KVR Audio

cymatics · Post by **cymatics** » Tue Jan 30, 2007 2:19 am

Stefan:

First, let me say that SM_Comp011a is the most exciting software comp I've used in LOOOOONG time. I am really addicted to how radically I am able to alter the sonic spectrum of whatever material is at hand. I am a hardware guy at heart who went to the digital domain slowly and grudgingly. I have found very few plugs that truly have 'vibe' or 'mojo' to them & I would place SM_Comp011a in my top 10.

Having said all of that, I have a weird CPU problem with SM_Comp011a.

I have two separate machines:

1) A 3.00 GHz P4 with 504 MB of RAM running XP Pro SP2
This machine clocks 3.6% CPU per instance of SM_Comp011a

2) A 3.4 GHz P4 with 1GB of RAM running XP Pro SP2
This machine clocks 19.4% CPU per instance of SM_Comp011a. It starts at about 1.5% then ramps up to 19.4% over the course of about 10 seconds, then stabilizes.

This was tested with both machines running REAPER 1.71.

Any idea what might be causing the faster machine (theoretically speaking) to suffer the CPU munch?

If I can provide any additional info to help debug please let me know.

Thanks for this outstanding freebie. It really puts a smile on my face.

- jon

stefancrs · Post by **stefancrs** » Tue Jan 30, 2007 7:32 am

Ouch.

Is that CPU readout from when it is being fed silence or when it actually is working on audio?

Thanks for the great feedback! And did you try the EQ too?

trickykid · Post by **trickykid** » Tue Jan 30, 2007 8:40 am

What is the latest versions of these plugins? Is the first post on this thread updated? I'm all confused.

stefancrs · Post by **stefancrs** » Tue Jan 30, 2007 9:08 am

trickykid wrote:What is the latest versions of these plugins? Is the first post on this thread updated? I'm all confused.

Yeah, I always put the latest released versions in the first post.

trickykid · Post by **trickykid** » Tue Jan 30, 2007 9:19 am

Ok. Thanks!

cymatics · Post by **cymatics** » Tue Jan 30, 2007 1:50 pm

stefancrs wrote:Is that CPU readout from when it is being fed silence or when it actually is working on audio?

When playback stops, the CPU returns to zero.

After my post above, I tried SM_Comp011a again on the machine with the CPU problem; this time running in Samplitude 9. I only got 5% CPU! So, I'm guessing it is a conflict between REAPER, SM_Comp011a & my PC (which happens to be my primary mix tool

). I'm really leaning toward REAPER as my mix platform, so I'm hoping this can be resolved somehow. If there is any other info about my PC that may helpful, please let me know

stefancrs wrote:Thanks for the great feedback! And did you try the EQ too?

Haven't had as much time to put it through it's paces yet, but the 'saturate' feature sounds unique and very promising.

stefancrs · Post by **stefancrs** » Tue Jan 30, 2007 2:05 pm

cymatics wrote:
stefancrs wrote:Is that CPU readout from when it is being fed silence or when it actually is working on audio?
When playback stops, the CPU returns to zero.

After my post above, I tried SM_Comp011a again on the machine with the CPU problem; this time running in Samplitude 9. I only got 5% CPU! So, I'm guessing it is a conflict between REAPER, SM_Comp011a & my PC (which happens to be my primary mix tool ). I'm really leaning toward REAPER as my mix platform, so I'm hoping this can be resolved somehow. If there is any other info about my PC that may helpful, please let me know

So the 19.4% readout was when playing back audio through the compressor and also the transport being in "play"? Sorry if I'm a bit of a nag, I just have to be sure about this.

cymatics wrote:
stefancrs wrote:Thanks for the great feedback! And did you try the EQ too?
Haven't had as much time to put it through it's paces yet, but the 'saturate' feature sounds unique and very promising.

The feedback might be even more fun actually

The saturation I mainly use for slight coloring or when I want to use the entire thing as an distortion unit (which I think works really well).

cymatics · Post by **cymatics** » Tue Jan 30, 2007 2:16 pm

stefancrs wrote: So the 19.4% readout was when playing back audio through the compressor and also the transport being in "play"? Sorry if I'm a bit of a nag, I just have to be sure about this.

No problem at all!

I don't mean this to sound sarcastic, but as far as I know, the only way for audio to pass through the comp is when the transport is in "play".

When playing back audio (transport is in "play"), the CPU ramps up to 19.4%. When I stop playback (transport is in "stop") the CPU promptly returns to 0%.

Please let me know if I misunderstood your question?

- jon

stefancrs · Post by **stefancrs** » Tue Jan 30, 2007 2:23 pm

cymatics wrote:
stefancrs wrote: So the 19.4% readout was when playing back audio through the compressor and also the transport being in "play"? Sorry if I'm a bit of a nag, I just have to be sure about this.
No problem at all!

I don't mean this to sound sarcastic, but as far as I know, the only way for audio to pass through the comp is when the transport is in "play".

When playing back audio (transport is in "play"), the CPU ramps up to 19.4%. When I stop playback (transport is in "stop") the CPU promptly returns to 0%.

