lossyWAV 1.1.0 Development Thread., Added noise WAV bit reduction method. Options

[Nick.C]

Following the release of lossyWAV 1.0.0b, I feel it is time to kick off development of the next minor release.

Items currently on the list for inclusion in 1.x.0:

1.1.0: STDIN input;
1.1.0: STDOUT output;
1.1.0: Channel independent bit removal;
1.1.0: Reversion to same bits-to-remove for all channels;
1.1.0: Noise shaping;
1.2.0: Checking of S (=L-R) channel for matrix surround content;

If you have any ideas, suggestions, code optimisations, etc, please post them here.

CODE

lossyWAV 1.1.0b, Copyright (C) 2007,2008 Nick Currie. Copyleft.

This program is free software: you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free Software
Foundation, either version 3 of the License, or (at your option) any later
version.

This program is distributed in the hope that it will be useful,but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.  See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with
this program.  If not, see <http://www.gnu.org/licenses/>.

Process Description:

lossyWAV adds white noise to the processed output. The amount of added noise is
based on analysis of the signal levels in the frequency range 20Hz to 16kHz.

If signals above the upper limiting frequency are at an even lower level, they
can be swamped by the added noise. This is usually inaudible, but the behaviour
can be changed by specifying a higher --limit (in the range 16kHz to 20kHz).

For many audio signals, there is little content at very high frequencies, and
forcing lossyWAV to keep the added noise level lower than the content at these
frequencies can increase the bitrate dramatically for no perceptible benefit.

Usage   : lossyWAV <input wav file> <options>

Example : lossyWAV musicfile.wav

Quality Options:

-I, --insane        highest quality output, suitable for transcoding;
-E, --extreme       high quality output, also suitable for transcoding;
-S, --standard      default quality output, considered to be transparent;
-P, --portable      good quality output for DAP use. Not considered to be fully
                    transparent, but considered fit for its intended purpose.

Standard Options:

-c, --check         check if WAV file has already been processed; default=off.
                    errorlevel=16 if already processed, 0 if not.
-C, --correction    write correction file for processed WAV file; default=off.
-f, --force         forcibly over-write output file if it exists; default=off.
-h, --help          display help.
-L, --longhelp      display extended help.
-M, --merge         merge existing lossy.wav and lwcdf.wav files.
-o, --outdir <t>    destination directory for the output file(s).
-v, --version       display the lossyWAV version number.

Advanced Options:

-                   if filename="-" then WAV input is taken from STDIN.
    --blockdist     show distribution of lowest significant bit of input
                    codec-blocks and bit-removed codec-blocks.
-D, --dither <n>    enable variable PDF dither of output; default=off;
                    0 = rectangular; 1 = triangular; 0.5 = half way between.
-l, --limit <n>     set upper frequency limit to be used in analyses to n Hz;
                    (16000<=n<=20000), default = 16000.
    --linkchannels  Revert to original single bits-to-remove value for all
                    channels rather than channel dependent bits-to-remove.
-q, --quality <n>   quality preset (10=highest quality, 0=lowest bitrate;
                    default = --standard = 5; --insane = 10; --extreme = 7.5;
                    --portable = 2.5)
    --sampledist    show distribution of lowest significant bit of input
                    samples and bit-removed samples.
    --scale <n>     scaling factor from WaveGain, etc; (0.0<n<=8.0),default=1.0
-s, --shaping <n>   enable fixed noise shaping; (0.00<=n<=1.00); default=q/10;
                    0.00 = off, 1.00 = 100% effectiveness, 0.50 = 50%, etc.
    --stdinname <t> pseudo filename to use when input from STDIN.
    --stdout        write processed WAV output to STDOUT.
-w, --writetolog    create (or append to) lossyWAV.log in the output directory.

System Options:

-B, --below         set process priority to below normal.
-d, --detail        enable detailed bits-to-remove information output mode
    --low           set process priority to low.
-n, --nowarnings    suppress lossyWAV warnings.
-Q, --quiet         significantly reduce screen output.
    --silent        no screen output.

Special thanks:

David Robinson      for the publication of his lossyFLAC method, guidance, and
                    the motivation to implement the method as lossyWAV.
Horst Albrecht      for ABX testing, valuable support in tuning the internal
                    presets, constructive criticism and all the feedback.
Sebastian Gesemann  for the noise shaping coefficients and help in using them
                    in the lossyWAV noise shaping implementation.
Don Cross           for the Complex-FFT algorithm used.

Link to the hydrogenaudio wiki article

Suggested foobar2000 converter setup:

lossyFLAC:

CODE

Encoder: c:\windows\system32\cmd.exe
Extension: lossy.flac
Parameters: /d /c c:\"program files"\bin\lossywav - --standard --silent --stdout|c:\"program files"\bin\flac - -b 512 -5 -f -o%d
Format is: lossless or hybrid
Highest BPS mode supported: 24

lossyTAK:

CODE

Encoder: c:\windows\system32\cmd.exe
Extension: lossy.tak
Parameters: /d /c c:\"program files"\bin\lossywav - --standard --silent --stdout|c:\"program files"\bin\takc -e -p2m -fsl512 -ihs - %d
Format is: lossless or hybrid
Highest BPS mode supported: 24

lossyWV:

CODE

Encoder: c:\windows\system32\cmd.exe
Extension: lossy.wv
Parameters: /d /c c:\"program files"\bin\lossywav - --standard --silent --stdout|c:\"program files"\bin\wavpack -hm --blocksize=512 --merge-blocks -i - %d
Format is: lossless or hybrid
Highest BPS mode supported: 24

There is a known problem within foobar2000 (although more likely to do with cmd.exe itself) when running an executable within the cmd.exe command line from a path which includes spaces. The suggested fix for this is to enclose the element of the path which contains spaces within double quotation marks ("), e.g. c:\"program files"\directory_where_executable_is\executable_name

Change log 1.1.0c: 30/04/2009
Exactly as 1.1.0b except that the WINE incompatibility issue has been fixed.

Executable here.
Source here.

Change log 1.1.0b: 03/08/08
FFT lengths will now increase for higher bitrate audio, i.e. 88.2/96kHz, 176.4/192kHz and 352.8/384kHz;
improved logfile output and --detail output;
reference threshold constants for rectangular dither and triangular dither have been calculated so added noise should be the same for dither off and any dither level between 0 and 1 - the number of bits-to-remove will however reduce with "increasing" dither.

