Manual.txt for Version 2.61 of ISO/MPEG Audio Layer 3 software only encoder/decoder for Unix. 1. ENCODER V2.61 ============= l3enc is an ISO/MPEG Layer-3 software only encoder. It takes audio data files as input and delivers Layer-3 coded bitstream files as output. Several options can be selected via command line switches. Usage: l3enc [-switch1 [-switch2 [...]]] PLEASE NOTE: ------------ o For non-registered users, ancillary data processing is not supported. o Non-registered users may use the encoder only with the following options (input must be 44.1 kHz!): 112 kbit/s stereo @ 44.1 kHz 56 kbit/s stereo @ 22.05 kHz 16 kbit/s mono @ 11.025 kHz o Registered users may use the encoder additionally with the following options: 8 kbit/s mono @ 8 kHz 16 kbit/s mono @ 11.025, 16 kHz 24 kbit/s mono @ 16, 22.05, 24 kHz. 32 kbit/s mono @ 16, 22.05, 24 kHz. 56 kbit/s stereo @ 16, 22.05, 24 kHz. 64 kbit/s stereo @ 16, 22.05, 24 kHz. 56 kbit/s mono @ 32, 44.1, 48 kHz 64 kbit/s mono @ 32, 44.1, 48 kHz 96 kbit/s stereo @ 32, 44.1, 48 kHz 112 kbit/s stereo @ 32, 44.1, 48 kHz 128 kbit/s stereo @ 32, 44.1, 48 kHz 256 kbit/s stereo @ 32, 44.1, 48 kHz If the input has a sampling frequency of x2, x3, x4 or x6, it is downsampled on the fly. If you need other bitrates, please contact layer3@iis.fhg.de. 1.1 : audio input file The first command line argument specifies the name for the PCM audio data file. Version 2.61 of the encoder accepts either raw PCM audio data files, PCM audio data files in RIFF/WAVE format as used by Microsoft Windows, PCM audio data files in the sun .au or PCM audio data files in the Apple AIFF format. The samples must be 16 bit signed integer values. for raw PCM audio data: By default the input file is assumed to contain raw PCM audio data. Stereo audio data is input in interleaved format, the first channel beeing the left channel. ... Mono audio data has the format .... Whether the input file is treated as mono or stereo audio data is set by the downmix switch (1.4). Default is stereo. Please see for the -sr, -tfc and -tfs switches below. PLEASE NOTE: Non-registered users may use the encoder only with .snd/.wav/.aiff files. 1.2 : Layer 3 output file The second command line argument specifies the name for the bitstream output file. The extension of the file name should be .mp3. The format of the bitstream is as defined in the ISO/MPEG publications IS11172-3 (MPEG-1) and IS13818-3 (MPEG-2). For very low bitrates a special Fraunhofer format called "MPEG 2.5" is used. 1.3 bitrate The bitrate of the bitstream output is selected via the '-br' switch. The bitrate is specified in bits/second. The bitrate is the total bitrate for all encoded channels, i.e. if you select 'br 112000' and 'stereo', both channels will be stuffed into one bitstream of 112000 bits/second. Valid bitrates are: o 8000 bit/s o 16000 bit/s o 24000 bit/s o 32000 bit/s o 56000 bit/s o 64000 bit/s o 96000 bit/s o 112000 bit/s o 128000 bit/s o 256000 bit/s The default bitrate is 112000 bit/s. 1.4 downmix If a stereo input file should be treated as mono, the '-dm' swich can be used. The mono signal is calculated by (l+r)/2. 1.5 high quality If the '-hq' option is specified, the encoder will try to produce higher audio quality, but at the cost of a reduced encoding speed. 1.6 crc check If '-crc' is asserted, ISO/MPEG crc checking is enabled. Without the 'crc' switch, crc checking is disabled. 1.7 ancillary data If the '-anc ' option is specified, the named file is is inserted as ancillary data in the bitstream. The rate is in bits/frame. 1.8 sampling rate If a raw PCM file is used as input, the '-sr' switch supports the encoder with the sampling rate. THIS IS NOT NEEDED FOR .wav/.snd/.aiff INPUT! 1.9 swap input samples If a raw PCM file is used as input, the '-tfs' switch swaps each 16 bit input sample prior to processing. THIS IS NOT NEEDED FOR .wav/.snd/.aiff INPUT! 1.10 number of channels If a raw PCM file is used as input, the '-tfc' switch indicates the number of channels (1=mono, 2=stereo). THIS IS NOT NEEDED FOR .wav/.snd/.aiff INPUT! 1.11 examples of switch settings l3enc infile.pcm out.mp3 -br 112000 -crc l3enc /home/music/pcm/newage.pcm /homem/music/mp3/newage.mp3 -br 64000 l3enc pop.wav pop.mp3 -br 96000 1.12 Encoding Recommendations Depending on the desired bitrate, the encoding process will be done with different parameter settings. 'l3enc' supports two versions of Layer-3 bitstreams called MPEG-1 and MPEG-2. The basic difference is the use of different sampling frequencies: MPEG-1 Layer 3 sampling frequencies 32, 44.1, 48 kHz MPEG-2 Layer 3 sampling frequencies 16, 22.05, 24 kHz MPEG-1 supports higher audio bandwidth and is therefore the best choice for high quality audio coding at bitrates >= 96 kbit/s (stereo) or >= 48 kbit/s (mono). For bitrates MPEG-2 offers better sound quality compared to MPEG-1. l3enc selects between MPEG-1 and MPEG-2 automatically depending on the bitrate switch (see section 1.3) For the coding of stereo files with bitrates will use the
