[gst-devel] Appsink output with audio/x-raw-float caps

[Alexey Chernov]

Hello,

I've got two more questions concerning appsink during futher development of my 
sound editor project:

1. Is it possible that appsink produces NaNs in its output with "audio/x-raw-
float,width=32,endianness=1234" caps? Or maybe I illegally convert to float from 
the output buffer of bytes (I just read the stored buffer as vector of floats 
with length divided by 4).. There're so many NaNs in the output flow after that 
and the question is whether it is normal for appsink output or it's my wrong 
convertion.

2. Could you please help me with any links to audio/x-raw-float format 
reference? Especially the maths of storing sample in floating point value and 
the way these values are normalized (I need to visualize the wave so it's 
important for me to know). I did my best to find it in the web but almost no 
results..

Thank you very much in advance.