Change log 1.1.0: 12/07/08
Certain advanced parameters removed for final release.

Change log 1.0.1x RC4: 12/07/08
Final release candidate prior to release of 1.1.0

Change log 1.0.1w RC3: 02/07/08
Code tidied up a bit more (yet again....);
--wine parameter modified to stop the program using Windows API function calls when using piped input (should hopefully stop crashing under Wine).

Change log 1.0.1v RC2: 30/06/08
Code tidied up a bit more (again....);
--wine parameter implemented to stop the program using the GetLastError Windows API call when using piped input (should stop crashing under Wine).

Change log 1.0.1u RC1: 20/06/08
Code tidied up a bit more;
--bitdist parameter introduced to allow user to "examine" the distribution of lowest set bit on a codec-block by codec-block basis, channels treated separately.

Change log beta 1.0.1t: 11/06/08
Revision to STDIN handling - bug found where last codec-block read from foobar2000 using STDIN input was not being written to the output file.

Change log beta 1.0.1s: 09/06/08
Revision to STDIN handling. Now (fingers crossed) should work successfully inside Foobar2000;
Code and help tidied up;
Dither function fixed and augmented. Taking on board a statement by SG with respect to using a dither function somewhere between rectangular (rand - 0.5) and triangular (rand-0.5)+(rand-0.5), i.e. (rand-0.5)+s*(rand-0.5) {0<=s<=1}. s=0 = rectangular dither; s=1 = triangular dither. -D, --dither now requires a supplementary <n> in the range 0<=n<=1.

Change log beta 1.0.1r: 03/06/08
Implementation of fast square root function using lookup tables for fxtract(ed) exponent and mantissa of input value;
--scale parameter corrected to accepted values in the range 0<n<=8.

Change log beta 1.0.1q: 30/05/08
Codec-block overflow bug (when codec-block-size=4096) corrected;

Change log beta 1.0.1p: 29/05/08
Quality synonym automatic noise shaping bug corrected;

Change log beta 1.0.1o: 29/05/08
Spreading function spread-zones and spreading-function string modified to allow finer control of high frequency zones;
Code "recovered" from 1.0.1e after a minor hardware failure

Change log beta 1.0.1n: 26/05/08
Implementation of -H, --highskew <n> parameter. Functionally identical to the internal skewing applied to the FFT results (-36dB @ 20Hz to 0dB at 3.45kHz) except applied from 3.45kHz upwards. Valid in the range 0 to 36 (0=default=no high skew applied).

Change log beta 1.0.1m: 25/05/08
reintroduction of max-inter-block-change implementation limits increase in bits-to-remove between codec-blocks to 1 bit.

Change log beta 1.0.1k: 23/05/08
static maximum_bits_to_remove limitation re-applied in serial with dynamic maximum_bits_to_remove limitation;
Automatic noise shaping now applied using a shaping-factor of quality-level / 10.

Change log beta 1.0.1j: 23/05/08
-q <n> quality selection moved to advanced settings;
-E, --excessive changed to --extreme; -I, --insane added, equivalent to -q 10;
--lowpass changed to -l, --limit in keeping with discussion;
Process Description text added to --longhelp.

Change log beta 1.0.1i: 23/05/08
-q <n> quality selection moved to advanced settings;
-E, --excessive; -N, --normal; -P, --portable quality "names" introduced following discussion in the development thread. These equate to -q 7.5; -q 5.0 and -q 2.5 respectively.

Change log beta 1.0.1h: 20/05/08
minimum bits to keep values changed for -q 0 and -q 1 to 2.333 and 2.667 respectively.

Change log beta 1.0.1g: 22/05/08
Reference_threshold > threshold_index > bits_to_remove calculation refined;
spreading function string modified;
minimum bits to keep values changed for -q 0 and -q 1;
--writetolog (-w) parameter implemented to write minimal output to "lossyWAV.log". Appends to existing file if already exists;
--lowpass <n> parameter re-implemented to allow users to set upper frequency limit of the range that lossyWAV uses in its analyses (16000<=n<=24000).

Change log beta 1.0.1f: 20/05/08
Filenaming logic "improved" when STDIN and STDOUT used together.

Change log beta 1.0.1e: 19/05/08
STDIN / STDOUT mode tidied up. Use the following as a flossy.bat file for foobar conversion:

CODE

@echo off
z:\bin\lossyWAV %1 --low --nowarnings --quiet %3 %4 %5 %6 %7 %8 %9 --stdout|z:\bin\flac - -5 -f -b 512 -o%2

Unfortunately, due to the nature of piped input to FLAC, the lossyWAV 'fact' chunk is lost. This means no record is kept within the file that is has been processed with lossyWAV (however, the lower the quality setting of the processing, the more likely the bitrate will be an obvious indicator that the file has indeed been processed with lossyWAV);
Minor error found and amended in revised remove_bits procedure, no minimum_bits_to_keep value was being applied, although this has little impact at -q >= 2;
New parameter --linkchannels implemented to revert to old remove_bits method whereby all channels share the same bits_to_remove. Implementing this, I found an error in the original which was forcing more bits to be lost to clipping prevention than should have been (i.e. output was more conservative).

Change log beta 1.0.1d: 18/05/08
STDIN / STDOUT mode modified again (use '-' as a filename to enable STDIN input, --stdout to enable STDOUT output).
Console output has been redirected to 'con', rather than STDOUT.

Change log beta 1.0.1c: 16/05/08
STDIN / STDOUT mode modified again (use '-' as a filename to enable STDIN input).

Change log beta 1.0.1b: 15/05/08
Channel independent bit-removal implemented;
STDIN / STDOUT mode modified - still very much a work in progress.

Change log beta 1.0.1: 14/05/08
STDIN / STDOUT mode commenced.

This post has been edited by Nick.C: May 14 2009, 13:10

[sauvage78]

halb27:
you're right -q 2 was near transparent with 1.0.1f & it is even more with 1.0.1h.
I tried to quickly ABX 1.0.1h -q2 vs original but I gave up after 5min as it was a boring task, I think it is still ABXable but it will take me 40 min just to prove to others something I am already sure for myself & even if -q2 is ABXable the fact that I can hear a difference doesn't mean that this difference sounds ugly.

