From the original SynthMaker forum (2004–2014), with its members’ names and dates.
stefan ·
In the latest version there's a new stereo two-band Compressor module added to the toolbox.
I would like to get some feedback regarding its features, audio quality and whatnot. Feel free to post any comments or engage any discussions you want to.
For the curious ones, here's what the module looks like on the inside:
And here's the actual "inner-most" compressor module, fairly simple stuff when all is considered 🙂
Thanks to Malc for helping me out with some GUI issues and making me those new fancy Meter modules and thanks to Lalo for some early feedback and testing 🙂
PS. Try automating the "Ahead" knob. DS.
skOre ·
luv the ahead knob! Makes great strangeness when applied at runtime.
Nice job!
lalo ·
stefan wrote: and thanks to Lalo for some early feedback and testing 🙂
and thanx to you for the compressor 🙂
stefan wrote: PS. Try automating the "Ahead" knob. DS.
you..nasty guy 😀 😉
Shifrin ·
I didn't expect to find this when I got home tonight 😀 😀 😀
Cheers!
Shif.
fritzman ·
Ah yeah!
Now we're talking! This is the basis for something really serious me thinks. Sounds ok already. I compared it to Marquis which is my fav comp right now and it sounds ok in comparisson. Different sound, good. It has some modulation sounds always going on all the time.
The most obvious prob we have now is plugin delay. We need to have a way to tell the host the plugin delay so that it can be compensated. The plugin should then have the max ahead time as delay (dunno, maybe more, I didn't look into the schematic yet) and then the ahead time must be compensated inside the plugin.
Thanx a lot for this coooool module!!! 😀 😀 😀
All the best, FRitz
fritzman ·
It has denormal problems.
fritzman ·
When driving the compressors output into the softsat the level is too low. Got -5dB here now.
fritzman ·
It can sound really good:
SM_compressor.wav
original.wav
Best wishes, FRitz
sambean ·
Cheers Stefan,
Just got a chance to play with the version posted at KvR.. Good set of presets there, quite distinct sounding. I'm not qualified to rate this as a 'subtle' compressor but as an effect it's definately got character and good sounds..
Nice example Fritz, it really sounds more chunky with the compression..
It may be good to have seperate controls for each band, added versatility and all that..
Sam
Tom7777 ·
Nice example Stefan 🙂
stefan ·
fritzman wrote: Ah yeah!
Now we're talking! This is the basis for something really serious me thinks. Sounds ok already. I compared it to Marquis which is my fav comp right now and it sounds ok in comparisson. Different sound, good. It has some modulation sounds always going on all the time.
The most obvious prob we have now is plugin delay. We need to have a way to tell the host the plugin delay so that it can be compensated. The plugin should then have the max ahead time as delay (dunno, maybe more, I didn't look into the schematic yet) and then the ahead time must be compensated inside the plugin.
Thanx a lot for this coooool module!!! 😀 😀 😀
All the best, FRitz
Indeed that is how the PDC would have to work. I think two versions, one without PDC so that it can operate without latency and one with PDC so that you can use the look-ahead properly is the way to go. The rest of the compressor introduces no latency, just the look-ahead.
stefan ·
fritzman wrote: It has denormal problems.
Indeed it has, the issue of working solely on AMD 🙂 I will look into it ASAP.
fritzman wrote: When driving the compressors output into the softsat the level is too low. Got -5dB here now.
But it goes up to 0dB when you drive it harder. It can't start at 0dB and then go above that when you drive it harder, because then you'd still get digital clipping when using it to tame your master bus, unless you'd turn the master down for instance 5dB.
Opinions on this?
fritzman ·
stefan wrote:
fritzman wrote: It has denormal problems.
Indeed it has, the issue of working solely on AMD 🙂 I will look into it ASAP.
Thanx, Stefan.
stefan wrote:
fritzman wrote: When driving the compressors output into the softsat the level is too low. Got -5dB here now.
But it goes up to 0dB when you drive it harder. It can't start at 0dB and then go above that when you drive it harder, because then you'd still get digital clipping when using it to tame your master bus, unless you'd turn the master down for instance 5dB.
Opinions on this?
Hm, let me say it like this: I want it to reach 0dB more often than I want this super overdriven sound it has when I want to reach 0dB right now. 😉
So maybe a seperate drive knob would be cool? Then the output knob could be a pure level knob. I think a "non distorting" limiter would be better suited for the "don't go above 0dB" task.