Please let me know if I misunderstood your question?

- jon

You could be monitoring live audio fed through the compressor and one can also have the compressor on a non-playing (silent) channel although the transport is in "play"

But I take this as you got the high CPU when actually feeding audio to the compressor and having the transport in play, and when not feeding audio to it and having transport in stop, the CPU went back down.

Do you get this weird high CPU behaviour no matter what settings you have on the compressor or is it setting dependent?

Thanks for helping me out here!

cymatics · Post by **cymatics** » Tue Jan 30, 2007 3:18 pm

Ah! I see. I'm thinking of things a little too literally

To answer your other question; yes, the CPU spike occurs regardless of compressor settings.

- jon

stefancrs · Post by **stefancrs** » Sun Feb 11, 2007 6:21 pm

Time for an update of the compressor. There's been a whole bunch of internal versions between 011 and this one, so I might have forgotten about any changes, bug fixes and whatnot that I've made. If so, I am sorry about that.

DLL Here

Main addition is the ability to shape the clipper from soft(ish) too hard. Click the green little curve there and drag the mouse.

Apart from that CPU usage has been decreased, at least on my machine, but I'd like to know how it has changed on your machines as well. Keep in mind that CPU usage is at its lowest when Ahead is set to its minimum and the clipper is at one of its boundaries (that is, fully hard or fully smooth(ish)).

And there were requests for making the GUI just slightly bigger, so I did that too. But to compensate and negate the effect of the request, I added that clipper shape thingy, sorry

BTW, watch your ears when changing Compress/Expand and Down/Up modes. If you're using a high Ratio setting you might end up with very loud settings.

Let us know of any issues!

Cheers!

DuX · Post by **DuX** » Sun Feb 11, 2007 6:57 pm

Wow, cool

, thanx Stefan!

stefancrs · Post by **stefancrs** » Mon Feb 12, 2007 8:06 am

*bump*
Anyone tried this out yet? What do you think about the control over the clipper? Good or bad?

And oh, I forgot to say, I have also changed how the Ratio is applied when doing expansion. Keep in mind however that if not using Auto Gain upwards expansion can get _quite_ dangerous. For instance, if set to 1:inf, and the signal is 1dB over the threshold, the signal will basically be expanded to inf dB over the threshold (or, well, not, but a _lot_ of dB's).

liars&ashes · Post by **liars&ashes** » Mon Feb 12, 2007 8:47 am

----I tried it out this afternoon. Pretty darn good for some uses, I don't know about for typical compression needs, or high quality requiring situations etc.
----For an SM plug, it works as well as any I've tried yet, and it's a pretty stompin' compressor to boot. The slanted line graph control, didn't seem to really do much, then again, I don't even know what it is, I was just adjusting it and listening for a change that never happened. The *off*/*sustain*/*peak* control really only seemed obvious at off or peak to me, and about the same for the stereo link, all the way right or left, but not much difference in the middle. It ran stable in FL, and didn't seem to use too much cpu.
----Is this one still in development then, or this is version like 1.3 or so or what ? Seems like a plug with really good potential actually.

Jeff

stefancrs · Post by **stefancrs** » Mon Feb 12, 2007 9:00 am

liars&ashes wrote:----I tried it out this afternoon. Pretty darn good for some uses, I don't know about for typical compression needs, or high quality requiring situations etc.

I find it good for both "ordinary" tasks and more eccentric ones as well, but it is of course kind of modeled after my needs, and from the feedback I get in this thread.

liars&ashes wrote: ----For an SM plug, it works as well as any I've tried yet, and it's a pretty stompin' compressor to boot. The slanted line graph control, didn't seem to really do much, then again, I don't even know what it is, I was just adjusting it and listening for a change that never happened.

Really? I find the audible difference quite apparent. But mainly when there's a small-ish amount of saturation going on, if there's lots of it both clipper shapes do sound quite the same. The "hard" clip gives a higher volume and seems to preserve some transients better, otoh it's not as harmonic as the smooth shape.

liars&ashes wrote:The *off*/*sustain*/*peak* control really only seemed obvious at off or peak to me, and about the same for the stereo link, all the way right or left, but not much difference in the middle. It ran stable in FL, and didn't seem to use too much cpu.

When doing limiting tasks you probably want the Auto Gain set at "Sustain" and let the peaks be caught by the clipper. What do you mean by "not much difference in the middle" for the stereo link? On material with lots of panned out sounds, there's a lot of difference in how the compression is applied with stereo link compared to without, and sometimes you want to "squash" the material a bit more than what you get to do with full stereo linking, but still want to preserve some, or a lot of, the stereo imaging. But these kind of adjustments only makes sense when there's quite a bit of dynamic difference between the left and the right channel.

liars&ashes wrote:----Is this one still in development then, or this is version like 1.3 or so or what ? Seems like a plug with really good potential actually.

Indeed it is. The EQ too. They are both getting nearer and nearer a stable feature set however, which is a good thing

BTW, thanks for your feedback!

Free SynthMaker Compressor+EQ/Multi-stage distortion unit