totally lost with audio !

Unread post by **oli_lab** » 26 Apr 2015, 18:15

the title says it all.

I'm trying to get my hands dirty with audio modules...

for a start, I wanted to start easy : an audio volume module with lots of gain followed by a soft clip algorithm...

but even if the module compiles OK, it sounds as if audio is going at modulation rate instead of audio rate.

I'm sure it is not that complicated to sort it out as I must have miss something, but what ???

looks like it is the "TPrecision" that is the problem

here the interesting bits

supabooster.h

private:
//-------------------------------------------------------------------------
// parameters events
UsineEventPtr audioInputs[AUDIO_INS_OUTS_MAX]; // audio input
UsineEventPtr audioOutputs[AUDIO_INS_OUTS_MAX]; // audio output
UsineEventPtr fdrGain;
UsineEventPtr switchMute;
TPrecision tempValue[AUDIO_INS_OUTS_MAX];

//-------------------------------------------------------------------------
static const int numOfParamAfterAudiotInOut = 2;

int queryIndex;
int numOfAudiotInsOuts;

////////////////////////////////////////////////

supabooster.cpp

void AudioVolumeExample::onProcess ()
{
for (int i = 0; i < numOfAudiotInsOuts; i++)
{
sdkCopyEvt (audioInputs, audioOutputs);
sdkMultEvt1 (coeffGain, audioOutputs);

//y = x / (1 + | x | ) softclipping formula
tempValue = sdkGetEvtData(audioOutputs) / (1 + abs(sdkGetEvtData(audioOutputs)));
sdkSetEvtData(audioOutputs, tempValue);

////////////////////////////////////////////////////////

also, is the audio values are between -1 and 1 ?

thanks for your help

Olivar

23fx23 · Unread post by **23fx23** » 26 Apr 2015, 19:10

yes audio values are going from -1 to 1 not yet fully understood audio, can't really help for the rest sry. is your module well defined as a process module "Don't Process = FALSE"

Unread post by **oli_lab** » 26 Apr 2015, 21:22

Thank you,
but "Don't Process = FALSE" by default so its not that...

can't wait it's working : I prepared a supaBooster with softclipping, another with hard insane clipping and a DC blocker, but still this problem with TPrecision...

martignasse · Unread post by **martignasse** » 27 Apr 2015, 11:07

hey oli_lab,

from the code you posted, you appli the soft clipping formula only to the first sample of each output sample block.

//y = x / (1 + | x | ) softclipping formula
tempValue = sdkGetEvtData(audioOutputs) / (1 + abs(sdkGetEvtData(audioOutputs)));
sdkSetEvtData(audioOutputs, tempValue);

remember that an audio parameter contain the whole block of sample corresponding to the audio tick for this process call, so it's literally an array of size = usineblock.

the simplest implementation is to add a for loop to apply your softclipping algo to each samples

something like this (not tested)

Code: Select all

void AudioVolumeExample&#58;&#58;onProcess &#40;&#41;
&#123;
    tempValue
    for &#40;int i = 0; i < numOfAudiotInsOuts; i++&#41;
    &#123;
        sdkCopyEvt &#40;audioInputs&#91;i&#93;, audioOutputs&#91;i&#93;&#41;;
        sdkMultEvt1 &#40;coeffGain, audioOutputs&#91;i&#93;&#41;;
        
        //y = x / &#40;1 + | x | &#41;  softclipping formula
        for &#40;int j = 0; j < sdkGetEvtSize&#40;audioOutputs&#91;i&#93;&#41;; j++&#41;
        &#123;
            tempValue&#91;i&#93; = sdkGetEvtArrayData&#40;audioOutputs&#91;i&#93;, j&#41; / &#40;1 + abs&#40;sdkGetEvtArrayData&#40;audioOutputs&#91;i&#93;, j&#41;&#41;&#41;;
            sdkSetEvtArrayData&#40;audioOutputs&#91;i&#93;, j, tempValue&#91;i&#93;&#41;;
        &#125;
    &#125;
&#125;

Unread post by **oli_lab** » 27 Apr 2015, 12:56

Thanx !
I'm not lost anymore.

I'll try that tonite !

Olivar

Unread post by **oli_lab** » 28 Apr 2015, 15:39

this "tempValue" could be as well a "tempValue" if I understand right

23fx23 · Unread post by **23fx23** » 29 Apr 2015, 00:38

mm think nope because it indexes a different audio input in case you got several channels, that would be same only if manipulating a mono channel,
but if got 4 channels with differents audio content, have to check volumes and process so that they got their own processing, thus the array stuff if i well understood.

Unread post by **oli_lab** » 29 Apr 2015, 16:03

if there is many channel, each channel is process by block, so, as I understand it, the "tempValue" will be used by each channel in turn, don't need to be memorise from one block to another of the same channel.

ceasless · Unread post by **ceasless** » 29 Apr 2015, 20:08

I think oli_lab is right. Since temp_value is only assigned and accessed within the innermost loop, you should be able to get away with a single placeholder variable. A small enhancement to the memory efficiency of your module.

Unless there is something I'm missing (and I've gotten used to that

).

The rest of the for-loop construct looks solid.

Unread post by **oli_lab** » 30 Apr 2015, 00:13

the module is working fine and is on the addons page !

Clearscreen · Unread post by **Clearscreen** » 30 Apr 2015, 03:26

Thanks for these.

ceasless · Unread post by **ceasless** » 01 May 2015, 16:32

Hey oli_lab,

were these also in the file you uploaded for the Grove?

This weekend I am going to be compiling OS X versions of all the modules

Unread post by **oli_lab** » 01 May 2015, 19:13

no not yet !

I am totally lost (again) with this github

can't sort out how to upload files.

here they are :
http://we.tl/TE4MLuahoY

Welcome to %s forums

totally lost with audio !

Login • Register