It's not so easy to dial in the wanted loudness right now, because of the heavy distortion. Don't get me wrong, there are times when I want the distortion (and it sounds quite good) but not always. So I think that for general purpose compression it's more in the way then for the good.
All the best, FRitz
stefan ·
What about a switch so that you can toggle the soft-saturation on and off?
fritzman ·
stefan wrote: What about a switch so that you can toggle the soft-saturation on and off?
Hm, then I still won't get to 0dB when overdriving the softsat only softly. 😉
stefan ·
fritzman wrote:
stefan wrote: What about a switch so that you can toggle the soft-saturation on and off?
Hm, then I still won't get to 0dB when overdriving the softsat only softly. 😉
I don't see how one would achieve that. What you're asking for is something that is at 0dB with no overdrive and _stays_ at 0dB when you increase the overdrive.
I've got an idea (using a totally different drive algorithm) that could work like this though, I'll investigate it later on 🙂
The Sound Scientist ·
Hi there... 😀 ... I really like the compressor 😉 ... As for loudness of a signal I have lerned to never compress a signal to make it louder ever.... My buddies and I always say "leave it to a great compressor to ruin a good engineer!!!" I often recieve over compressed mixes that need some serious help cause the whole mix is "Squashed"... I am yet to compare the stereo images of the compressed and uncompressed signal but will get around to that shortly.... the compressor sounds good to the ear but we ALL know that the numbers speak clearer than the ear.... I'm not so sure if i like the new meters.... however all in all so far so good..... I will get back with some comparitave feedback of this compressors image alongside my trusty Ozone and our analog studio gear..... Nothing touches analog......Nothing..... I LOVE 😍 Analog....... some nice optical compression..... ooooohhhh boooyyyy!!!!
Just for the record if anybody cares:
>We should never use compression to make signals louder... Thats what the volume knob is for... 😉 ( or normalize, maximaize)
>Compression should be used on a track by track basis .. Rarely on a WHOLE mix....
>If used on a track by track basis properly, there should be no need for a MASTER compressor to "pull" the mix together.. It will already be together and just need some great brickwall limiting with a soft clip..
>The only time compression really should be used for a final mastering touch is if you do TV related music or internet audio... Compress the signal so the "little tiny speakers" everybody uses will match and support the frequency of your track or sound..
>I just would like to help and make sure we all can see the best results from our software and hardeware...
Much Love 😍 ... Peace and Science 🙂 😀 😉
stefan ·
I disagree on quite a lot about the usage of compression 🙂 I'm not a big fan of volume based dynamics, since it makes listening at a low volume (and very high, you might get some unexpected high volume thingies if there's too much volume based dynamics) difficult 🙂 It's all a matter of taste. I prefer spectral and arrangement based dynamics, but that's just me.
skOre ·
I think the original sense of "dynamics" is quite tied to volume. Yet its always good to please every using style.
stefan ·
And, ofcourse, compression isn't the only way to control the dynamics of your track 🙂
The Sound Scientist ·
stefan
PostPosted: Wed Nov 16, 2005 4:59 pm Post subject:
I disagree on quite a lot about the usage of compression Smile I'm not a big fan of volume based dynamics, since it makes listening at a low volume (and very high, you might get some unexpected high volume thingies if there's too much volume based dynamics) difficult Smile It's all a matter of taste. I prefer spectral and arrangement based dynamics, but that's just me.
Well of course... 😉 We all have our own niche I'm sure.... lol 😀 I'm not sure what part you disagree with though? Much Love...Peace and Science
stefan ·
Well. I think compressors are useful as a master bus effect 🙂 It's often a matter of music genre as well. I'm not for compressing the shit out of for instance classical music, because it will destroy the "expressiveness" of the tune and the played instruments. But I am for compressing "the shit" out of for instance some types of electronic music. It's part of the composition 🙂
The Sound Scientist ·
I see.. 😉 Sounds good! We don't really do a lot of electronic music in our studio, but i definately see what you are talking about and agree..... As for a master bus effect... side-chaining 1or 2 good comps. is awfully nice!!!!!! Peace and Science 😉 😀
stefan ·
fritzman wrote:
stefan wrote: What about a switch so that you can toggle the soft-saturation on and off?
