey Vaneev · Post by **Aleksey Vaneev** » Fri Jul 07, 2006 6:10 pm

defjamm wrote: one thing, this goes to all developers: i don't want sound sacrificed for cpu. there are 1000 plugins which do that. if you can make it sound great without much cpu-consumption than ok, but if not than i take quality over quantity any day.

Sure, this is reasonable, but according to my practice a lot of people would like to have plug-ins with minimal CPU load.

Klemperer · Post by **Klemperer** » Fri Jul 07, 2006 6:39 pm

Aleksey Vaneev wrote:
Sure, this is reasonable, but according to my practice a lot of people would like to have plug-ins with minimal CPU load.

That's right, for example me

. As long as fine new things are optional, and don't force the effect to overheat your CPU (always thinking in those categories now in my flat

), it may be nice. Thanks for the explanation about HarmoniEQ which I own - I don't have CurveEQ (but demo-ed it) and now I know it's lighter on the CPU like the wonderful HarmoniEQ.

Stupid American Pig · Post by **Stupid American Pig** » Fri Jul 07, 2006 6:42 pm

Aleksey Vaneev wrote:
defjamm wrote: one thing, this goes to all developers: i don't want sound sacrificed for cpu. there are 1000 plugins which do that. if you can make it sound great without much cpu-consumption than ok, but if not than i take quality over quantity any day.
Sure, this is reasonable, but according to my practice a lot of people would like to have plug-ins with minimal CPU load.

This forum has proven it. just look at all the cpu hog threads. and it seems to me that most of the time you only get a 5% improvment in sound with a 50% increase in cpu usage, so most people dont hear the difference anyway...

Aleksey Vaneev · Post by **Aleksey Vaneev** » Fri Jul 07, 2006 7:20 pm

Nevertheless, people should not forget about the fact that minimizing CPU load minimizes creativity. Efficient audio processing in most aspects means simplistic audio processing (especially if the most efficient building blocks were used - like FFT).

For example, I may be interested in creating a 32-band 'classic' stereo vocoder, but its CPU load will be pretty high - something like 30% on Athlon XP 3000+. I doubt many users will be happy with such performance. But 5-band vocoder with 5% load is not the same thing at all.

Dandruff · Post by **Dandruff** » Fri Jul 07, 2006 7:24 pm

vocoder -> yeah! need it

phrauge · Post by **phrauge** » Fri Jul 07, 2006 7:28 pm

Aleksey Vaneev wrote:I may be interested in creating a 32-band 'classic' stereo vocoder, but its CPU load will be pretty high - something like 30% on Athlon XP 3000+.

I'll gladly take a 30% CPU hit if your Vocoder works as well as your other plugs.

kylen · Post by **kylen** » Fri Jul 07, 2006 7:52 pm

I really like the optional and selectable oversampling feature in Elephant. I thought I'd heard everything the plug had to offer till I turned on 4x oversampling just now. In fact I was wondering if there might be use for a plug that would encode/decode oversample kind of like MSED - assuming the host could handle it. I currently don't have the firepower (2.6MHz P4 w/1gb ram) to handle 96KHz native wave files, but 441.KHz wav files with certain oversampled plugs seems to sound real good.

Cybotron · Post by **Cybotron** » Fri Jul 07, 2006 8:39 pm

phrauge wrote:
Aleksey Vaneev wrote:I may be interested in creating a 32-band 'classic' stereo vocoder, but its CPU load will be pretty high - something like 30% on Athlon XP 3000+.
I'll gladly take a 30% CPU hit if your Vocoder works as well as your other plugs.

Me too!

Aleksey Vaneev · Post by **Aleksey Vaneev** » Sat Jul 08, 2006 1:51 am

kylen, VST spec does not allow upsampled/downsampled streams, so 'sort of MSED' is not possible to implement. Such universal 'oversampler' has to be a VST host itself.

kylen · Post by **kylen** » Sat Jul 08, 2006 3:02 am

Oh - OK thanks Aleksey I didn't know. I guess you would have already considered making it if it were possible!

Sort of dumb Voxengo question...