Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[scthom]

I posted new versions of both the Wavpack and FLAC decoders on the http://mcplugins.sourceforge.net page.

The Wavpack decoder upgrades to Wavpack 4.41 and fixes a memory leak with cover art.

The FLAC decoder adds an option to not Remove Empty Tags.

[GHammer]

Thank you Scot!
Both are working well here in initial use.

[Alex B]

Thanks!

However, the Wavpack plugins may have some problems. I converted three short Wavpack test samples as follows: source Wavpack > Monkey's Audio > new Wavpack

For checking the results I used foobars' Bit-Compare Tracks and Verify Integrity tools. The following tests failed:

Bit compare: source Wavpack vs. converted Monkey's Audio

Comparing:
"U:\test\LCD_SS\Get _Innocuous.wv"
"C:\test\mc12\Get _Innocuous.ape"
Comparing failed (length mismatch : 0:17.033900 vs 0:16.033900, 751195 vs 707095 samples).

Comparing:
"U:\test\Klaus Schulze\Return_Of_The_Tempel.wv"
"C:\test\mc12\Return_Of_The_Tempel.ape"
Comparing failed (length mismatch : 0:30.000000 vs 0:29.000000, 1323000 vs 1278900 samples).

Comparing:
"U:\test\Klaus Schulze\Wild_And_Blue.wv"
"C:\test\mc12\Wild_And_Blue.ape"
Comparing failed (length mismatch : 0:30.000159 vs 0:29.000159, 1323007 vs 1278907 samples).

Verify Integrity: new Wavpack

Item: "C:\test\mc12\Get _Innocuous.wv"
Warning: Reported length is inaccurate : 0:24.049002 vs 0:16.033900 decoded

Item: "C:\test\mc12\Return_Of_The_Tempel.wv"
Warning: Reported length is inaccurate : 0:43.500000 vs 0:29.000000 decoded

Item: "C:\test\mc12\Wild_And_Blue.wv"
Warning: Reported length is inaccurate : 0:43.500000 vs 0:29.000159 decoded

All items decoded successfully.
3 items decoded with minor problems.

List of decodable but problematic items:
"C:\test\mc12\Get _Innocuous.wv"
"C:\test\mc12\Return_Of_The_Tempel.wv"
"C:\test\mc12\Wild_And_Blue.wv"

After this a tried a couple of standard ripped audio CD discs with similar results.

[scthom]

Not quite clear from your message if I'm short or long.  Do they play OK?  Does wvunpack say anything?

I'll not be able to work on it until next week/weekend.  Conference in Toronto starting on Sunday (well, travel on Sunday).

[GHammer]

Hi Scot,

I spoke too soon it seems.
I am unable to import wv files successfully.
Manual import or Auto, no tags are found.
These are Wavpack with no correction files.
I have both WV and Wavpack selected in MC's options for associations and for import/auto import.
The files are found by MC, but no tags are being read.


While looking at various settings trying to solve this, I took a look in the plugin options and see that Remove Empty Tags is selected but grayed out. Should it be?

I am moving to the US in 2 days. I'll check in again when I get there and get a computer set up.

[scthom]

In my testing, the encoder rips seem to be fine.  Conversions seem to be problematic.  So I think I can track that down.  I did a wv->ape->wv to see and the first wv (ripped) is fine but the 2nd is not.  Could be either decoder or encoder plugin.  It plays fine at the beginning but the durations and # of bytes is messed up.

Imports and tagging are fine here, so ?

The Remove Empty Tags is disabled because it's not standard APE functionality and I'm going to have to work on it quite a bit to make it work.  I'm debating about removing it.

[Alex B]

Not quite clear from your message if I'm short or long.  Do they play OK?  Does wvunpack say anything?

I started with "good" wavpack files that were created outside MC.

Wv to ape conversion with MC12 produced slightly short ape files. Otherwise these ape files were correct.
For checking the Monkey's Audio encoding part separately I tried wav to ape conversion. This had no problems.

Ape to wv conversion produced files that can be decoded with correct pcm data, but the reported file lenghts were incorrect in foobar. (e.g. 43.5 s instead of 29 s)  "wvunpack -v" reports this:

E:\X\Wavpack\wp44>wvunpack -v *.wv

 WVUNPACK  Hybrid Lossless Audio Decompressor  Win32 Version 4.41.0
 Copyright (c) 1998 - 2007 Conifer Software.  All Rights Reserved.


Get _Innocuous.wv:
file is missing 353466 samples!

Return_Of_The_Tempel.wv:
file is missing 639450 samples!

Wild_And_Blue.wv:
file is missing 639443 samples!