Hm, then I still won't get to 0dB when overdriving the softsat only softly. 😉
Feel free to check the new version out, I posted about it on kvr:
http://www.kvraudio.com/forum/viewtopic ... 60#1503860
Does this saturation method suite you better? For me it works just as good, all I do is turn it fully on when I want it to behave like the old one 🙂. And are there still denormal issues in there?
lalo ·
an idea stef 🙂
sidechain ... very very hype 😉
lalo ·
some weird stuff with the VST version...
i'm testing on a really old machine (but's enough for a compressor 😉 )
PIII 933Mhz
with good ASIO drivers...
using eXT as a test host...
1. when automating a single parameter the cpu goes almost 100% 😮
2. when the sequencer is stopped the cpu is higher when it's running... like the compressor uses more resources when not processing than when processing (everything go "normal" if you change preset)
3. when changing preset the CPU goes at 100% for a while
lalo
skOre ·
is it the same when you don't display the gui?
lalo ·
the same when the gui is hidden....at least not a noticeable difference.....
it seems that SM automation has to be improved 😇
and maybe even the idle status....the plug consumes more when not processing audio 🙂
even on my powerful dothan 2.0
the above is on eXT
on tracktion 2 the automation seems a bit better...but still the idle problem...
when tracktion is stopped the cpu goes madly hi 😮 same as in eXT
and the same on podium...
automation cost seems a bit high (far better on podium and tracktion than eXT...the absolute best on podium 😍 )
the idle cost is around 40% on podium 😮
that said for a compressor...is pretty high on cpu...
there should be some optimisazion possible in SM as jorge told me...
my comparison is always Synth1 (try to automate all Synth1 parameters and you'll see 😉 )
huge testing today 🙂
fritzman ·
Hm,
the extra softclip knob is a step backwards for me. In the last version it was predictable as I knew that I can't have a level beyond 0dB and I had only to raise the host channel fader if I didn't go into lotsa overdrive.
Now it's easy to get huge levels with the output knob too high with little or no overdrive. The softclip isn't independant of the output level and it's knob so that I have to deal with three controls now: output, softclip and my channel volume fader.
I've got a few ideas about solving that prob, I'll try something later. Will post something if I have any interesting results.
BTW, I have no probs here with CPU usage, automation and denormals. CPU usage is a bit high actually, 10% on P4 @3.5. The first version used half(!) of the CPU of the new version. The anti-denormal code?
Best wishes, FRitz
stefan ·
The anti-denormal code can't be what's causing such a big cpu usage increase. I compared the two versions (the two released dll's) and had the same CPU usage on them both.
Please explain how you would want the softclip to work. Do you want the compressor to _always_ clip at 1.0, and the saturation knob would just change the shaping curve for the output?
lalo ·
fritzman wrote:
BTW, I have no probs here with CPU usage, automation and denormals. CPU usage is a bit high actually, 10% on P4 @3.5. The first version used half(!) of the CPU of the new version. The anti-denormal code?
wich hosts fritz?
fritzman ·
lalo wrote:
fritzman wrote:
BTW, I have no probs here with CPU usage, automation and denormals. CPU usage is a bit high actually, 10% on P4 @3.5. The first version used half(!) of the CPU of the new version. The anti-denormal code?
wich hosts fritz?
Nuendo 2.2. I was wrong about the CPU usage. I should shut down everything in the background when measuring CPU usage. 😊 The CPU usage is about the same. Now I couldn't measure a difference that is beyond the tolerance. Sorry for that.
fritzman ·
stefan wrote: The anti-denormal code can't be what's causing such a big cpu usage increase. I compared the two versions (the two released dll's) and had the same CPU usage on them both.
Please explain how you would want the softclip to work. Do you want the compressor to _always_ clip at 1.0, and the saturation knob would just change the shaping curve for the output?
Hm,
it could be cool to change the overdrive characteristic. That would add more versatility, cool.
I thought that the drive knob could act in the same way as the output knob in the first version, driving the output into the softclip. Then we could have a real output knob after the clip stage. Like in a guitar amp: preamp gain and master volume. Is that clear?
Best wishes, FRitz
stefan ·
fritzman wrote:
stefan wrote: The anti-denormal code can't be what's causing such a big cpu usage increase. I compared the two versions (the two released dll's) and had the same CPU usage on them both.
Please explain how you would want the softclip to work. Do you want the compressor to _always_ clip at 1.0, and the saturation knob would just change the shaping curve for the output?
