1CELT is a very low delay audio codec designed for high-quality communications. 2 3Traditional full-bandwidth codecs such as Vorbis and AAC can offer high 4quality but they require codec delays of hundreds of milliseconds, which 5makes them unsuitable for real-time interactive applications like tele- 6conferencing. Speech targeted codecs, such as Speex or G.722, have lower 720-40ms delays but their speech focus and limited sampling rates 8restricts their quality, especially for music. 9 10Additionally, the other mandatory components of a full network audio system— 11audio interfaces, routers, jitter buffers— each add their own delay. For lower 12speed networks the time it takes to serialize a packet onto the network cable 13takes considerable time, and over the long distances the speed of light 14imposes a significant delay. 15 16In teleconferencing— it is important to keep delay low so that the participants 17can communicate fluidly without talking on top of each other and so that their 18own voices don't return after a round trip as an annoying echo. 19 20For network music performance— research has show that the total one way delay 21must be kept under 25ms to avoid degrading the musicians performance. 22 23Since many of the sources of delay in a complete system are outside of the 24user's control (such as the speed of light) it is often only possible to 25reduce the total delay by reducing the codec delay. 26 27Low delay has traditionally been considered a challenging area in audio codec 28design, because as a codec is forced to work on the smaller chunks of audio 29required for low delay it has access to less redundancy and less perceptual 30information which it can use to reduce the size of the transmitted audio. 31 32CELT is designed to bridge the gap between "music" and "speech" codecs, 33permitting new very high quality teleconferencing applications, and to go 34further, permitting latencies much lower than speech codecs normally provide 35to enable applications such as remote musical collaboration even over long 36distances. 37 38In keeping with the Xiph.Org mission— CELT is also designed to accomplish 39this without copyright or patent encumbrance. Only by keeping the formats 40that drive our Internet communication free and unencumbered can we maximize 41innovation, collaboration, and interoperability. Fortunately, CELT is ahead 42of the adoption curve in its target application space, so there should be 43no reason for someone who needs what CELT provides to go with a proprietary 44codec. 45 46CELT has been tested on x86, x86_64, ARM, and the TI C55x DSPs, and should 47be portable to any platform with a working C compiler and on the order of 48100 MIPS of processing power. 49 50The code is still in early stage, so it may be broken from time to time, and 51the bit-stream is not frozen yet, so it is different from one version to 52another. Oh, and don't complain if it sets your house on fire. 53 54Complaints and accolades can be directed to the CELT mailing list: 55http://lists.xiph.org/mailman/listinfo/celt-dev/ 56 57To compile: 58% ./configure 59% make 60 61For platforms without fast floating point support (such as ARM) use the 62--enable-fixed argument to configure to build a fixed-point version of CELT. 63 64There are Ogg-based encode/decode tools in tools/. These are quite similar to 65the speexenc/speexdec tools. Use the --help option for details. 66 67There is also a basic tool for testing the encoder and decoder called 68"testcelt" located in libcelt/: 69 70% testcelt <rate> <channels> <frame size> <bytes per packet> input.sw output.sw 71 72where input.sw is a 16-bit (machine endian) audio file sampled at 32000 Hz to 7396000 Hz. The output file is already decompressed. 74 75For example, for a 44.1 kHz mono stream at ~64kbit/sec and with 256 sample 76frames: 77 78% testcelt 44100 1 256 46 intput.sw output.sw 79 80Since 44100/256*46*8 = 63393.74 bits/sec. 81 82All even frame sizes from 64 to 512 are currently supported, although 83power-of-two sizes are recommended and most CELT development is done 84using a size of 256. The delay imposed by CELT is 1.25x - 1.5x the 85frame duration depending on the frame size and some details of CELT's 86internal operation. For 256 sample frames the delay is 1.5x or 384 87samples, so the total codec delay in the above example is 8.70ms 88(1000/(44100/384)). 89