Embedded Audio Coder (EAC) - Encoder

Fig. 1 Functionality of the EAC encoder.

The EAC encoder a highly efficient scalable audio encoder. It converts an audio waveform to a master embedded bitstream, which can be scaled in terms of coding bitrate, number of audio channels, and audio sampling rate later. If the EAC bitstream is to be scaled, a companion file should be generated to aide the scaling.

DOWNLOAD

The EAC encoder will not be available for download now. If the reader is interested to test the your own clips, please contact the author, and the author will be happy to encode music clip of reader's choice with the EAC encoder.

You will need the Visual C++ runtime library (including msvcrt.dll etc) to use the EAC. The runtime libtrary can be downloaded here. (If you have question regarding the Visual C runtime library, please refer to this article. )

Currently, EAC encoder is only available as a command line program in Windows platform. It may be run under cmd.exe, or cygwin (a UNIX environment developed by Redhat under Windows). Please contact the author if you request support for other platforms.

USAGE

EAC encoder is a command line program in the following format

eac_encoder -i INPUT.WAV -o OUTPUT.EAC -comp COMPANION.COMP -rate RATE [OPTIONS]

The following parameters are mendatory, you must specify the options in the encoder:

• -i INPUT.WAV : The input audio waveform. Currently, EAC encoder only supports .wav audio file (mono/stereo, 16bit PCM encoded, sampling rate 8, 16, 32, 11.025, 22.05, 44.1kHz). If your audio waveform is of a different format or sampling rate, you may want to use tools to convert the audio to the .wav file. Some available tools are AFSP (free), Cooledit.
• -o OUTPUT.EAC : The output master EAC bitstream. The bitstream can be decoded (playered) by itself. For bitstream reshaping, a companion file should be used.

The following parameters are optional. But you should set the parameter:

• -comp COMPANION.COMP : The companion file. It is used in bitstream scaling/reshaping.
• -rate RATE : The coding bitrate, in number of bits per second.
• -mx MIXER : For stereo coding, MIXER specifies how the stereo channels are multiplexed.
  • MIXER=0[default] : L and R channel are separately encoded.
  • MIXER=1 : L+R and L-R channels are encoded.
• -fs SPLIT : Section split control.
  • SPLIT=0[default] : no section split.
  • SPLIT=1: split into 2 sections (allow 1/2 sampling rate reduction)
  • SPLIT=16: split into 3 sections (allow 1/2, 1/4 sampling rate reduction)

The following parameters are optional, you may simply let the program uses its default, or play with your own parameter.

• -en ENTROPY : The entropy coder in use.
  • ENTROPY=5[default] : implicit auditory masking.
  • ENTROPY=3 : implicit auditory masking with a simplified JND rule.
  • ENTROPY=0 : Sub-bitplane entropy coder without auditory masking.
• -bk BKSIZE : The size of the MLT long window. The default is 2048. You may choose other values. BKSIZE should be 2^n.
• -qs QSTEP : The quantization step size. The default is 1.0.
• -dz DEADZONE : The quantization dead zone, in proportion to the quantization step size. The default is 0.5.
• -ns SPLIT : Number of MLT long windows in a timeslot. The default is 16.
• -bark BARK : Quite threshold setting.
  • BARK=1 : default quite threshold.
  • BARK=0 : alternative quite threshold.

You should keep the other parameters at default.

Examples

eac_encoder -i mpeg_44m.wav -o mpeg_44m_64kbps.eac -comp mpeg_44m.comp -rate 65536

Encode audio waveform mpeg_44m.wav (mono, 16bit PCM coded, 44.1kHz sampling) to master bitstream mpeg_44s_64kbps.eac, generate a companion file at mpeg_44m.comp, the coding bitrate is 64kbps.

eac_encoder -i mpeg_44s.wav -o mpeg_44s_512kbps.eac -comp mpeg_44s.comp -rate 524288 -mx 1 -fs 16.

Encode audio waveform mpeg_44s.wav (stereo, 16bit PCM coded, 44.1kHz sampling) to master bitstream mpeg_44s_512kbps.eac, generate a companion file at mpeg_44s.comp, the coding bitrate is 512kbps. Use 3 section split (enable resampling at 11.025 and 22.05kHz), and encode audio as L+R and L-R channels (enable conversion to mono).

©Copyright by Jin Li, Apr 23, 2002