Hm,
it could be cool to change the overdrive characteristic. That would add more versatility, cool.
I thought that the drive knob could act in the same way as the output knob in the first version, driving the output into the softclip. Then we could have a real output knob after the clip stage. Like in a guitar amp: preamp gain and master volume. Is that clear?
Best wishes, FRitz
Sure, but then you'd still get the scenario that you'd have to pull the volume up when using only low softclip drive to get a 0dB output, which basicly is like the first version but with a volume control added after the softclip.
fritzman ·
stefan wrote:
fritzman wrote: Hm,
it could be cool to change the overdrive characteristic. That would add more versatility, cool.
I thought that the drive knob could act in the same way as the output knob in the first version, driving the output into the softclip. Then we could have a real output knob after the clip stage. Like in a guitar amp: preamp gain and master volume. Is that clear?
Best wishes, FRitz
Sure, but then you'd still get the scenario that you'd have to pull the volume up when using only low softclip drive to get a 0dB output, which basicly is like the first version but with a volume control added after the softclip.
Oh, sure you're right. 😊 Was too late yesterday. I've got no cool idea right now. I'll think about it.
Best wishes, FRitz
sambean ·
Hey Stefan, I noticed a couple of comments about cpu consumption of the compressor over at kvr, if you don't mind I have a small suggestion..
I've found with the compressor most the cpu is being taken with the conversion to and from logarithmic scale.. I at first found this was a bit too much to include it as an insert effect on my Grainskan synth (the gate is constructed from the compressor..) so I switched the exp and log10 functions over to lookup tables instead using this:
Lookup table Generator
This totally reduced the cpu.. Hope this helps 🙂
Sam
stefan ·
I've already started using a look-up table for the dB->linear conversion. The log10 doesn't use nearly as much cpu as the pow(x,y) did. This change alone about halfed the cpu usage of the compressor 🙂
Thanks anyway though 🙂
tormodulator ·
Maybe you could call the knob Dry/Wet, because dry is to the left, no? Just for clarity?
I have a little problem with the module, but I'm going to investigate a bit more before posting it. Make sure I'm not just being dumb. 🙂
stefan ·
tormodulator wrote: Maybe you could call the knob Dry/Wet, because dry is to the left, no? Just for clarity?
I have a little problem with the module, but I'm going to investigate a bit more before posting it. Make sure I'm not just being dumb. 🙂
Sure, no worries!
As you might now ( http://www.kvraudio.com/forum/viewtopic ... highlight= ) I have made quite a lot of updates to the compressor (and EQ). The suitable changes will be transferred over to the toolbox versions as well.
tormodulator ·
stefan wrote: As you might now ( http://www.kvraudio.com/forum/viewtopic ... highlight= ) I have made quite a lot of updates to the compressor (and EQ). The suitable changes will be transferred over to the toolbox versions as well.
Yes - my problem is that while the compressor you posted at KvR works very much as expected (it's a very cool plug - I'm gonna get some real use out of this), the module when simply exported as a vst does not do what I expect. But like I said, I'll have to make some more a/b tests and make sure my assumptions are valid before I claim there's a bug or two in it. 🙂
Would it be possible to have both a minimal version *and* the more elaborate version in the toolbox? It would be too bad to have to strip the module of otherwise useful stuff if I'm not interested in all the features in a project...
stefan ·
tormodulator wrote:
stefan wrote: As you might now ( http://www.kvraudio.com/forum/viewtopic ... highlight= ) I have made quite a lot of updates to the compressor (and EQ). The suitable changes will be transferred over to the toolbox versions as well.
Yes - my problem is that while the compressor you posted at KvR works very much as expected (it's a very cool plug - I'm gonna get some real use out of this), the module when simply exported as a vst does not do what I expect. But like I said, I'll have to make some more a/b tests and make sure my assumptions are valid before I claim there's a bug or two in it. 🙂
Would it be possible to have both a minimal version *and* the more elaborate version in the toolbox? It would be too bad to have to strip the module of otherwise useful stuff if I'm not interested in all the features in a project...
There are some issues with the toolbox version when using very short attack or release times, maybe other issues as well. This things have been taken care of but are not yet merged into the toolbox.
Two versions is a very good idea. What should the simple version look like? Should we remove the different modes and just have downwards compression in it? Single band? No saturator? etc. Any help is appreciated.
Wrote something here? To have a post removed or your name hidden, write to [email protected].
SynthMaker