So yes -q2 was listenable for DAP with this specific sample, specially in a noisy environnement, but my personal interest for lossywav is not in DAP even if I like the flexibility of using any lossless format.

Nick.C:
I was thinking lossywav was just removing unearable random noise in the back that takes much space to losslessly encode, but obviously it also add its own noise. I was thinking lossywav was as a kind of "lossless audio cleaner" by optimizing wav before encoding to lossless, I was thinking lossywav was not introducing "quantization like noise", but I was wrong it seems that even without quantization it add lossy-like additionnal noise. So I haven't followed all the lossywav developments, but now I am interested in knowing if lossywav applies some encoding strategy at some settings & not at some other settings. Even without understanding in deep the used techniques, just to understand the logical step within the scale.

Why is Q0 to Q4 for DAP, Q5 magically called "transparent", Q6 & Q7 as good as lossless & Q8 to Q10 for transcoding ??? My technical knowledge is very limited but I would like to know if these claims are rationnal or just arbitrary. Is it just me or lossywav have become more lossy than it was supposed to be in the beginning ? If there is no different techniques at different settings isn't the scale a little too wide ? Wouldn't it be logic to have only 2 recommended settings within the scale covering each specific use:

"-near lossy" for DAP, covering Q0 to Q5 use, targetting transparency.
"-near lossless" for archiving, covering Q6 to Q10 use, targetting transcodable archiving.

I don't mean that the actual scale is bad, I like it, freedom is always good. But it is missleading for newbies as I ended using Q0 for archive thinking lossywav was overkill anyway no matter the setting. I dunno but

"-near lossy" could be the same as let's say -q 3 (ABXing can help here)
"-near lossless" could be the same as ... well I don't know but the least agressive possible while still optimizing the lossless encoding. (ABXing is useless here, only technical knowledge so I let the codecs developpers decide together, not only you Nick.C ... Monty, Ivan, Gabriel may have an hintfull opinion here)

I know this is still very arbitrary but it would help newbies using lossywav as it should, & not lowering its reputation ...

for me for there is only 4 uses for hybrid:
100% lossless for Archive + "-near lossy" for DAP.
"-near lossless" for Archive + 100% lossy for DAP.
"-near lossless" for Archive + "-near lossy" for DAP.
hybrid lossless part for Archive + hybrid lossy part for listening.

so in the end two well tuned settings can cover it all.

This post has been edited by sauvage78: May 22 2008, 21:29

[Nick.C]

halb27:
you're right -q 2 was near transparent with 1.0.1f & it is even more with 1.0.1h.
I tried to quickly ABX 1.0.1h -q2 vs original but I gave up after 5min as it was a boring task, I think it is still ABXable but it will take me 40 min just to prove to others something I am already sure for myself & even if -q2 is ABXable the fact that I can hear a difference doesn't mean that this difference sounds ugly.

So yes -q2 was listenable for DAP with this specific sample, specially in a noisy environnement, but my personal interest for lossywav is not in DAP even if I like the flexibility of using any lossless format.

Nick.C:
I was thinking lossywav was just removing unearable random noise in the back that takes much space to losslessly encode, but obviously it also add its own noise. I was thinking lossywav was as a kind of "lossless audio cleaner" by optimizing wav before encoding to lossless, I was thinking lossywav was not introducing "quantization like noise", but I was wrong it seems that even without quantization it add lossy-like additionnal noise. So I haven't followed all the lossywav developments, but now I am interested in knowing if lossywav applies some encoding strategy at some settings & not at some other settings. Even without understanding in deep the used techniques, just to understand the logical step within the scale.

Why is Q0 to Q4 for DAP, Q5 magically called "transparent", Q6 & Q7 as good as lossless & Q8 to Q10 for transcoding ??? My technical knowledge is very limited but I would like to know if these claims are rationnal or just arbitrary. Is it just me or lossywav have become more lossy than it was supposed to be in the beginning ? If there is no different techniques at different settings isn't the scale is little too wide ? Wouldn't it be logic to have only 2 recommended settings covering each specific use:

-near lossy for DAP, covering Q0 to Q5, targetting transparency.
-near lossless for archiving, covering Q6 to Q10, targetting transcoding.

I don't mean that the actual scale is bad, I like it, freedom is always good. But it is missleading for newbies as I ended using Q0 for archive thinking lossywav was overkill anyway no matter the setting. I dunno but

-near lossy could be the same as -q 3 (ABXing can help here)
-near lossless could be the same as ... well I don't know but the least agressive possible while still optimizing the lossless encoding. (ABXing is useless here, only technical knowledge so let the codecs developpers decide together, not only Nick.C)

I know this is very arbitrary but it would help newbies using lossywav as it should, & not lowering its reputation ...

lossyWAV has always added noise to the output, see the original lossyFLAC post started by David (2Bdecided) - the method used is his.

The Delphi transcode of the Matlab original started as just that - an attempt to duplicate David's method in a form that more users could make use of - Matlab is after all rather expensive....

Once agreement between both versions was achieved (about 0.3.0 roughly from memory), other ideas blossomed and added to the method.

The preset scale started as -1, -2 and -3. -1 = best, -2 = transparent, -3 = more aggressive. This evolved over time into -1 to -5, -1 to -7 and finally -q 10 to -q 0. As it states in the help:

CODE

Quality Options:

-q, --quality <n>   quality preset (10=highest quality, 0=lowest bitrate;
                    -q 5 is generally accepted to be transparent)
                    default=-q 5.

On the subject of transparency, as many problem samples as have been pointed out to halb27 and me have been included in the refinement of the quality presets. v0.6.4 RC1 was thought to be good until the day after release Alex B found a problem sample that halb27 also ABX'ed.

Transparency at -q 5 is not an arbitrary claim - although not every possible sample has been tested, therefore it is still "only" a claim (it should be noted that a very small group of users has been involved in the developmental ABX'ing process).

However, unless the optional (and still beta) noise shaping is selected, the actual bit removal process only rounds off the least significant <bits-to-remove> bits of each sample in a codec-block, thus adding white noise. If noise shaping is selected, the added noise is modified and a lot of it is pushed much higher in the spectrum (thanks to SebastianG's noise shaping coefficients).

Another, more advanced, method of noise shaping is in the early stages of development at present - but don't hold your breath.

Personally, I only use lossyWAV & FLAC as a DAP transcode - my lossless FLAC archive will remain intact.

