I am working on audio localization, and I can't figure out this format, now it's near to listenable. Maybe you can help me with this.

 

Links:

JOURNO.RIB

GASPOUR.RIB

 

Closest .txth I got.

Quote

codec = IMA
channels = 2
interleave = 0x10000
sample_rate =44100
num_samples = data_size

 

The file is a stream of samples encoded by a variation of the ADPCM IMA algorithm (in FFMPEG this algorithm is identified as ADPCM_IMA_QT, but the implementation differs in detail).

Each of the channels is in turn encoded in an interleave of 0x10000 bytes in size into frames of 0x400 bytes in size. Thus, in one interleave there are up to 64 channel frames, encoded by a variation of the ADPCM_IMA_QT algorithm. The key difference from the original implementation is the storage of data for encoder initialization: while ADPCM_IMA_QT uses packed storage of predicor and step_index in a 16-bit value, in the modified algorithm predictor is stored in the first and second bytes of the frame as 16-bit number (little-endian), and step_index is in the third byte as an 8-bit number. The structure is aligned to four bytes, leaving the fourth byte unused and equal to zero.

If there is fewer data than the frame size at the end of the file, the remaining frame and interleave space remains filled with zeros.

via winterheart ManhuntRIBber

Thanks, we already contacted with winterheart and solved this.