FLAC 1.2.1 is out, the major addition being support for all RIFF/AIFF metadata, including Broadcast Wave Format (BWF) via the new --keep-foreign-metadata option. There are many other small improvements and bug fixes; see the changelog entry for complete details.

Thanks Josh! I hope a Universal Mac version is due out soon

FLAC becomes better and better - compression rate and speed is growing, additional RIFF/AIFF metadata can be stored. For de'facto standard lossless codec (at least in open source world) it is great.

Good work, Josh, thank you very much.

BTW, it's time to update wiki page of "Lossless comparison" and remove 'unable to support RIFF chunks' from cons.

Me too! Thanks for all your hard work Josh!

Is the new option --keep-foreign-metadata recommended for files generated by EAC? I think it is only needed for WAV files with metainformation such as timestamp points etc used by Wavelab and other Sound-Editting apps.

Great work! For my purposes FLAC really is the most complete lossless codec available, thanks jcoalsen.
(a really minor gripe though: could the windows installer please ask me if I would like to have a desktop icon?)

No, it isn't.

Thanks for the great work Josh!

However, I am not too sure what SF# 1764105 - "Fixed bug where importing non-CDDA cuesheets would cause an invalid lead-out track number" means. Can someone please tell me what this means in plain English?

Also, is this a serious enough bug to warrant a trancode of all my FLAC images from v1.1.4 to v1.2.1? I am totally confused

Read this post. Previously, when embedding a cue sheet in a flac file that was created from non-CD audio, the lead-out track number would be incorrect. Starting with this new version of FLAC, the lead-out track number for non-CD audio will always be 255.
You don't have to transcode anything unless you have such files. In other words, if your FLAC images are CD-DA rips (44.1 kHz, 16 bit, stereo), you're fine.

Universal binary ready for download at my website...
Enjoy!

I wasn't sure what your URL for downloads of the Mac version was, but just found it in your profile:

http://homepage.mac.com/krmathis/

Thanks! Downloaded it...

just got a report that if you use "flac --replay-gain" but no padding option, the resulting file will have a very small padding block. I'll upload a new installer soon, or the workaround is to use a -P option with your preferred amount of padding.

uploaded new installer with --replay-gain fix here:
http://downloads.sourceforge.net/flac/flac-1.2.1b.exe

Thanks for the new version.

Seems the problems reading WAVE_FORMAT_EXTENSIBLE headers by STDIN, repported in this post, are solved.

Josh. Is this a fix at source code level or part of the MS Windows installer package only?
Meaning, will we see a new source code tarball?

By the way. I appreciate the work you put into this.
Thanks!

haven't decided on that yet, full releases are quite a pain. if other urgent bugs pop up I might. in the meantime you can pick up all urgent fixes to 1.2.1 via the maintenance branch, i.e. check out by branch tag FLAC_RELEASE_1_2_1_MAINTENANCE_BRANCH

Josh, these releases might "be a pain", but having several 1.2.1 versions with different code branches and possible problems should not be allowed. I recommend you make the fixes 1.2.2 and not 1.2.1 branch. People need to rely on a version, and having bugs causes people to be leery to use that version I believe. Just my opinion.

I see..
WIll check out the maintenance branch, and follow closely in case you decide to put out a 1.2.1b tarball.

Now padding with replaygaining works well, thank you.

--keep-foreign-metadata seems to be fine:

046335deef981b35f3227f0315cc9bc8 *file.wav
046335deef981b35f3227f0315cc9bc8 *file_decoded.wav

a. should i expect any problems?
b. what exactly is the latest cmd version? (i have extracted one from installer flac-1.2.1b.exe, is that the latest?)

tnx

a. This release is considered to be stable, so expect no problems
b. Yes, the one from the installer is the latest. By the way, the only difference is a really small fix (use of the --replay-gain option caused no padding).

egor, tnx, the problems with '--keep-foreign-metadata' is the question?

Thank you, Josh

Not only thanks for your work on FLAC, this is also a textbook example of a good news post. I would like to see more news submissions like that from other people; brief and to the point.




Done.

"RIFF/AIFF metadata, including Broadcast Wave Format (BWF) via the new --keep-foreign-metadata"

So can I use FLAC to compress wav files used in multitrack projects in programs like SONAR, Cubase, Pro Tools etc and maintain the time position data and stuff like that? This would be really great

should work if they are valid wave files. wave64 (sound forge) does not work yet though.

it has been beta tested but it is also a new feature, that's why the --delete-input-file option has been disabled when --keep-foreign-metadata is used. it's a good idea to compare like this for a while until you're satisfied it is working for you.

I would say to treat anything that says flac 1.2.1 as having that bug. I fixed the windows installer because it's quick and is the most widely used version, so that most of the people who don't know about the bug won't get bit.

The new Winamp plugin doesn't play Ogg FLAC files.
I have tried it with the ogg and the oga extension.

'--keep-foreign-metadata' is really exciting news. This feature opens the door to the world of sample libraries where for example loop points needs to be stored. I won't be surprised to find support for flac in future versions of Native Instruments Kontakt, for example.

This way, musicians may benefit from flac where loading times of huge libraries will decrease significantly!

Is the zipped download updated as well?

VUPlayer 2.49, AudioTester 1.4, and FLAC Adobe Audition Filter all updated by James Chapman for FLAC 1.2.1. support.

