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.
Topic: FLAC compression improvement patch (Read 50434 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

FLAC compression improvement patch

Reply #75
What about setting -8 -e as --best? Before --best was the same compression as -8. This way --best makes much sense imho.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

FLAC compression improvement patch

Reply #76
I am kind of torn on that, considering -8 was same as --best, but you could still squeeze some more bytes by specifying -A tukey something flattop (sorry don't remember the command).
It does make sense for --best to be equivalent with maximum compression though.

Comparing a single 661 MB file is quite different than ~150 GB of material.
I don't see anything wrong with that, the encoding speed doesn't increase or decrease with size. Besides, my 812 files (~27GB in wav) showed the same encoding speeds as my 661MB file

FLAC compression improvement patch

Reply #77
What about setting -8 -e as --best? Before --best was the same compression as -8. This way --best makes much sense imho.

IMHO people got used to -8 and --best equivalence.
And maybe some will start to complain that FLAC 1.3.1 --best is much slower than 1.3.0 --best.

FLAC compression improvement patch

Reply #78
What about setting -8 -e as --best? Before --best was the same compression as -8. This way --best makes much sense imho.

IMHO people got used to -8 and --best equivalence.
And maybe some will start to complain that FLAC 1.3.1 --best is much slower than 1.3.0 --best.

Just because people got used different things behave the same i mean someone who knows to entering --best and creates smaller files has to expect more processing time.
On the other hand with entering --best you simply can print out a link to the flacCL homepage 
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

FLAC compression improvement patch

Reply #79
What about setting -8 -e as --best? Before --best was the same compression as -8. This way --best makes much sense imho.

Yes, but why -8 -e as best? Why not -8 -p? Or -8 -e -p? Here some results from my standard test set: (with the new presets)

Code: [Select]
-8:       ratio 56.51%, encoding: 52.5x, decoding: 400.8
-8 -p:    ratio 56.47%, encoding: 13.1x, decoding: 396.4
-8 -e:    ratio 56.45%, encoding: 13.4x, decoding: 394.5
-8 -e -p: ratio 56.40%, encoding: 1.6x,  decoding: 394.3

You could even add more apodizations. Or you can re-enable rice parameter search (which was disabled and the switch deprecated quite some time ago because it's gains are very, very small)  and exact rice bits calculation. You see where this is going? --best can be made as slow as you want with very little gain. If --best is made different from -8, it opens the floodgates, we can keep on slowing down FLAC with little gain.
Music: sounds arranged such that they construct feelings.

FLAC compression improvement patch

Reply #80
Now that makes sense. Wonder where this --best comes from then.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

FLAC compression improvement patch

Reply #81
Flac 1.3.1 is released.
http://www.phoronix.com/scan.php?page=news...&px=MTg0ODk
The FLAC 1.3.1 release improves decoding efficiency for all bit-depths but in particular for 24-bits for IA32. The FLAC 1.3.1 update also has faster encoding using SSE and AVX CPU insruction set extensions, various fixes, new adopization functions, I/O buffering improvements on Windows, build system improvements, and a wide range of fixes.

Good job.
Best thanksgiving gift.
Thanks for all the hard work.

FLAC compression improvement patch

Reply #82
Flac 1.3.1 is released.
http://www.phoronix.com/scan.php?page=news...&px=MTg0ODk
The FLAC 1.3.1 release improves decoding efficiency for all bit-depths but in particular for 24-bits for IA32. The FLAC 1.3.1 update also has faster encoding using SSE and AVX CPU insruction set extensions, various fixes, new adopization functions, I/O buffering improvements on Windows, build system improvements, and a wide range of fixes.

Good job.
Best thanksgiving gift.
Thanks for all the hard work.

There is currently a topic on the 1.3.1 release here in the News Submissions section waiting to be validated. http://www.hydrogenaud.io/forums/index.php?showtopic=107611
Every night with my star friends / We eat caviar and drink champagne
Sniffing in the VIP area / We talk about Frank Sinatra
Do you know Frank Sinatra? / He's dead

FLAC compression improvement patch

Reply #83
Necromancing the previous discussion on this page, I'd say that --best is arguably too sacred to change, but what about good (?) old --super-secret-totally-impractical-compression-level? ;-)

FLAC compression improvement patch

Reply #84
but what about good (?) old --super-secret-totally-impractical-compression-level? ;-)

It's not that difficult to use -8ep instead of -8 right? And, as a bonus, it's not that secret but still totally impractical!
Music: sounds arranged such that they construct feelings.

FLAC compression improvement patch

Reply #85
Latest CUETools build (2.1.6, http://www.cuetools.net/wiki/CUETools_Download) includes a new version of CUETools.Flake.exe that makes heavy use of new window functions. At compression levels -7 and -8 it uses 1 tukey window + 2 partial tukey windows + 3 punchout tukey windows + 4 smaller punchout tukey windows, to a total of 10 window functions. Results on my sample files have been quite good, but i'll leave it for somebody else to compare it to new libFLAC.
CUETools 2.1.6

 

FLAC compression improvement patch

Reply #86
Latest CUETools build (2.1.6, http://www.cuetools.net/wiki/CUETools_Download) includes a new version of CUETools.Flake.exe that makes heavy use of new window functions. At compression levels -7 and -8 it uses 1 tukey window + 2 partial tukey windows + 3 punchout tukey windows + 4 smaller punchout tukey windows, to a total of 10 window functions. Results on my sample files have been quite good, but i'll leave it for somebody else to compare it to new libFLAC.

Nice! I hope the CL version will come around soon
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!