cbagwell / sox Goto Github PK

Is it possible to influence the reading operation, so that the volume correction occurs exactly when reading, and not after?

I noticed there hasn't been a release since 2015...

hi,
I trying suppress background noise in my waves (call record). at
I find Sox very helpful for batch pre-processing.
My issue is with the noisered amount argument, after I generate a noise profile from the original wav (a trim of silence ), the noisered give an output with weird echo and the speech quality change within the output.
is there an specific way for choosing the right amount argument?
can it be generalize for batch pre-processing?

thanks

Hey Mansr, the compiled sox in MacOS doesn't show the version after typing ./sox --version

./sox --version

SoX v

hello,Very rare. You opened issue under the project！！！
Well, I want to compile a static library for Mac OS and a static library for windows for developing sound effects, but I can't compile the code I downloaded from SourceForge successfully. In my Mac OS M1; So I hope to get your help. I wonder if you can provide some compilation tutorials or projects?

While sox_delete_effect is used internally and works correctly in that context, it is not usable from user code:

// This will double-free e->priv when chain is deleted
sox_effect_t * e = sox_create_effect(sox_find_effect("input"));
sox_add_effect(chain, e, in, out);
sox_delete_effect(e);

// This is the way to do it according to the examples.
// However it is not possible from user code when libsox is dynamically linked (at least on Windows)
sox_effect_t * e = sox_create_effect(sox_find_effect("input"));
sox_add_effect(chain, e, in, out);
free(e);  // memory could have been allocated on another heap

// And by the way, this (admittedly not very useful example)
// will leak memory since e->priv is never freed
sox_effect_t * e = sox_create_effect(sox_find_effect("input"));
sox_delete_effect(e);

sox_delete_effect should probably be fixed so that it works as a proper counterpart to sox_create_effect. As a plan B, it could be removed from the public API to avoid confusion (but that way there would still be no way to use the effects chain from user code without leaking memory).