Beta release 1 of TAK 1.0.4 ((T)om's lossless (A)udio (K)ompressor)

It consists of:

- TAK Applications 1.0.4

Winamp plugin, SDK and decoding library will come with the final release.

Download:

TAK Applications 1.0.4

containing the GUI and command line compressor:

Download link removed. TAK 1.0.4 final has been released.

What's new

New Features:

- Support for pipe decoding.

Improvements:

- Decoding speed improvements of 7 (preset 5) to 13 (preset 0) percent on my system.
- Modification of the presets -p0 and -p1 to make them even faster: -p0 is now encoding 17, -p1 10 percent faster on my system.
- I have replaced some of the Delphi libraries with own code. This will help a later translation to standard C, which can't use delphi libraries. A nice side effect: Most of the binaries are much smaller now.

Modifications:

- The read buffer used to decode presets -p0 to -p3 is considerably smaller now. In most cases this should improve the decoding speed.

Known issues:

- If you use pipe decoding and the application reading the pipe is beeing terminated before the whole file has been read, TAKC may get into an endless loop and has to be manually killed with the task manager. I don't think this is a big issue but i will try to fix it in one of the next versions. BTW: Big thanks to shnutils for testing the pipe decoding!

Beta testing

The beta version has already gone through extensive testing performed by my automatic scripts. But especially because of the many changes for 1.0.4 rare bugs are still possible (as always...). Please try the beta release and report any bugs in this thread.

I would also be happy about tests of compression efficiency and speed. Because the final release will have identical performance (there may be a speed variation of 1 to 3 percent because of different code alignment of another build), it does make sense to test the beta. Exception: If the smaller decoder read buffer does hurt performance, i will switch back to the old size.

Thanks for testing and have fun

Thomas

Your time and effort is appreciated as always. Thanks.

Consequently I've released shntool 3.0.7, which is able to decode TAK files. Thank you Thomas for your hard work!

Awesome!!!

Thanks so much to both of you guys!

Tbeck, thank you for beta release.
I noticed that the variance of bitrate per track for p0m is smaller than of FLAC 8. For example:
p0m - Min. 874 kbit/s Max 1113 kbit/s
FLAC 8 - Min 872 kbit/s Max 1117 kbit/s
For average p0m has tiny bit higher compression ratio than FLAC 8.

You are fast! Thank you very much!


Thanks for the info. Since FLAC 8 is using 12 instead of TAK -p0's 4 predictors and because especially going from 4 to 8 predictors can significantly increase the compression on some files, i am sure FLAC -8 will sometimes compress better than TAK -p0m. Possibly TAK -p1 would be a good choice for you. Decoding is only slightly slower than with -p0.

Results for 1.0.4b1 have been added to my comparison.

Thanks a lot!
It would be interesting to know what do the others developers think about TAK !
Come on guys n girls, so long...

Great Thomas

Thank you so much!

And nice results

Decoding got significantly faster and i think new -p0 is a nice addition for people (or applications) who (which) need ultra fast encoding.

Thomas

Thanks for the update. i converted a bunch of my older Tak files (v 1.01 stuff) to the newer format, and saved about 500 kb per file. Thank you very much for the hard work.

Here's to even stronger compression presets

If you have encountered any bugs, please report them now!

I would like to release TAK 1.0.4 final within the next days...

I'm not sure it's fair to call it a bug, but you can't pipe the output directly to crc1632.exe, though you can with flac (Monkey's Audio can't do it either). Luckily I'm only interested in a CRC32 of the raw PCM data so I use sox as an intermediary. For md5 hashes I pipe to md5sum, which works great!

As a workaround, you can use shntool's cat mode to send the raw PCM data of any supported format to stdout:

% shntool cat -c -e file.tak | crc1632

Of course, there is also hash mode which calculates MD5 and SHA1 hashes internally. But the above command is more flexible.

I'll have to try your workaround to see which runs faster on my system.

I'm currently using:

takc -d "%1" - | sox -t wav - -t raw - | crc1632