The bottom line on the quality presets: 10 = best, 0 = worst - use one which you are comfortable with.

Thanks for taking the time to post.

Nick.

With this sample and -q 0 setting, it looks like level of added noise changes very frequently, and this modulation is more noticeable than the noise itself.

I'll have a look at the bits-to-remove detail for the sample at -q 0 and see if it is large stepping in the bits-to-remove (i.e. added noise) - from what you say, it would seem to be.

I had a look:

CODE

Z:\WAV\tmp>lossywav "..\_swav\01 - Ginnungagap - The Black Hole.original.wav" -q 0 --detail -f
lossyWAV beta 1.0.1h, Copyright © 2007,2008 Nick Currie. Copyleft.
This is free software under the GNU GPLv3+ license; There is NO WARRANTY, to
the extent permitted by law. <http://www.gnu.org/licenses/> for details.
%lossyWAV Warning% : Detailled output mode enabled
Processing : 01 - Ginnungagap - The Black Hole.original.wav
Format : 44.10kHz; 2 ch.; 16 bit.
Progress :
Block Time 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 Tot.
====================================================================
0 0.00s. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
16 0.19s. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
32 0.37s. 5 7 10 10 10 9 9 10 9 9 9 8 7 7 9 8 136
48 0.56s. 9 11 10 10 10 10 9 10 9 9 9 9 9 9 8 9 150
64 0.74s. 7 9 8 8 6 7 6 8 7 7 7 7 7 7 8 10 119
80 0.93s. 10 10 10 9 9 9 9 8 8 8 8 9 8 8 9 11 143
96 1.11s. 10 11 11 11 11 11 11 11 11 10 11 10 10 10 10 10 169
112 1.30s. 10 10 10 9 10 10 9 9 10 10 10 10 9 9 9 10 154
128 1.49s. 9 10 9 9 10 10 10 9 9 10 10 9 8 9 9 8 148
144 1.67s. 9 9 9 9 9 8 9 9 9 9 9 9 9 9 8 8 141
160 1.86s. 8 9 8 9 9 9 9 9 9 8 9 9 9 9 9 8 140
176 2.04s. 8 8 9 8 8 8 9 9 8 9 8 8 9 8 8 8 133
192 2.23s. 7 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 127
208 2.41s. 7 8 7 7 8 8 7 8 8 7 8 8 7 7 7 8 120
224 2.60s. 7 8 7 7 6 8 7 8 7 7 7 7 7 7 7 6 113
240 2.79s. 7 7 8 6 7 6 7 6 7 7 6 7 7 7 7 7 109
256 2.97s. 7 7 7 7 7 7 6 7 7 7 7 6 6 7 7 7 109
272 3.16s. 6 6 7 7 7 7 7 8 8 7 8 8 8 8 8 8 118
288 3.34s. 7 8 7 8 7 7 8 7 7 7 7 7 7 7 7 7 115
304 3.53s. 7 7 7 8 8 8 7 7 7 7 7 8 7 7 8 6 116
320 3.72s. 7 7 7 7 7 7 7 7 7 7 7 7 7 6 6 7 110
336 3.90s. 7 7 5 7 7 6 7 6 6 6 7 6 6 7 7 7 104
352 4.09s. 7 7 7 7 7 7 6 7 6 6 7 7 6 7 7 7 108
368 4.27s. 6 7 6 6 6 6 6 7 7 7 7 6 7 6 7 6 103
384 4.46s. 6 7 6 6 6 7 6 6 6 6 6 6 6 6 7 6 99
400 4.64s. 6 6 7 7 6 6 7 6 6 7 7 7 7 6 6 6 103
416 4.83s. 6 6 5 6 6 6 6 6 6 6 6 6 6 5 6 5 93
432 5.02s. 6 6 5 6 6 6 6 6 6 6 6 5 6 5 6 6 93
448 5.20s. 5 6 6 6 6 6 6 6 5 5 5 5 6 5 5 6 89
464 5.39s. 6 6 6 5 5 6 6 6 6 6 5 6 6 5 5 5 90
480 5.57s. 5 6 5 5 6 6 6 6 6 6 5 5 6 6 6 6 91
496 5.76s. 6 5 5 5 6 5 5 6 6 6 6 6 5 6 5 6 89
512 5.94s. 6 5 5 5 5 5 6 6 10 10 10 10 9 9 9 10 120
528 6.13s. 9 8 8 8 9 9 8 10 10 10 9 10 10 9 9 9 145
544 6.32s. 9 9 8 8 8 7 8 7 8 8 8 7 7 7 7 7 123