intensity stereo technique. Note, however, that the use of intensity stereo may demage information which is needed for sound processing schemes like Dolby Surround. For bitrates >= 112 kbit/s, intensity stereo is not used. The following table summarizes the recommendations. - Coding of Mono Input bitrate coding standard ----------------------------- >= 48 kbit/s MPEG-1 - Coding of Stereo Input bitrate coding standard use of intensity stereo ------------------------------------------------------ 96 kbit/s MPEG-1 on >=112 kbit/s MPEG-1 off 2. DECODER V2.10 ============= l3dec is an ISO/MPEG Layer 3 software only decoder. It takes Layer 3 bitstream files as input and delivers PCM audio data files as output. A number of options can be selected via command line switches. Usage: l3dec [-switch1 [switch2 [...]]] If you specify no output file name and use the -sto option, the audio data is written to stdout. If you specify -sti, the decoder reads from stdin instead of the bitstream file. 2.1 : bitstream input file The format of the bitstream input file must comply with ISO/IEC IS11172-3 or IS 13818-3. The decoder will process all valid MPEG1 Layer-3 bitstream data without restrictions to bitrate or sampling frequency. It supports also MPEG2 Layer-3 low sampling frequencies. For very low bitrates an special Fraunhofer format called "MPEG 2.5" is used. 2.2 : audio data output file Audio data is output as samples of 16 bit signed integer PCM data. The default format is raw PCM data and can be either one channel or two interleaved channels. format of one (mono) channel PCM audio data: .... format of two channel (stereo) PCM audio data: ... If one or two audio channels are used depends on the encoded information in the bitstream. For stereo output data the first channel is the left channel. Information about sampling frequency and number of used channels is displayed at the beginning of the decoding process. 2.3 RIFF/WAVE format If selected by the '-wav' switch, audio data is output in RIFF/WAVE format (*.WAV) as used by Microsoft Windows. The audio data itself is still written as 16 bit PCM data as described in 2.2 but it is preceded by a WAVE-header. The WAVE-Header contains information about the number of channels (1 or 2), sampling frequency (32k/44.1k/48k) and used bits per sample (16). 2.4 SND format If selected by the '-snd' switch, audio data files are output in the SND format used on SUN and NeXT-Workstations. 2.5 AIFF format If selected by the '-aif' switch, audio data files are output in the AIFF format. 2.6 AIFC format If selected by the '-aic' switch, audio data files are output in the AIFC format. 2.7 skip frames With the '-fb' option you can skip a number of frames in the bitstream before the decoding starts. '-fb nnn' skips the first nnn frames. Each frame contains 1152 (MPEG-1) or 576 (MPEG-2) samples of audio data. Depending on the sampling frequency used, the duration of a frame is calculated as 24 msec (@ 48kHz, 24kHz), 26.1 msec (@ 44.1kHz, 22.05kHz) or 36 msec (@ 32kHz, 16 kHz). 2.8 decode only nnn frames If you want to decode only a certain number of frames, specify the '-fn' option. '-fn xxx' will decode only xxx frames (see also 2.6). 2.9 search again after loss of synchronisation Normally the decoding process is stopped, if a loss of synchronisation is detected, i.e. the synch information is incorrect. To enable decoding of partially damaged bitstream files, you may assert the '-sa' option. In this mode the decoding is not stopped and the file is searched for valid synch information until the end of file is encountered. 2.10 write audio data as ascii hex 24bit output file If the option '-h24 xxx' is specified an (additional) output file with name 'xxx' is opened. PCM Audio data is output as 24 bit ascii hex values followed by carriage return and line feed. Accuracy of the output values is 24 bit compared to the 16 bits raw output mode. Files output in 'h24' format take four times the storage capacity necessary for raw 16bit output format. 2.11 ignore error messages If errors in the bitstream are detected, the decoding process is normally halted. If the '-ign' option is specified, the decoder tries to continue with the decoding process. 2.11 accept free format bitstream If the '-ff' option is specified, a free format bitstream is accepted. 2.11 ancillary data If the bitstream contains ancillary data (user data integrated into the bitstream) the decoder can write this data into an ancillary data file. Use the switch '-a file' to specify the filename for the ancillary data. The default alignment of ancillary data is byte aligned ('-aba'). You can also use the switch '-afh' for the FhG mode. In FhG-mode, ancillary data is framed, beginning with a Sync, a length byte and has a trailing checksum. 2.12 write to stdout If the '-sto' option is specified, the PCM data output is written to stdout. 2.13 read from stdin If the '-sti' option is specified, the bitstream input is read from stdin. All brand names are registered trade marks of their respective owners.
