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 1.3.0 has been released (Read 189460 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

FLAC 1.3.0 has been released

Almost 6 years after the prior release, the new FLAC devs at Xiph.org, headed by Erik de Castro Lopo, have released a new version. Excerpts from the changelog:
  • Supports RF64+Sony Wave64
  • Can ignore timestamp/permissions from infile
  • Allow MM:SS:FF/.SS times in non-CDDA cuesheets
  • Fixed ReplayGain on stdin
  • Appropriate channel masks for 6.1/7.1 input
  • Analyse gain of 56–192 kHz files
  • Handle UTF-8 filenames on Windows
  • Support input files > 2/4 GB
  • Command-line tools can now use wildcards
Home: http://xiph.org/flac/
Compiles: see post #7 onwards

FLAC 1.3.0 has been released

Reply #1
Changelog is updated, but download links go to old FLAC 1.2.1 version
ghostman

FLAC 1.3.0 has been released

Reply #2
Which links, where? If you mean on either http://xiph.org/flac/ or the no longer updated http://flac.sourceforge.net/, probably the developers simply have not yet compiled official binaries. Perhaps some reader here has done this for his/her personal usage and platform and might be willing to share the results.

As per the penultimate item on the bulleted list of changes, the sources are now maintained using git at Xiph, not SourceForge. The relevant link: https://git.xiph.org/?p=flac.git;a=tree Download all files at once by clicking on the link entitled “snapshot”.

Finally, moving to Validated News! Congratulations to everyone who helped to bring FLAC back from its cryosleep.

FLAC 1.3.0 has been released

Reply #3
FWIW, Arch Linux already has i686 and x86_64 binaries.

FLAC 1.3.0 has been released

Reply #4
Changelog is updated, but download links go to old FLAC 1.2.1 version

Yeah, this is more of a developers release. There are no big changes, (edit: besides the UTF-8 support in Windows of course, thank Case for that) it's been made easier to compile libFLAC. There's no real need to upgrade for end users, except if you want to encode RF64, W64 or 7- or 8-channel files. (edit: or if you want to encode files with non-latin characters in the filenames) There's a checked and tested Windows binary shipped with the new FLAC frontend however, if you're curious. On other platforms you'll have to compile it yourself for now I guess.
Music: sounds arranged such that they construct feelings.

FLAC 1.3.0 has been released

Reply #5
The official download location for the source tarballs is at http://downloads.xiph.org/releases/flac/ (that is linked from the “download” page, and 1.3.0 is there).

FLAC 1.3.0 has been released

Reply #6
Further to my earlier comments about the availability of compiled binaries and packages of the source, see the posts starting here in the thread about prereleases of 1.3.0:
http://hydrogenaudio.org/forums/?showtopic...125#entry835436
Replies directed specifically at previous posts within that thread should probably go there. However, if they have significant relevance to the final version as a whole, posting them here might be more prudent.

Other posts about the final release of 1.3.0 should be kept in this thread, in the hope that discussions about it can be easily located by readers, rather than being fragmented across various locations.



FLAC 1.3.0 has been released

Reply #9
assuming everything compiles OK, I'll post a set of binaries at Rarewares later today and post here when they're available.

FLAC 1.3.0 has been released

Reply #10
I don't know much about programming and don't know much about all these binary versions, I just know they are all different (MD5 etc.).

How is the official 1.2.1b build compiled? Is it possible to have a 1.3.0 built the same way of the official one? I just want to make sure it will work with everything.

I am currently use this one: http://www.saunalahti.fi/~cse/temp/flac-1.3.0-win32.zip

Thanks.

FLAC 1.3.0 has been released

Reply #11
How is the official 1.2.1b build compiled? Is it possible to have a 1.3.0 built the same way of the official one? I just want to make sure it will work with everything.


Old official FLAC was compiled with Visual Studio 6.0. It made small binaries as it could dynamically link against msvcrt.dll that came with Windows since Windows 2000. I don't think anyone has the compiler in use anymore.

Quote
I am currently use this one: http://www.saunalahti.fi/~cse/temp/flac-1.3.0-win32.zip


That version is compiled with MSVC 2012 Update 2 using the Windows XP toolset and static linking. It works on Windows XP and newer and doesn't require any additional dlls.

 

FLAC 1.3.0 has been released

Reply #12
I noticed that FLAC compile from Case was added into foobar2000 free encoder pack.

FLAC 1.3.0 has been released

Reply #13
What about the issue (problem) regarding FLAC and file sizes greater than 2GB . . . has this been addressed?

FLAC 1.3.0 has been released

Reply #14
Yes, not sure why it was left out of the changelog. Another problem size was 4 GB. Both issues are solved and I have tested over 20 GB FLAC files succesfully. Such large FLAC actually revealed bugs in foobar2000 that are fixed in latest versions.

FLAC 1.3.0 has been released

Reply #15
Confirmed as well on Linux x86_64 with a 5GB FLAC file.

FLAC 1.3.0 has been released

Reply #16
AFAIK 2GB problem was related to MSVC:

libFLAC uses the C stdlib for file i/o.  even on my XP box with VS 2005, microsoft's stdlib implementation is still limited to 2 GB (i.e. no 64-bit off_t).

I'm reluctant to add win-specific calls to libFLAC just because MS is intentionally sabotaging portability.  every other build of flac works with large files.


FLAC 1.3.0 has been released

Reply #17
Good point about the fixes for large input files and the curious fact of these being omitted from the changelog. I have added those and some other things to the list of changes in the OP, alongside a couple of new links that are useful to have in the same place for quick access.

FLAC 1.3.0 has been released

Reply #18
Yes, not sure why it was left out of the changelog. Another problem size was 4 GB. Both issues are solved and I have tested over 20 GB FLAC files succesfully. Such large FLAC actually revealed bugs in foobar2000 that are fixed in latest versions.


So I assume continuously streaming flac will now also work?

FLAC 1.3.0 has been released

Reply #19
32 and 64 bit compiles are at Rarewares now - ICL 12.1 compiles.


FLAC 1.3.0 has been released

Reply #21
Crash.
Do these ICL 12.1 compiles require a SSE2 cpu by any chance?

The Intel 12.1 compiler defaults to arch SSE2. Therefore, my guess is the answer to your question is yes.

FLAC 1.3.0 has been released

Reply #22
I tested these compiles on my Intel Core2. A CD image (44.1/16/stereo, 53min 50sec) was encoded with -8 setting.

Encoding time (smaller is better):

Code: [Select]
Case               76.2 s
ktf                77.0 s
john33 32bit       79.6 s
lamedude SSE 32bit 79.5 s

john33 64bit       77.6 s
lamedude SSE 64bit 76.6 s

Case: flac built by Case using MSVC 2012 (see post #12 in this thread)
john33: flac built by john33 using ICC 12.1 (see post #20 in this thread)
ktf: flac built by ktf using MinGW (see post #134)
lamedude: flac built by lamedude using MSVC 2012 (see post #132)

FLAC 1.3.0 has been released

Reply #23
And for FLAC -5:

Code: [Select]
Case               21.0 s
ktf                23.2 s
john33 32bit       21.2 s
lamedude SSE 32bit 21.1 s

john33 64bit       26.3 s
lamedude SSE 64bit 25.3 s

FLAC 1.3.0 has been released

Reply #24
32 and 64 bit compiles are at Rarewares now - ICL 12.1 compiles.

Thank you very much, John

I can confirm that the 64-bit encoder is slower than the 32-bit version. Running Win 8 Pro, i5 2500k, using foobar2000 1.2.6 as a front-end, 32-bit averages around 245x real-time, while 64-bit is around 222x.