0 imperceptible -1 perceptible but not annoying -2 slightly annoying -3 annoying -4 very annoying
The measurement procedure has only been calibrated for a sampling rate of 48 kHz. The measurement procedure works for monaural or binaural (stereo) signals. Use ResampAudio to resample audio files with other sampling rates to 48 kHz. The measurement procedure also assumes that the files have been time and gain aligned. Use CompAudio to determine the required gain and delay compensation and CopyAudio to actually modify the gain and time-align the files.
This program implements the basic version of PEAQ. The measurement results on a standard database do not fall within the tight bounds specified in the standard (see the report cited below for a discussion of the difficulties in interpreting the standard). However, the results are close enough to be useful for quality impairment measurements.
T. Thiede et al, "PEAQ - The ITU Standard for Objective Measurement of Perceived Audio Quality", J. Audio Eng. Soc., vol. 48, pp. 3-29, Jan.-Feb. 2000.
P. Kabal, "An Examination and Interpretation of ITU-R BS.1387: Perceptual Evaluation of Audio Quality", TSP Lab Technical Report, Dept. Electrical & Computer Engineering, McGill University, May 2002. (http://www.TSP.ECE.McGill.CA/MMSP/Documents)
"clip_MOV" or "no_clip_MOV" - Clip MOV values (default "no_clip_MOV")
"PC_init=V" - Initial value for the pattern correction factors
(default 0)
"PD_factor=V" - Forgetting factor for the maximum probability of
detection calculation (default 1).
"overlap_delay" or "no_overlap_delay" - Overlap warmup frames (frames
before the data boundary) and the delay for calculating modulation
difference and noise loudness values (default "overlap_delay")
"data_bounds" or "no_data_bounds" - Ignore frames with small data
values at the beginning and end of the reference (default
"data_bounds")
"end_min=N" - Stop processing at the frame that contains at least
N samples (default value of N is 1024). The "-l" command line option
can be used to shift the entire processing. For instance "-l -1024:"
will add half a frame of zero padding before the start of data.
"EHS_lag_start=N" where N is 0 or 1. The default is 1.
"auto" - determine the input file type from the file header "AU" or "au" - AU audio file "WAVE" or "wave" - WAVE file "AIFF" or "aiff" - AIFF or AIFF-C sound file "noheader" - headerless (non-standard or no header) audio file "SPHERE" - NIST SPHERE audio file "ESPS" - ESPS sampled data feature file "IRCAM" - IRCAM soundfile "SPPACK" - SPPACK file "INRS" - INRS-Telecom audio file "SPW" - Comdisco SPW Signal file "CSL" or "NSP" - CSL NSP file "text" - Text audio file
"Format, Start, Sfreq, Swapb, Nchan, ScaleF"
"undefined" - Headerless files will be rejected "mu-law8" - 8-bit mu-law data "A-law8" - 8-bit A-law data "unsigned8" - offset-binary 8-bit integer data "integer8" - two's-complement 8-bit integer data "integer16" - two's-complement 16-bit integer data "integer24" - two's-complement 24-bit integer data "integer32" - two's-complement 32-bit integer data "float32" - 32-bit floating-point data "float64" - 64-bit floating-point data "text" - text data
"native" - no byte swapping "little-endian" - file data is in little-endian byte order "big-endian" - file data is in big-endian byte order "swap" - swap the data bytes as the data is read
"default" - Scale factor chosen appropriate to the type of data. The
scaling factors shown below are applied to the data in the file.
8-bit mu-law: 1/32768
8-bit A-law: 1/32768
8-bit integer: 128/32768
16-bit integer: 1/32768
24-bit integer: 1/(256*32768)
32-bit integer: 1/(65536*32768)
float data: 1
"<number or ratio>" - Specify the scale factor to be applied to
the data from the file.
"undefined, 0, 8000., native, 1, default"