Download web page links below for each FLAC 1.2.1 supporting program/utility:

VUPlayer: http://www.vuplayer.com/vuplayer.php

AudioTester: http://www.vuplayer.com/other.php

Audition FLAC File Filter: http://www.vuplayer.com/audition.php

unfortunately no.

my temporary mirror of zipped latest (hopefully) version:
http://somestuff.org/downloads/flac121b/

Thanks for all your great work Josh. Flac is where it's at!

Running the Windows installer under Parallels on a MacBook, Flac frontend seems to fail to install. Installation wizard gives the error:
"there was an error registering a needed file with the operating system: flac frontend may not run."
and on attempting to run flac frontend, I get the error:
"Component 'MSCOMTL.OCX' or one of its dependencies not correctly registered: a file is missing or invalid."

I presume this is a problem with Parallels (build 3214), but thought I should report it.

My other question is regarding the universal binaries posted by krmathis: Where should I place these so that mac programs that use flac such as xact or cog, will find them and use the newer version?

Thanks again!

I know this was mentioned before, but it would be great to have the Nero Flac plugin included in the downloads. The proper installation is to copy the plugin to the C:\Program Files\Common Files\Ahead\AudioPlugins folder. It would really be grand if the plugin could be included in the zip download. It's a pain having to run the installer just for the nero plugin.

Or you could just provide a link to afterdawns neroplugin site:

http://neroplugins.afterdawn.com/

To best of my knowledge, the Nero FLAC 3rd party plugin hasn't been updated since FLAC 1.1.0 (August 6, 2003), and I reported a bug in loading or saving mono FLAC 44.1khz 16 bit files, but it was never fixed. I can't trust that 3rd party Nero plugin. Would be nice if the Nero developers would include native FLAC read and write support in Nero...

Let me see...
* xACT are self contained. Just replace the FLAC binary inside the .app and you are all set.
* Cog use the FLAC framework, and not the command line binary.

I agree that FLAC should be supported natively by NERO. I wonder if the bloated Nero 8 does so.

Thanks.

If I understand, this means I have to wait for a future version of Cog that uses the new framework?

What about slimserver, (which is what I most use to listen to flacs)?

And is there any point in transcoding from flacs encoded in 1.1.3 or 1.1.4? Would I gain much space? I have about 1.3 TB of flacs.

Somebody with the connections/authority/"ability to code" might want to get on the ffmpeg guys' cases. At least with ffmpeg 0.4.9_p20070616-r1 from Gentoo Portage, playing of 1.2.1 flacs appears to be broken. (Using ffplay actually segfaults.) Thankfully, xine-lib, audacious, and mpd all use libFLAC, but mplayer uses ffmpeg's flac decoder. And it was me just going to preview a 1.2.1 encoded flac with mplayer that I noticed the problem.

Yes. Either that or compiling the FLAC framework yourself, for use in Cog

Thanks Josh,

Another mingw32 compile note: make fails due to this conditional at the top of \include\share\alloc.h, which prevents stdint.h from being #included, leaving SIZE_MAX undefined.



-Bahman

IMPORTANT NOTE:

some people have had problems with 24bit files encoded with flac-1.2.1. what is happening is that the encoder is taking advantage of 1.2.x features to encode the 24bit file more efficiently. this was not supposed to be released yet to give s/w more time to update to a 1.2.x decoder, but was picked up by a CVS merge that was too broad. specifically, the files are using RICE2 partitions. 1.2.x decoders handle this fine so you can get back the original using a 1.2.x decoder if this is causing a problem.

the change to the decoder needed to support RICE2 partitions is pretty trivial; I will try and get in touch with ffmpeg about it.

Josh

I'm a little bit hesitant to include a shaky plugin but now that you've described the procedure it's at least possible for me to do it! thanks.

there was a compression improvement in 1.1.4, otherwise, no. see the FLAC changelog

thanks, fixed in CVS.

is this happening on a 24bps stream? that's the only thing I can think of. if so, read the previous important note. if not, can you host or upload a small flac file that causes a crash?

Josh

Whoa are you serious there? Should I re-encode all of my CDS with the new version then? I just finished encoding all of my CDs with the last version of FLAC!

Hmm, well, I did a quick test, and it appears that ffmpeg chokes on my flacs that have the album art in the tags. (Added using "metaflac --import-picture-from=cover.jpg") Removing the album art and resulting padding block fixes the problem.

not necessary, ripped WAVs are not affected.

I checked out the ffmpeg source, it looks like they use a 64k buffer to try and read in all the metadata and the first FLAC frame. so if you've got a big jpeg in there that method will fail. that should be fixed.

I did manage to make a 3 line patch though to fix the 1.2.x stuff which I'll be submitting shortly.

Josh

In the Partitioned Rice 2 documentation, I've noticed the "Encoding Parameter" section reads "0000-11110 : Rice parameter." when it should probably read "00000-11110 : Rice parameter." in order to cover the full five bits. I'm sure that's just a cut & paste typo, but I thought I'd mention it just the same.

@jcoalson:
Is the Rice2 coding method only in use for 24 bit files?

currently it's only enabled if the input is more than 16bps, and 24bps is the only resolution >16 that the command-line encoder supports right now, so yes. it will probably stay that way since it's extremely unlikely to be utilized in 16bps audio anyway.

Josh

Has 1.2.1's 24-bit problem been fixed yet? Foobar2k won't play 24-bit flac files I made with 1.2.1 downloaded from flac.sourceforge.net last week...