560 6.50s. 7 6 7 7 7 10 10 10 10 10 9 8 9 9 9 9 137
576 6.69s. 8 8 8 8 8 11 11 11 11 11 10 11 10 11 10 11 158
592 6.87s. 10 11 11 10 10 10 10 10 10 10 10 9 10 10 9 9 159
608 7.06s. 9 9 9 10 10 10 9 9 9 10 10 10 10 10 10 9 153
624 7.24s. 10 9 9 9 9 9 9 9 10 10 9 9 10 9 9 9 148
640 7.43s. 9 9 10 9 9 9 7 9 9 9 9 9 8 9 9 8 141
656 7.62s. 9 9 9 8 9 9 9 8 9 8 9 8 8 8 8 8 136
672 7.80s. 9 8 8 8 8 8 7 7 8 9 9 8 8 9 8 7 129
688 7.99s. 7 7 8 8 8 8 8 8 9 7 7 8 7 7 8 7 122
704 8.17s. 8 8 8 8 7 7 7 7 8 6 7 7 6 7 8 6 115
720 8.36s. 7 7 8 8 7 7 7 7 7 8 7 7 6 6 7 6 112
736 8.54s. 7 7 6 7 5 6 7 7 6 6 5 7 6 7 7 7 103
752 8.73s. 7 7 6 6 6 7 7 7 7 7 7 8 8 7 8 8 113
768 8.92s. 8 8 8 8 7 7 7 7 7 7 7 6 6 7 8 7 115
784 9.10s. 8 7 7 7 7 6 8 7 7 7 7 7 7 8 8 7 115
800 9.29s. 7 7 7 7 7 7 7 8 8 7 7 7 7 7 7 7 114
816 9.47s. 7 7 6 6 7 7 7 7 8 8 7 7 8 7 7 7 113
832 9.66s. 7 7 7 7 7 7 6 7 7 8 7 7 7 6 6 7 110
848 9.85s. 7 6 7 5 7 7 6 6 7 7 7 7 7 7 7 7 107
864 10.03s. 6 6 6 7 6 7 7 6 7 5 6 6 4 6 7 7 99
880 10.22s. 7 6 6 7 8 7 8 7 8 7 8 8 8 6 8 8 117
896 10.40s. 8 8 8 7 7 7 8 7 8 7 7 7 7 8 7 7 118
912 10.59s. 7 7 6 7 7 7 7 7 5 7 6 6 7 6 6 7 105
928 10.77s. 7 7 7 6 6 7 7 7 7 7 7 7 7 7 8 7 111
944 10.96s. 7 8 7 7 7 7 8 8 7 7 8 8 7 7 7 7 117
960 11.15s. 7 7 7 7 7 7 7 7 7 7 7 7 8 7 7 7 113
976 11.33s. 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 112
992 11.52s. 7 7 7 6 6 7 7 7 7 6 7 7 7 6 8 11 113
1008 11.70s. 10 10 10 10 10 9 9 9 8 8 8 8 8 9 10 9 145
1024 11.89s. 10 9 9 10 10 9 9 9 9 8 8 9 8 8 9 7 141
1040 12.07s. 7 7 8 8 7 8 7 7 7 6 7 8 10 10 10 9 126
1056 12.26s. 10 9 9 9 9 8 8 8 8 8 8 8 11 11 11 11 146
1072 12.45s. 11 11 10 11 10 11 11 10 11 10 11 10 10 10 10 10 167
1088 12.63s. 10 10 10 9 9 9 7 8 9 10 9 10 9 9 9 10 147
1104 12.82s. 10 8 10 10 9 10 10 10 9 10 9 9 9 10 9 10 152
1120 13.00s. 7 9 9 8 9 9 8 9 9 9 9 9 9 8 9 9 139
1136 13.19s. 8 9 9 9 9 9 9 9 7 9 9 9 9 9 8 7 138
1152 13.37s. 9 7 8 8 8 8 8 8 8 8 8 8 8 8 8 8 128
1168 13.56s. 7 8 8 7 8 5 8 8 8 8 8 7 7 7 8 8 120
1184 13.75s. 7 8 8 8 8 8 8 8 8 8 7 8 7 7 8 7 123
1200 13.93s. 7 7 8 7 7 7 7 7 6 8 8 7 7 7 7 6 113
1216 14.12s. 6 7 6 7 7 7 7 6 7 8 6 6 7 7 7 7 108
1232 14.30s. 7 7 7 7 7 7 6 7 7 7 7 7 7 7 7 7 111
1248 14.49s. 7 7 7 8 8 8 8 8 7 8 6 8 7 6 6 7 116
1264 14.68s. 8 7 8 7 6 8 7 8 8 8 7 8 6 7 8 7 118
1280 14.86s. 7 8 8 7 7 7 7 7 7 7 7 7 6 6 7 6 111
1296 15.05s. 7 7 7 7 7 6 7 7 6 6 7 7 7 7 7 7 109
1312 15.23s. 7 6 7 7 7 7 7 7 7 7 6 6 7 6 7 6 107
1328 15.42s. 7 7 6 6 7 7 6 7 7 7 7 7 7 7 7 7 109
1344 15.60s. 7 5 6 6 6 7 7 6 6 6 6 7 7 6 6 5 99
1360 15.79s. 6 7 6 7 6 6 6 6 6 6 6 6 6 7 7 7 101
1376 15.98s. 7 6 7 5 5 6 6 6 6 6 6 6 6 6 6 6 96
1392 16.16s. 7 6 7 5 6 6 5 7 6 5 5 6 6 5 6 6 94
1408 16.35s. 6 6 6 6 6 6 6 5 6 6 6 6 6 6 5 5
Average : 2.78MB; 7.4112 bits; 32.93x; 0.50s;
%lossyWAV Warning% : 0.0112 bits not removed due to clipping.
%lossyWAV Warning% : 17 sample(s) clipped to limiting amplitude.

