Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.

Poll

FLAC POLL

=========== Question 1° "What FLAC encoder(s) do you use?" ===========
Reference FLAC encoder - https://xiph.org/flac/
FFmpeg (libavcodec) FLAC encoder - https://www.ffmpeg.org/
FLACCL (cuetools) encoder - http://cue.tools/wiki/FLACCL
Flake - https://flake-enc.sourceforge.net/
I DON'T ENCODE INTO FLAC FORMAT
=========== Question 2° "What settings of FLAC encoder(s) do you use?" ===========
Reference FLAC encoder    -0
Reference FLAC encoder    -1
Reference FLAC encoder    -2
Reference FLAC encoder    -3
Reference FLAC encoder    -4
Reference FLAC encoder    -5
Reference FLAC encoder    -6
Reference FLAC encoder    -7
Reference FLAC encoder    -8
Reference FLAC encoder    -8 -p
Reference FLAC encoder    -8 -p -e
Reference FLAC encoder    with other custom settings (please comment)
FLACCL (cuetools) encoder    -0
FLACCL (cuetools) encoder    -1
FLACCL (cuetools) encoder    -2
FLACCL (cuetools) encoder    -3
FLACCL (cuetools) encoder    -4
FLACCL (cuetools) encoder    -5
FLACCL (cuetools) encoder    -6
FLACCL (cuetools) encoder    -7
FLACCL (cuetools) encoder    -8
FLACCL (cuetools) encoder    -9
FLACCL (cuetools) encoder    -10
FLACCL (cuetools) encoder    -11
FLACCL (cuetools) encoder    with other custom settings (please comment)
Flake -0
Flake -1
Flake -2
Flake -3
Flake -4
Flake -5
Flake -6
Flake -7
Flake -8
Flake -9
Flake -10
Flake -11
Flake -12
Flake with other custom settings (please comment)
FFmpeg FLAC encoder    -0
FFmpeg FLAC encoder    -1
FFmpeg FLAC encoder    -2
FFmpeg FLAC encoder    -3
FFmpeg FLAC encoder    -4
FFmpeg FLAC encoder    -5
FFmpeg FLAC encoder    -6
FFmpeg FLAC encoder    -7
FFmpeg FLAC encoder    -8
FFmpeg FLAC encoder    -9
FFmpeg FLAC encoder    -10
FFmpeg FLAC encoder    -11
FFmpeg FLAC encoder    -12
FFmpeg FLAC encoder    with other custom settings (please comment)

Voting closes: 2027-01-01 00:31:40

Topic: FLAC encoders Poll  (Read 10813 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

FLAC encoders Poll

This poll is dedicated to FLAC encoders and compression options. Poll has two questions. Members can vote up to 8 options. 8 votes is bigger than usual 5 votes so it can cover situations where member uses different FLAC encoders AND different compression options.  

Poll will be active for 2 years.

Enjoy  :)

Re: FLAC encoders Poll

Reply #1 –
I would like to ask people who use -8 -p -e whether it compresses substantially better than -8p.  

-p squizes an extra 0.1% more, while -e just does 0.01%. I can't even notice difference in 1 kbit of bitrate for 44.1/16 in comparison -8pe vs -8p.
Or does -e help on specific genres, high definition 24/96 etc...?

Re: FLAC encoders Poll

Reply #2 –
I use FLAC compiled with mutithreading support:

flac.exe -8 --threads=12 -V
TAPE LOADING ERROR

Re: FLAC encoders Poll

Reply #3 –
my utility script uses both ffmpeg and reference flac depending on the task, just because it's easy to, respectively
(1) convert from other formats
(2) seek by samples (accurately splitting whole-disk images)
and the result is lossless flac either way
a fan of AutoEq + Meier Crossfeed

Re: FLAC encoders Poll

Reply #4 –
I use reference FLAC with these settings: -s --ignore-chunk-sizes --lax -8pr15 -q15 -A -V -p --no-padding --preserve-modtime --no-replay-gain - -o %d

Re: FLAC encoders Poll

Reply #5 –
Quote from: TuNk77 on
I use reference FLAC with these settings: -s --ignore-chunk-sizes --lax -8pr15 -q15 -A -V -p --no-padding --preserve-modtime --no-replay-gain - -o %d
Looks like a foobar2000 line  (the "%d", the -s and the --ignore-chunk-sizes, which you n ormally wouldn't use in a command-line on a file).
--no-replaygain is the default, so you don't need to type that. --preserve-modtime is also a default, but you cannot get modtime preserved when encoding from stdin I think.

