Re: [SoX-users] Problem with using LADSPA plugin

[Eric W. <nor...@yh...>]

Paul Arnold <pa...@pi...> wrote:
> Hi Eric. The command I used for Ecasound that got it going was 
> 
> ecasound -i vocoder_test_short_formant_left.wav -o result.wav -el:vocoder,16,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1 -erc 1,2
> 
> So that's inputting a file with formant on left (speech) and carrier on right (synth). You get mono vocoded output.

Hi Paul, thank you for that command-line, it was useful for
understanding and testing my SoX results vs ecasound.  Apologies for
taking so long to get back to you.

I've made two more patches on top of my "ladspa-multiport" branch
which should improve SoX LADSPA support for plugins like vocoder
which remap channels.

> Regarding what I tried with Sox, my notes are slightly confused. I had thought I got a file which was blank, but all I can get just now is 
> 
> # sox vocoder_test_short_formant_left.wav result.wav ladspa vocoder vocoder 16,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1
> sox FAIL ladspa: not enough arguments for control ports

LADSPA parameters for sox are separate shell parameters, so you should
separate parameters with " " (space) instead of ",".

Extracted from my commit message, I've managed to get the vocoder
LADSPA plugin working with my patches:

    Using the formant.wav and carrier.wav samples from
    http://www.sirlab.de/linux/descr_vocoder.html the following
    command-line is successful:
    
    sox -M formant.wav carrier.wav \
      -b 16 result.wav rate 44100 \
      ladspa vocoder 16 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
    
    Using 16-bit (or higher) output reduces the noise-floor hiss
    greatly since LADSPA processing is done as 32-bit float.

More info in the SF tracker:

https://sourceforge.net/tracker/index.php?func=detail&aid=3534109&group_id=10706&atid=310706

(TL;DR: git pull git://bogomips.org/sox.git ladspa-multiport)