CODE

Z:\WAV\tmp>lossywav "..\_swav\01 - Ginnungagap - The Black Hole.original.wav" -q 0 --detail -f --minbits 3
lossyWAV beta 1.0.1h, Copyright © 2007,2008 Nick Currie. Copyleft.
This is free software under the GNU GPLv3+ license; There is NO WARRANTY, to
the extent permitted by law. <http://www.gnu.org/licenses/> for details.
%lossyWAV Warning% : Detailled output mode enabled
Processing : 01 - Ginnungagap - The Black Hole.original.wav
Format : 44.10kHz; 2 ch.; 16 bit.
Progress :
Block Time 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 Tot.
====================================================================
0 0.00s. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
16 0.19s. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
32 0.37s. 5 7 10 10 10 9 9 10 9 9 9 8 7 7 9 8 136
48 0.56s. 9 10 10 10 10 10 9 10 9 9 9 9 9 9 8 9 149
64 0.74s. 7 9 8 8 6 7 6 8 7 7 7 7 7 7 8 10 119
80 0.93s. 10 10 10 9 9 9 9 8 8 8 8 9 8 8 9 10 142
96 1.11s. 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 160
112 1.30s. 9 10 10 9 10 9 9 9 10 10 10 10 9 9 9 10 152
128 1.49s. 9 10 9 9 9 10 10 9 9 10 10 9 8 9 9 8 147
144 1.67s. 9 9 9 9 9 8 9 9 9 9 9 9 9 9 8 8 141
160 1.86s. 8 9 8 9 9 9 9 9 9 8 9 9 9 9 9 8 140
176 2.04s. 8 8 9 8 8 8 9 9 8 9 8 8 9 8 8 8 133
192 2.23s. 7 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 127
208 2.41s. 7 8 7 7 8 8 7 8 8 7 8 8 7 7 7 8 120
224 2.60s. 7 8 7 7 6 8 7 8 7 7 7 7 7 7 7 6 113
240 2.79s. 7 7 8 6 7 6 7 6 7 7 6 7 7 7 7 7 109
256 2.97s. 7 7 7 7 7 7 6 7 7 7 7 6 6 7 7 7 109
272 3.16s. 6 6 7 7 7 7 7 8 8 7 8 8 8 8 8 8 118
288 3.34s. 7 8 7 8 7 7 8 7 7 7 7 7 7 7 7 7 115
304 3.53s. 7 7 7 8 8 8 7 7 7 7 7 8 7 7 8 6 116
320 3.72s. 7 7 7 7 7 7 7 7 7 7 7 7 7 6 6 7 110
336 3.90s. 7 7 5 7 7 6 7 6 6 6 7 6 6 7 7 7 104
352 4.09s. 7 7 7 7 7 7 6 7 6 6 7 7 6 7 7 7 108
368 4.27s. 6 7 6 6 6 6 6 7 7 7 7 6 7 6 7 6 103
384 4.46s. 6 7 6 6 6 7 6 6 6 6 6 6 6 6 7 6 99
400 4.64s. 6 6 7 7 6 6 7 6 6 7 7 7 7 6 6 6 103
416 4.83s. 6 6 5 6 6 6 6 6 6 6 6 6 6 5 6 5 93
432 5.02s. 6 6 5 6 6 6 6 6 6 6 6 5 6 5 6 6 93
448 5.20s. 5 6 6 6 6 6 6 6 5 5 5 5 6 5 5 6 89
464 5.39s. 6 6 6 5 5 6 6 6 6 6 5 6 6 5 5 5 90
480 5.57s. 5 6 5 5 6 6 6 6 6 6 5 5 6 6 6 6 91
496 5.76s. 6 5 5 5 6 5 5 6 6 6 6 6 5 6 5 6 89
512 5.94s. 6 5 5 5 5 5 6 6 10 10 10 10 9 9 9 10 120
528 6.13s. 9 8 8 8 9 9 8 10 10 10 9 10 10 9 9 9 145
544 6.32s. 9 9 8 8 8 7 8 7 8 8 8 7 7 7 7 7 123
560 6.50s. 7 6 7 7 7 10 10 10 10 10 9 8 9 9 9 9 137
576 6.69s. 8 8 8 8 8 11 11 11 10 10 10 10 10 10 10 10 153
592 6.87s. 10 10 10 10 10 10 10 10 10 10 10 9 10 10 9 9 157
608 7.06s. 9 9 9 10 10 10 9 9 9 10 10 10 10 10 10 9 153
624 7.24s. 10 9 9 9 9 9 9 9 10 9 9 9 9 9 9 9 146
640 7.43s. 9 9 10 9 9 9 7 9 9 9 9 9 8 9 9 8 141
656 7.62s. 9 9 9 8 9 9 9 8 9 8 9 8 8 8 8 8 136
672 7.80s. 9 8 8 8 8 8 7 7 8 9 9 8 8 9 8 7 129
688 7.99s. 7 7 8 8 8 8 8 8 9 7 7 8 7 7 8 7 122
704 8.17s. 8 8 8 8 7 7 7 7 8 6 7 7 6 7 8 6 115
720 8.36s. 7 7 8 8 7 7 7 7 7 8 7 7 6 6 7 6 112
736 8.54s. 7 7 6 7 5 6 7 7 6 6 5 7 6 7 7 7 103
752 8.73s. 7 7 6 6 6 7 7 7 7 7 7 8 8 7 8 8 113
768 8.92s. 8 8 8 8 7 7 7 7 7 7 7 6 6 7 8 7 115
784 9.10s. 8 7 7 7 7 6 8 7 7 7 7 7 7 8 8 7 115
800 9.29s. 7 7 7 7 7 7 7 8 8 7 7 7 7 7 7 7 114
816 9.47s. 7 7 6 6 7 7 7 7 8 8 7 7 8 7 7 7 113
832 9.66s. 7 7 7 7 7 7 6 7 7 8 7 7 7 6 6 7 110
848 9.85s. 7 6 7 5 7 7 6 6 7 7 7 7 7 7 7 7 107
864 10.03s. 6 6 6 7 6 7 7 6 7 5 6 6 4 6 7 7 99
880 10.22s. 7 6 6 7 8 7 8 7 8 7 8 8 8 6 8 8 117
896 10.40s. 8 8 8 7 7 7 8 7 8 7 7 7 7 8 7 7 118
912 10.59s. 7 7 6 7 7 7 7 7 5 7 6 6 7 6 6 7 105
928 10.77s. 7 7 7 6 6 7 7 7 7 7 7 7 7 7 8 7 111
944 10.96s. 7 8 7 7 7 7 8 8 7 7 8 8 7 7 7 7 117
960 11.15s. 7 7 7 7 7 7 7 7 7 7 7 7 8 7 7 7 113
976 11.33s. 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 112
992 11.52s. 7 7 7 6 6 7 7 7 7 6 7 7 7 6 8 11 113
1008 11.70s. 10 10 10 10 10 9 9 9 8 8 8 8 8 9 10 9 145
1024 11.89s. 10 9 9 10 9 9 9 9 9 8 8 9 8 8 9 7 140
1040 12.07s. 7 7 8 8 7 8 7 7 7 6 7 8 10 10 10 9 126
1056 12.26s. 10 9 9 9 9 8 8 8 8 8 8 8 11 11 11 10 145
1072 12.45s. 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 160
1088 12.63s. 10 10 10 9 9 9 7 8 9 10 9 10 9 9 9 10 147
1104 12.82s. 10 8 9 10 9 10 10 10 9 10 9 9 9 10 9 10 151
1120 13.00s. 7 9 9 8 9 9 8 9 9 9 9 9 9 8 9 9 139
1136 13.19s. 8 9 9 9 9 9 9 9 7 9 9 9 9 9 8 7 138
1152 13.37s. 9 7 8 8 8 8 8 8 8 8 8 8 8 8 8 8 128
1168 13.56s. 7 8 8 7 8 5 8 8 8 8 8 7 7 7 8 8 120
1184 13.75s. 7 8 8 8 8 8 8 8 8 8 7 8 7 7 8 7 123
1200 13.93s. 7 7 8 7 7 7 7 7 6 8 8 7 7 7 7 6 113
1216 14.12s. 6 7 6 7 7 7 7 6 7 8 6 6 7 7 7 7 108
1232 14.30s. 7 7 7 7 7 7 6 7 7 7 7 7 7 7 7 7 111
1248 14.49s. 7 7 7 8 8 8 8 8 7 8 6 8 7 6 6 7 116
1264 14.68s. 8 7 8 7 6 8 7 8 8 8 7 8 6 7 8 7 118
1280 14.86s. 7 8 8 7 7 7 7 7 7 7 7 7 6 6 7 6 111
1296 15.05s. 7 7 7 7 7 6 7 7 6 6 7 7 7 7 7 7 109
1312 15.23s. 7 6 7 7 7 7 7 7 7 7 6 6 7 6 7 6 107
1328 15.42s. 7 7 6 6 7 7 6 7 7 7 7 7 7 7 7 7 109
1344 15.60s. 7 5 6 6 6 7 7 6 6 6 6 7 7 6 6 5 99
1360 15.79s. 6 7 6 7 6 6 6 6 6 6 6 6 6 7 7 7 101
1376 15.98s. 7 6 7 5 5 6 6 6 6 6 6 6 6 6 6 6 96
1392 16.16s. 7 6 7 5 6 6 5 7 6 5 5 6 6 5 6 6 94
1408 16.35s. 6 6 6 6 6 6 6 5 6 6 6 6 6 6 5 5
Average : 2.78MB; 7.3919 bits; 38.17x; 0.43s;
%lossyWAV Warning% : 0.0056 bits not removed due to clipping.
%lossyWAV Warning% : 3 sample(s) clipped to limiting amplitude.