But also:
 * You use -q15 to force max-length predictor coefficients, but -p to brute-force optimize it. There are even two "-p" (the other is in "-8pr15").
 * You use "-A" with no functions; since there is -V following it, it doesn't err out. But it means you replace -8's windowing by -5's, weakening compression. 
 * Curious how often you actually get finer partitions than 8. If you are on Windows, you can check for partition order 9 on all .flac in a directory by flac -ac *.flac | find "partition_order=9" and see if it outputs anything. In any case, it doesn't do much worse than take up some extra time.
 * --no-padding? If you intend to tag, that just makes you wait for longer.

Since you are in for heavier compression than -8p, I suggest something like
-s --ignore-chunk-sizes -V8pr8 -A "subdivide_tukey(5)" - -o %d
The "subdivide_tukey(5)" takes time, but not as much as the "p".
If you are doing stuff on command-line rather than through fb2k, that shortens to -V8pr8 -A "subdivide_tukey(5)"

Re: FLAC encoders Poll

Reply #6 –
Quote from: Porcus on
* --no-padding? If you intend to tag, that just makes you wait for longer.

Since you are in for heavier compression than -8p, I suggest something like
-s --ignore-chunk-sizes -V8pr8 -A "subdivide_tukey(5)" - -o %d
The "subdivide_tukey(5)" takes time, but not as much as the "p".
If you are doing stuff on command-line rather than through fb2k, that shortens to -V8pr8 -A "subdivide_tukey(5)"
Thank you very much for the tips, Porcus :)

I use cue + flac, so i don't need to tag the flac-file ;) That's why i use --no-padding

As for the for heavier compression, your commandeline options squeezed a bit extra, compared to my previous commandline setting.
From the foobar2000 log: Total encoding time: 0:28.578, 14.59x realtime.
Not bad, if you ask me.

Thanks again

Re: FLAC encoders Poll

Reply #7 –
Reference FLAC encoder -7

Fairly sure it was the work done by @Porcus within the last couple of years that showed using -7 with the 1.4+ versions was about the same compression rate as older version -8 setting, but then you get a slightly faster encoding and decoding performance.

Re: FLAC encoders Poll

Reply #8 –
For decoding, any significant difference between -7 and -8 would be a surprise, but yeah @bennetng produced such figures on an old telephone: https://hydrogenaud.io/index.php/topic,123025.msg1016437.html#msg1016437 .
Surprise because maximum complexity is the same for those two presets, although it could of course happen that -8 can find an estimated predictor that outdoes the fixed ones, when -7 cannot.
But flac decodes already blitz fast anyway.

As for encoding, long story not-short-but-at-least-shorter:
 * @ktf found out that one reason flake (which ffmpeg's flac encoder is based on) compressed better than reference version 1.3.x, was doing the least-squares math in double precision.
 * That was ported into reference flac version 1.4.0, and upon release we already knew that it would improve compression, at a modest cost in speed (double precision means more bits to compute, taking more time). It would not affect fixed-predictor computation, hence not presets 0 to 2 (which got a slightly faster due to better code somewhere else in the encoder).
 * Of course, if you don't like the slowdown - myself I don't mind, having upgraded computer once over ten years - you could instead step one preset down -8 to -7 in your case. That's what I tested. On that particular computer, -7 took a five percent slowdown, which I don't mind - but size-wise it even beat the three times as slow old "-8p".
And the worst exception was 0.06 percent. I.e., Laibach's "NATO" album was 0.06 percent smaller with old -8 than with new -7, and none of the 37 other albums was as "bad for 1.4.0 -7".

Re: FLAC encoders Poll

Reply #9 –
Thanks for the explanation and clarification Porcus.