Encoding Formats

Supported Formats

The DSP Series 2 card supports the following file formats:

Raw

• G.711 A-law

• G.711 µ-law

• G.726 32 Kbps ADPCM

• 32 Kbps OKI ADPCM (Dialogic)

• 24 Kbps OKI ADPCM (Dialogic)

.wav

• G.711 A-law

• 16-bit linear, 11025 Khz mono

• 8-bit linear, 11025 Khz mono

* Note that 16 bit uncompressed data requires twice the processor and DSP bandwidth, and will adversly affect channel densities accordingly.

ADPCM

Excel supports Adaptive Differential Pulse Code Modulation (ADPCM) for Playback files. ADPCM is a compression algorithm for voice that records only the difference between samples, and dynamically adjusts the coding scale to accommodate large and small differences.

Because OKI requires fewer resources than G.726, OKI is more suitable for resource-intensive applications, such as speeding up or slowing down voice files. And OKI provides voice quality comparable to G.726.

When you choose either OKI or G.726, the caches on both the DSP and the 8260 chip store the data in that encoding format only. Both the Skip Ahead / Skip Back feature and the Speed Up / Slow Down feature take this file format choice into consideration. The Play File Offset and Length feature does not take this file format choice into consideration. Instead, the host application is responsible for considering that choice.

Compression Benefits

Compared to standard PCM, ADPCM effectively doubles your cache capacity. For two seconds of voice with standard PCM, 16 KB are required. For two seconds of voice with ADPCM, only 8KB are required. Because only half the data is read or written compared to standard PCM, ADPCM also increases performance over the Network File System (NFS) server link.

Compression is performed as the voice is initially recorded, and decompression is performed just before the file is played out. ADPCM does not degrade the performance of the playback or record functionality.