The DCT codec

The DCT transform has been widely used for in standard codecs such has the MPEG-2,H26x and JPEG standards. I developed two codecs Codec 1, maximising the compression ratio, whereas the output of Codec 2 is more robust. For both codecs, the bitrate is programmable and remains CONSTANT, in contrast to the above mentioned standard codecs, where intra frames exceed ten times the inter frame size and therefore delays of seconds are possible. This is the reason why the codecs need to run for about 10 to 20 frames to establish the quality. Example sequences for Codec one at 6.7, 8, 9.6 and 13 kbps and for Codec 2 at 11.36 kbs is shown here:

The Claire Sequence

  1. The Claire Sequence at 6.7 kbps , DCT Codec 1,
  2. The Claire Sequence at 8 kbps, DCT Codec 1,
  3. The Claire Sequence at 9.6 kbps, DCT Codec 1,
  4. The Claire Sequence at 13 kbps, DCT Codec 1,
  5. The Claire Sequence at 11.4 kbps, DCT Codec 2.
The codecs are proposed as a direct replacement of voice channels in second generation mobile systems such as the Pan-European GSM, the American IS-54 and IS-95 as well as Japanese systems.

The following two sequences are Coded with Codec 1:

The Miss America Sequence

  1. The Missa Sequence , at 125 byte per frame (compression ratio 202:1)
  2. The Missa Sequence , at 100 byte per frame (compression ratio 253:1)
  3. The Missa Sequence , at 62.5 byte per frame (compression ratio 406:1)

The 'Lab' Sequence

  1. The Lab Sequence , at 125 byte per frame (compression ratio 202:1)
  2. The Lab Sequence , at 100 byte per frame (compression ratio 253:1)
  3. The Lab Sequence , at 62.5 byte per frame (compression ratio 406:1)