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: TAK 2.0.0 (Read 155096 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

TAK 2.0.0

Reply #25
- Put any remaining encoder option into MAX. But always keep an eye on the speed of -p4m because many users are judging TAK's performance by its strongest preset.


So at some point we might see p3m being stronger than p4m?

TAK 2.0.0

Reply #26
I'm converting to TAK with foobar2000 using parameters -e -md5 -ihs -Pmax - %d
But when I try to test with
takc.exe -t
or
takc.exe -t -md5
I get something like
196* Wave meta data missing
What's the problem and how this can be solved?

TAK 2.0.0

Reply #27
- Put any remaining encoder option into MAX. But always keep an eye on the speed of -p4m because many users are judging TAK's performance by its strongest preset.


So at some point we might see p3m being stronger than p4m?

Shush... That's totally secret! 

I'm converting to TAK with foobar2000 using parameters -e -md5 -ihs -Pmax - %d
But when I try to test with
takc.exe -t
or
takc.exe -t -md5
I get something like
196* Wave meta data missing
What's the problem and how this can be solved?

It's no problem. By default TAK saves the header of a wave file in the compressed file. On decoding it will restore this original wave file header. Basically this would not be necessary: TAK could create a new wave file header when decoding. Unfortunately there are several ways to construct a valid header for wave file containing identical audio data. So the header cretaed by TAK could be different from the original header and a binary compare of the decoded file with the source file could fail, although the audio data is identical!

If you use foobar for pipe encoding, you have to specify the -ihs option: (I)gnore (H)eader (S)ize. That's because foobar will sent am incomplete wave file header, that is lacking the size of the file. With no valid header available, TAK can't store the header data and then has to create an own header on decoding. The message will tell you about it. As i wrote above, the audio data itself isn't affected by all this.

TAK 2.0.0

Reply #28
Is there a guide or set of instructions for re-encoding older TAK files?

TAK 2.0.0

Reply #29
Code: [Select]
takc -e -p4e -md5 -wm0 "3613 - Gloria Rv588, 11. Coro- Cum Sancto Spiritu.wav"
3613 - Glo...Cum Sancto Spiritu.wav .......... Wave file not supported

1 files failed.

This time it's serious.
How can I know what's wrong and what can I do with it?

TAK 2.0.0

Reply #30
Quote
Sample rates from 8000 to 192000 Hz. Bit depths of 8, 16 or 24. Channels: Mono or Stereo.


What format does this file have?

TAK 2.0.0

Reply #31
Just a regular redbook, 44100, 16, stereo. Other tracks from this CD work flawlessly.
ADDED:
I noticed that if I decode it with flac, then tak doesn't work. But if I use bass library (via Total Commander plugin), it does.
I compared both files, there are differences in 2 places, header and the end. bass generates a file that's round 8 KB longer(13303852 vs. 13312044 B), with additional data looking random to me.
The differences in the header:
FLAC:

bass:

TAK 2.0.0

Reply #32
Thanks for the screen shots.

The 4 bytes following the RIFF id tell, how long the RIFF data will be.

For FLAC:  33,135,012 bytes !!!
For bass: 13,312,036 bytes

Since your FLAC decoded file is only 13303852 bytes long, the header is obviously invalid.

TAK 2.0.0

Reply #33
Thanks for the screen shots.

The 4 bytes following the RIFF id tell, how long the RIFF data will be.

For FLAC:  33,135,012 bytes !!!
For bass: 13,312,036 bytes

Since your FLAC decoded file is only 13303852 bytes long, the header is obviously invalid.


Thanks for the answer. So what should I write to have a correct wav w/out the extra data? I guess I mix something with endianess (because I barely understand it), but I can't sum / convert the numbers to get  33,135,012.

Also, I suggest making TAK more error resilient. Though the wav is incorrect, it's not really a problem to encode it.

TAK 2.0.0

Reply #34
Also, I suggest making TAK more error resilient. Though the wav is incorrect, it's not really a problem to encode it.


Well, generally this is not TAK's problem. As TBeck wrote:

Since your FLAC decoded file is only 13303852 bytes long, the header is obviously invalid.


I also encode with TAK sandwich with raw adpcm/pcm & ogg. This was done without any problems. And decoding too.


TAK 2.0.0

Reply #35
Also, I suggest making TAK more error resilient. Though the wav is incorrect, it's not really a problem to encode it.


Well, generally this is not TAK's problem. As TBeck wrote:

Since your FLAC decoded file is only 13303852 bytes long, the header is obviously invalid.


I also encode with TAK sandwich with raw adpcm/pcm & ogg. This was done without any problems. And decoding too.




I don't call it a problem either. Just a misfeature. It deals with errors in a correct, but the least welcome way.

BTW, I noticed that the audio track is cut. Scheduled for reripping.

TAK 2.0.0

Reply #36
I don't call it a problem either. Just a misfeature. It deals with errors in a correct, but the least welcome way.

...so TAK should silently produce incorrect files?

TAK 2.0.0

Reply #37
I don't call it a problem either. Just a misfeature. It deals with errors in a correct, but the least welcome way.

...so TAK should silently produce incorrect files?



Produce? Definitely no. It should correct wav errors when it's possible and just give a warning about decompressibility, if bit-bit correctness isn't preserved.
In the command line that I used there's -wm0 switch, so I made it clear that I don't care about metadata preservation anyway.

TAK 2.0.0

Reply #38
Looking through the TAK threads, I didn't notice ReplayGain being mentioned. So TAK doesn't support it? Or did I miss it somewhere?

TAK 2.0.0

Reply #39
Replay Gain is supported. RG values are written into tags and work like any other apev2 tags.

TAK 2.0.0

Reply #40
But does the TAK encoder actually support it? I don't see RPG as an option on the command line.

TAK 2.0.0

Reply #41
The encoder doesn't calculate any RG values. You need another app like foobar for that.

TAK 2.0.0

Reply #42
Thanks!


TAK 2.0.0

Reply #44
Hello,

I was interested if TAK has any downsides against FLAC.
Except that I'd appreciate more plugins, at least CoolEdit filter and WhereIsIt tag reader plugin.
Thanks for TAK.

TAK 2.0.0

Reply #45
Hello,

I was interested if TAK has any downsides against FLAC.
Except that I'd appreciate more plugins, at least CoolEdit filter and WhereIsIt tag reader plugin.
Thanks for TAK.

No support for multichannel music.


TAK 2.0.0

Reply #47
Bug: Takc fails if there's any Unicode character in path.
Code: [Select]
E:\music\Jean Michel Jarre\1975-03-30 Les Granges brulées>"D:\TC\lib\packers\tak
c.exe" -e 14.wav "1.tak
14.wav                              .......... Error reading source

1 files failed.

TAK 2.0.0

Reply #48
Bug: Takc fails if there's any Unicode character in path.
Code: [Select]
E:\music\Jean Michel Jarre\1975-03-30 Les Granges brulées>"D:\TC\lib\packers\tak
c.exe" -e 14.wav "1.tak
14.wav                              .......... Error reading source

1 files failed.


I just finished recompressing most of my library from flac -8 (sometimes stronger settings used) to tak -pMax -wp0 -md5.
Music styles are highly variable, there's practically everything.
Savings:
~3.4%.

Very good.

TAK 2.0.0

Reply #49
A noob question: What could be a commandline for encoding files to tak using foobar2000.
I tried:
Code: [Select]
-e -md5 -ihs -p4m - %d
it gives me a window  showing me the encoding progress (very quickly) and then another saying "the file was not found" and really no file is created.