 **** warning: errors occurred in 3 of 3 files! ****

The original source wv files are fine:

E:\X\Wavpack\wp44>wvunpack -v *.wv

 WVUNPACK  Hybrid Lossless Audio Decompressor  Win32 Version 4.41.0
 Copyright (c) 1998 - 2007 Conifer Software.  All Rights Reserved.


Get _Innocuous.wv:
verified Get _Innocuous.wv in 0.24 secs (lossless, 29.33%)

Return_Of_The_Tempel.wv:
verified Return_Of_The_Tempel.wv in 0.36 secs (lossless, 32.71%)

Wild_And_Blue.wv:
verified Wild_And_Blue.wv in 0.34 secs (lossless, 37.27%)

 **** 3 files successfully processed ****

I used the default Wavpack (lossless, normal, 1) and Monkey's Audio (normal) compression settings in my tests.

[scthom]

Yeah, that verifies what I saw too.  Short some bytes/samples.  Maybe missing a flush on the stream somewhere.

As for not importing, I'm thinking about a tag giving me trouble, or ID3v1 or something like that getting in the way.

[Alex B]

Yeah, that verifies what I saw too.  Short some bytes/samples.  Maybe missing a flush on the stream somewhere.

Actually the files converted from wv to ape seem to be exactly 1 s shorter (= 44100 samples).

My bit-compare test had a couple of files in wrong order.
Here are the correct results (I fixed the previous post too):

Bit compare: source Wavpack vs. converted Monkey's Audio

Comparing:
"U:\test\LCD_SS\Get _Innocuous.wv"
"C:\test\mc12\Get _Innocuous.ape"
Comparing failed (length mismatch : 0:17.033900 vs 0:16.033900, 751195 vs 707095 samples).

Comparing:
"U:\test\Klaus Schulze\Return_Of_The_Tempel.wv"
"C:\test\mc12\Return_Of_The_Tempel.ape"
Comparing failed (length mismatch : 0:30.000000 vs 0:29.000000, 1323000 vs 1278900 samples).

Comparing:
"U:\test\Klaus Schulze\Wild_And_Blue.wv"
"C:\test\mc12\Wild_And_Blue.ape"
Comparing failed (length mismatch : 0:30.000159 vs 0:29.000159, 1323007 vs 1278907 samples).

[mobyfrag]

Thanks scthom one more time

[scthom]

I just wanted to confirm the problems noted above are NOT in the encoder.  The encoder creates good files from either Encode-While-Ripping, Rip-And-Encode-Separately, and Conversions.  The decoder seems to be short 44100 bytes (0.5 second).

Still working on it.

[Alex B]

I just wanted to confirm the problems noted above are NOT in the encoder. The encoder creates good files from either Encode-While-Ripping, Rip-And-Encode-Separately, and Conversions.

Did you try wvunpack and foobar's "verify integrity" tool?

Perhaps my previous post wasn't clear. I tested conversion from correct ape and wave source files. MC12 with Enc_Wavpack.dll created incorrect wavpack files according to wvunpack and foobar2000:

Ape to wv conversion produced files that can be decoded with correct pcm data, but the reported file lenghts were incorrect in foobar. (e.g. 43.5 s instead of 29 s)  "wvunpack -v" reports this:

Code: [Select]

E:\X\Wavpack\wp44>wvunpack -v *.wv

 WVUNPACK  Hybrid Lossless Audio Decompressor  Win32 Version 4.41.0
 Copyright (c) 1998 - 2007 Conifer Software.  All Rights Reserved.


Get _Innocuous.wv:
file is missing 353466 samples!

Return_Of_The_Tempel.wv:
file is missing 639450 samples!

Wild_And_Blue.wv:
file is missing 639443 samples!

 **** warning: errors occurred in 3 of 3 files! ****

The decoder seems to be short 44100 bytes (0.5 second).

Hmm... in my test the difference was 1 second (= 44.100 samples of decoded audio):

... the files converted from wv to ape seem to be exactly 1 s shorter (= 44100 samples).

Code: [Select]

Bit compare: source Wavpack vs. converted Monkey's Audio
Comparing:
"U:\test\LCD_SS\Get _Innocuous.wv"
"C:\test\mc12\Get _Innocuous.ape"
Comparing failed (length mismatch : 0:17.033900 vs 0:16.033900, 751195 vs 707095 samples).

Comparing:
"U:\test\Klaus Schulze\Return_Of_The_Tempel.wv"
"C:\test\mc12\Return_Of_The_Tempel.ape"
Comparing failed (length mismatch : 0:30.000000 vs 0:29.000000, 1323000 vs 1278900 samples).

Comparing:
"U:\test\Klaus Schulze\Wild_And_Blue.wv"
"C:\test\mc12\Wild_And_Blue.ape"
Comparing failed (length mismatch : 0:30.000159 vs 0:29.000159, 1323007 vs 1278907 samples).

[scthom]

See post for decoder release 2.2.1.