CODE

Z:\WAV\tmp>lossywav "..\_swav\01 - Ginnungagap - The Black Hole.original.wav" -q 0 --detail -f --minbits 3.5
lossyWAV beta 1.0.1h, Copyright © 2007,2008 Nick Currie. Copyleft.
This is free software under the GNU GPLv3+ license; There is NO WARRANTY, to
the extent permitted by law. <http://www.gnu.org/licenses/> for details.
%lossyWAV Warning% : Detailled output mode enabled
Processing : 01 - Ginnungagap - The Black Hole.original.wav
Format : 44.10kHz; 2 ch.; 16 bit.
Progress :
Block Time 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 Tot.
====================================================================
0 0.00s. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
16 0.19s. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
32 0.37s. 5 7 10 10 10 9 9 9 9 9 9 8 7 7 9 8 135
48 0.56s. 9 9 10 10 10 9 9 9 9 9 9 9 9 9 8 9 146
64 0.74s. 7 9 8 8 6 7 6 8 7 7 7 7 7 7 8 10 119
80 0.93s. 10 10 10 9 9 9 9 8 8 8 8 9 8 8 9 10 142
96 1.11s. 10 10 10 9 10 9 9 10 9 10 9 9 9 9 10 9 151
112 1.30s. 9 9 9 9 9 9 9 9 9 9 9 10 9 9 9 9 145
128 1.49s. 9 9 9 9 9 9 9 9 9 9 9 9 8 9 9 8 142
144 1.67s. 9 9 9 9 9 8 9 9 9 9 9 9 9 9 8 8 141
160 1.86s. 8 9 8 9 9 9 9 9 9 8 9 9 9 9 9 8 140
176 2.04s. 8 8 9 8 8 8 9 9 8 9 8 8 9 8 8 8 133
192 2.23s. 7 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 127
208 2.41s. 7 8 7 7 8 8 7 8 8 7 8 8 7 7 7 8 120
224 2.60s. 7 8 7 7 6 8 7 8 7 7 7 7 7 7 7 6 113
240 2.79s. 7 7 8 6 7 6 7 6 7 7 6 7 7 7 7 7 109
256 2.97s. 7 7 7 7 7 7 6 7 7 7 7 6 6 7 7 7 109
272 3.16s. 6 6 7 7 7 7 7 8 8 7 8 8 8 8 8 8 118
288 3.34s. 7 8 7 8 7 7 8 7 7 7 7 7 7 7 7 7 115
304 3.53s. 7 7 7 8 8 8 7 7 7 7 7 8 7 7 8 6 116
320 3.72s. 7 7 7 7 7 7 7 7 7 7 7 7 7 6 6 7 110
336 3.90s. 7 7 5 7 7 6 7 6 6 6 7 6 6 7 7 7 104
352 4.09s. 7 7 7 7 7 7 6 7 6 6 7 7 6 7 7 7 108
368 4.27s. 6 7 6 6 6 6 6 7 7 7 7 6 7 6 7 6 103
384 4.46s. 6 7 6 6 6 7 6 6 6 6 6 6 6 6 7 6 99
400 4.64s. 6 6 7 7 6 6 7 6 6 7 7 7 7 6 6 6 103
416 4.83s. 6 6 5 6 6 6 6 6 6 6 6 6 6 5 6 5 93
432 5.02s. 6 6 5 6 6 6 6 6 6 6 6 5 6 5 6 6 93
448 5.20s. 5 6 6 6 6 6 6 6 5 5 5 5 6 5 5 6 89
464 5.39s. 6 6 6 5 5 6 6 6 6 6 5 6 6 5 5 5 90
480 5.57s. 5 6 5 5 6 6 6 6 6 6 5 5 6 6 6 6 91
496 5.76s. 6 5 5 5 6 5 5 6 6 6 6 6 5 6 5 6 89
512 5.94s. 6 5 5 5 5 5 6 6 10 10 10 10 9 9 9 9 119
528 6.13s. 9 8 8 8 9 9 8 10 10 10 9 10 9 9 9 9 144
544 6.32s. 9 9 8 8 8 7 8 7 8 8 8 7 7 7 7 7 123
560 6.50s. 7 6 7 7 7 10 10 10 10 10 9 8 9 9 9 9 137
576 6.69s. 8 8 8 8 8 10 10 11 10 10 9 10 9 10 10 10 149
592 6.87s. 10 10 10 9 9 10 9 10 9 9 9 9 9 9 9 9 149
608 7.06s. 9 9 9 9 10 9 9 9 9 9 9 9 9 9 9 9 145
624 7.24s. 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 144
640 7.43s. 9 9 9 9 9 9 7 9 9 9 9 9 8 9 9 8 140
656 7.62s. 9 9 9 8 9 9 9 8 9 8 9 8 8 8 8 8 136
672 7.80s. 9 8 8 8 8 8 7 7 8 9 9 8 8 9 8 7 129
688 7.99s. 7 7 8 8 8 8 8 8 8 7 7 8 7 7 8 7 121
704 8.17s. 8 8 8 8 7 7 7 7 8 6 7 7 6 7 8 6 115
720 8.36s. 7 7 8 8 7 7 7 7 7 8 7 7 6 6 7 6 112
736 8.54s. 7 7 6 7 5 6 7 7 6 6 5 7 6 7 7 7 103
752 8.73s. 7 7 6 6 6 7 7 7 7 7 7 8 8 7 8 8 113
768 8.92s. 8 8 8 8 7 7 7 7 7 7 7 6 6 7 8 7 115
784 9.10s. 8 7 7 7 7 6 8 7 7 7 7 7 7 8 8 7 115
800 9.29s. 7 7 7 7 7 7 7 8 8 7 7 7 7 7 7 7 114
816 9.47s. 7 7 6 6 7 7 7 7 8 8 7 7 8 7 7 7 113
832 9.66s. 7 7 7 7 7 7 6 7 7 8 7 7 7 6 6 7 110
848 9.85s. 7 6 7 5 7 7 6 6 7 7 7 7 7 7 7 7 107
864 10.03s. 6 6 6 7 6 7 7 6 7 5 6 6 4 6 7 7 99
880 10.22s. 7 6 6 7 8 7 8 7 8 7 8 8 8 6 8 8 117
896 10.40s. 8 8 8 7 7 7 8 7 8 7 7 7 7 8 7 7 118
912 10.59s. 7 7 6 7 7 7 7 7 5 7 6 6 7 6 6 7 105
928 10.77s. 7 7 7 6 6 7 7 7 7 7 7 7 7 7 8 7 111
944 10.96s. 7 8 7 7 7 7 8 8 7 7 8 8 7 7 7 7 117
960 11.15s. 7 7 7 7 7 7 7 7 7 7 7 7 8 7 7 7 113
976 11.33s. 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 112
992 11.52s. 7 7 7 6 6 7 7 7 7 6 7 7 7 6 8 10 112
1008 11.70s. 10 10 10 9 10 9 9 9 8 8 8 8 8 9 10 9 144
1024 11.89s. 10 9 9 9 9 9 9 9 9 8 8 9 8 8 9 7 139
1040 12.07s. 7 7 8 8 7 8 7 7 7 6 7 8 10 10 10 9 126
1056 12.26s. 9 9 9 9 9 8 8 8 8 8 8 8 10 10 10 10 141
1072 12.45s. 9 10 9 10 9 10 10 9 9 9 9 9 9 9 9 9 148
1088 12.63s. 9 10 9 9 9 9 7 8 9 9 9 9 9 9 9 9 142
1104 12.82s. 9 8 9 9 9 9 9 9 9 9 9 9 9 9 9 9 143
1120 13.00s. 7 9 9 8 9 9 8 9 9 9 9 9 9 8 9 9 139
1136 13.19s. 8 9 9 9 9 9 9 9 7 9 9 9 9 9 8 7 138
1152 13.37s. 9 7 8 8 8 8 8 8 8 8 8 8 8 8 8 8 128
1168 13.56s. 7 8 8 7 8 5 8 8 8 8 8 7 7 7 8 8 120
1184 13.75s. 7 8 8 8 8 8 8 8 8 8 7 8 7 7 8 7 123
1200 13.93s. 7 7 8 7 7 7 7 7 6 8 8 7 7 7 7 6 113
1216 14.12s. 6 7 6 7 7 7 7 6 7 8 6 6 7 7 7 7 108
1232 14.30s. 7 7 7 7 7 7 6 7 7 7 7 7 7 7 7 7 111
1248 14.49s. 7 7 7 8 8 8 8 8 7 8 6 8 7 6 6 7 116
1264 14.68s. 8 7 8 7 6 8 7 8 8 8 7 8 6 7 8 7 118
1280 14.86s. 7 8 8 7 7 7 7 7 7 7 7 7 6 6 7 6 111
1296 15.05s. 7 7 7 7 7 6 7 7 6 6 7 7 7 7 7 7 109
1312 15.23s. 7 6 7 7 7 7 7 7 7 7 6 6 7 6 7 6 107
1328 15.42s. 7 7 6 6 7 7 6 7 7 7 7 7 7 7 7 7 109
1344 15.60s. 7 5 6 6 6 7 7 6 6 6 6 7 7 6 6 5 99
1360 15.79s. 6 7 6 7 6 6 6 6 6 6 6 6 6 7 7 7 101
1376 15.98s. 7 6 7 5 5 6 6 6 6 6 6 6 6 6 6 6 96
1392 16.16s. 7 6 7 5 6 6 5 7 6 5 5 6 6 5 6 6 94
1408 16.35s. 6 6 6 6 6 6 6 5 6 6 6 6 6 6 5 5
Average : 2.78MB; 7.3322 bits; 32.50x; 0.51s;
%lossyWAV Warning% : 0.0014 bits not removed due to clipping.

