DYN4M1X

Nu Audio Science · 23 Jun 2009

Very cool How the hell did you get the load time that quick ?

NAS

aliasant · 23 Jun 2009

Nu Audio Science wrote: Very cool How the hell did you get the load time that quick ?

NAS

Lol Yes it only took about ten seconds to load it =)

But seriously. It sounds pretty good and has some interesting ideas. The double Knee. But its slightly hard to control.

I guess there is still a lot to do. I saw that bmanic responded to your post at kvr. Listen to that guy. He knows everything there is to know about compressors. And then some...

Working on my own compressors. Treying to get them to V.2 but it takes time. Im sooo slow. Btw. The Hold envelope. Which one did you end up using? I cant remember which one of the ones I kept actually worked.

Nu Audio Science · 23 Jun 2009

Load time here was instant Im getting really annoyed with my plugins taking 10-20 seconds to load

NAS

aliasant · 23 Jun 2009

Nu Audio Science wrote: Load time here was instant Im getting really annoyed with my plugins taking 10-20 seconds to load

NAS

That is odd. For me it loaded very very slowly. My own load pretty fast. Less then a sec at least.

Update: I tried again. Made a new empty Reaper 3.03 project and created 1 track. Loaded the plug and counted slowly to 10. 10 seconds?

Closed it and loaded my own Inerthia. Didnt take less then a sec. More like 1.5 -2 seconds. But still thats alot faster. And I think mine has more heavy graphics but maybe there are some strange arrays in the Dynamix?

platinumears · 23 Jun 2009

Interesting. It loads in about 1/2 second here.

I have an array of 2048 floats for the transfer LUT, plus another of only about 60 for the custom scale on the Range knob. Would that account for it?

I wrote my own envelope code:

streamin in; streamin attcoeff; streamin hldcoeff; streamin relcoeff; streamout out; float level,up,down,holdcount; stage(0) { level=1; } stage(1) { holdcount=(level>in)&(holdcount+1);

up =level*attcoeff;

down=(holdcount<hldcoeff)&level +(holdcount>=hldcoeff)&(level*relcoeff);

level=(in<level)&(max(in,down)) +(in>=level)&(min(in,up));

out=level;

<edit> got the following bit wrong:

Att coeff = dB/sample (expressed as linear scaler value, ie: slightly larger than 1) Hold coeff = no. of samples Rel coeff = -dB/sample (expressed as linear scaler value, ie: slightly smaller than 1)

I converted this to ASM in the plug so that I could change the float holdcount to an int, but I'm still not sure what I'm doing with ASM so any other optimisation suggestions would be welcome!

aliasant · 23 Jun 2009

platinumears wrote: Interesting. It loads in about 1/2 second here.

I have an array of 2048 floats for the transfer LUT, plus another of only about 60 for the custom scale on the Range knob. Would that account for it?

I have no idea why it loads so slow. Im running Reaper 3.03 on Windows 7 It cant be SM since all my other plugs loads quick. There are other problems with VST in the latest beta but this isnt one of them I think.

Strange.

About asm. I would love to do asm on this. i think its a pretty fast envelope already but I havnt figured out the knob ranges. When you write (db/sample) Do you db / samplerate?) And dB. what dB is that? I hooked it up to an oscilloscope and ran an osc thru it but I could get anything out of it at all. Flat lined on me 😉 What would the attack knob range look like?

platinumears · 23 Jun 2009

The envelope smoothing comes after the dynamics transfer LUT (so is operating on the gain modulation signal rather than the audio input) and works on the basis that the attack and release stages have a constant rate of gain change. That's why the attack and release sliders are calibrated in dB/ms and dB/s respectively.

So to get the attack coefficient I divide the required rate of gain change in dB/s by the samplerate to calculate the gain change per sample, then convert that dB value to a linear scaler. The attack coefficient should end up as a value slightly larger than 1. The release coeff is the same, except the dB value should be negative so the coeffiecient ends up as a value slightly smaller than 1. Hope that makes sense..?

(I just realized I posted total rubbish above! Will edit it...)

infuzion · 22 Jul 2009

aliasant wrote: That is odd. For me it loaded very very slowly. My own load pretty fast. Less then a sec at least. Update: I tried again. Made a new empty Reaper 3.03 project and created 1 track. Loaded the plug and counted slowly to 10.10 seconds? Closed it and loaded my own Inerthia. Didnt take less then a sec. More like 1.5 -2 seconds. But still thats alot faster. And I think mine has more heavy graphics but maybe there are some strange arrays in the Dynamix?

Maybe the SM.DLL was still in the memory after the first VST. Try loading your VST directly after a fresh reboot.

platinumears wrote: I converted this to ASM in the plug so that I could change the float holdcount to an int, but I'm still not sure what I'm doing with ASM so any other optimisation suggestions would be welcome!

I can help out with ASM, but I don't think it matters for float vs int for the counter?

platinumears · 25 Jul 2009

Hi infuzion.

Any optimisation help would be great! ASM is a bit of a mystery to me still... I've also been getting some apparent denormal problems when using this in a mix, despite the denormal remover module in the sidechain, but I've not had time to investigate that yet.

I guess I might as well upload the osm file. See attached.

infuzion · 26 Jul 2009

platinumears wrote: Any optimisation help would be great! ASM is a bit of a mystery to me still... I've also been getting some apparent denormal problems when using this in a mix, despite the denormal remover module in the sidechain, but I've not had time to investigate that yet.

Check your email; Santa came in July this year!

platinumears · 26 Jul 2009

nothing in my inbox...? 😕

infuzion · 26 Jul 2009

platinumears wrote: nothing in my inbox...?

[email removed] is not you?

platinumears · 27 Jul 2009

infuzion wrote: platinumears wrote: nothing in my inbox...?

[email removed] is not you?

Yes thats me, but I haven't recieved anything. Try "dan" instead of "info"? Or [email removed]

infuzion · 27 Jul 2009

platinumears wrote: Try "dan" instead of "info"? Or [email removed]

OK, turn off your gmail & FWD filters please. 🙂