The issue might be the 11 bits removed around 1.1 seconds. They are removed because the signal is considered by the algorithm to be strong enough to mask the noise - in this instance obviously not. More analysis of the spectral makeup of the sample around the 1.1 second mark is required.

This post has been edited by Nick.C: May 22 2008, 21:32

[lvqcl]

The issue might be the 11 bits removed around 1.1 seconds. They are removed because the signal is considered by the algorithm to be strong enough to mask the noise - in this instance obviously not. More analysis of the spectral makeup of the sample around the 1.1 second mark is required.

Well, the issue I hear is (I think) quick changes in noise intensity. They're highlighted in this picture:

It's the difference between original and lossy signals.

(btw: if you delete first 16463 samples (about 0.38 sec) then Adobe Audition says that the rest signal has only 15 bits )

[Nick.C]

The issue might be the 11 bits removed around 1.1 seconds. They are removed because the signal is considered by the algorithm to be strong enough to mask the noise - in this instance obviously not. More analysis of the spectral makeup of the sample around the 1.1 second mark is required.

Well, the issue I hear is (I think) quick changes in noise intensity. They're highlighted in this picture:

It's the difference between original and lossy signals.

(btw: if you delete first 16463 samples (about 0.38 sec) then Adobe Audition says that the rest signal has only 15 bits )

How long are the steps in the signal - if it's a multiple of 11.6msec (512 samples) then the steps are caused by different bits-to-remove being applied to the output and adding different intensities of white noise.