A comparison of MP3-decoders for MM3

[Big_Berny]

Hi guys,
there are a lot decoders out there and most people don't know which is the best. So I tried to do a little comparison of different codecs and the newest MM 3.0.3 RC9.

First about quality:
The quality of all decoders are very similar because MP3 is a standard, so there aren't different ways how to decode it. But there are some difference about the bit-depth ("detail") and noiseshaping which might affect the quality - at least theoretically! Practically there is no hearable difference under normal listening circumstances. And a higher resolution doesn't mean automatically that the sound is better!
Anyway there is a Comparison regarding quality on hydrogenaudio which shows some differences. But the tester had to use some special tricks to hear the difference so you can't just compare it to normal listening situations. And btw, the tested MAD-plugin is older than the one I mention later.

So the main difference of the codecs are other features like gapless playback:

Default MM3-decoder
The default decoder is very basic and has no special features. For VBR-songs it shows the average bitrate which cant be changed. Unfortunately the decoder doesn't decode perfectly gapless. There is no pause between the tracks but there's a little click which sometimes is very noticeable.
Functionality: *
Gapless: *

MAD 0.14.2b
This MAD-plugin which is linked on the MediaMonkey-homepage has some more features: It is able to show the average or the realtime-bitrate, supports different resolutions, clipping protection and you can set the channels to mono, stereo and others. The gap between songs is a bit better than in the default decoder: For some songs it's perfect on some songs you hear a slight click or 'jump'.
BUT: The seekbar doesn't work in MM 3.0.3 RC9.
Functionality: **
Gapless: **

MAD v2 (by MoSPDude)
This decoder has tons of options: IT has the same like the older MAD but also a lot additions like gap-removal (nothing to do with gapless-mode!) different noise-shaping-algorithms, storage of streams and much more. There are no gaps between different songs - much better than the other plugins.
BUT: Like with the other MAD-plugin the seekbar doesn't work in MM 3.0.3 RC9.
Functionality: ****
Gapless: ****

Buggy MediaMonkey 3:
Unfortunately MediaMonkey 3 is really buggy regarding MP3-Input-Plugins. While the MAD-plugins have the problem that the seekbar doesn't work - which is REALLY annoying - other plugins just don't work at all. This affects all in_mpg123-plugins I tried: Neither the one linked one the MediaMonkey-homepage nor Otachans newest plugin for Winamp 2.x/3.x were able to play songs. But also another MAD-plugin had the same problems. The songs were accessibble (not greyed) but they just skipped when you wanted to play them.

Conclusion:
At the moment there is no perfect MP3-decoder for MediaMonkey. The original one works normally but has no features, isn't perfectly gapless and some users mention problems with playback or songlength. And there are good alternatives but unfortunately most of them don't work with MM3. And these which work have the problem that the seekbar doesn't work.

I really hope that the developers are able to find the different bugs which lead to the mentioned problems. Because in older versions other plugings worked great!

Big_Berny

[jiri]

I reviewed both issues reported:

1. in_mpg123.dll - it's a problem of how this plug-in handles playback start, which isn't compatible with some thread processing in MM. There is a very simple fix directly in in_mpg123.dll, in in_mpg123.cpp change

if((OpenStatus ? ::SendMessage(hPlay2ndWindow, WM_USER, 0, 0) != 0 : false) == false) {

to

if((OpenStatus ? Play2nd() : false) == false) {

We will try to also provide the compiled dll and also to find a way of how to prevent this problem.

2. in_mad.dll - this plug-in sends one parameter that isn't compatible with out_MMDS.dll plug-in. We will provide a fix soon, until then you can use another output plug-in.

Jiri

[dypsis]

I thnk if someone could develop a BASS Input Plug-in for MM, then that would be the best path to take.

[Big_Berny]

@jiri: Thanks for your reply!

Yeah, you're right. All the mad-issues are due to the output-plugin. With the kernelstreaming-output-plugin the last MAD-plugin works and all the seekbar-issues are fixed.
So that's a good solution for me. But anyway I'm pretty sure that it worked in older versions.

About the in_mpg123.dll: Since Otachan seems not not develope this plugin any further it really would be useful to create a compatible plugin of the last version. Anyway it would be nice if the compatibility to winamp would be improved. But here again I'm sure that it worked before.
Btw there's some progress going on: http://mpg123.org/download.shtml . Unfortunately there's no input-plugin of it available yet.

[DaKo]

When will a new version of the in_mpg123.dll be available? It seems like the solution for the problem isn't too difficult.

[ReverendEntity]

I thnk if someone could develop a BASS Input Plug-in for MM, then that would be the best path to take.

That would be pretty awesome.

[gpzbc]

There is a newer version of the MAD plugin. Here is the link to the forum with instructions on how to download and set it up.
http://www.mediamonkey.com/forum/viewto ... 07&start=0

[Nebbin]

There is a newer version of the MAD plugin. Here is the link to the forum with instructions on how to download and set it up.
http://www.mediamonkey.com/forum/viewto ... 07&start=0

Have you found this v15b version of the plugin much better than v14? I tried v14 and although the seeking was fine, there were too many other niggles which made me switch back to MM default, including:

* Song now playing would skip backwards when changing any tag details;
* Evaluated a different "track volume" value for most songs I tested. (Not sure why, but if I was to continue to use this plug-in, I'd have to re-analyse many thousands of songs...)

[gpzbc]

Have you found this v15b version of the plugin much better than v14? I tried v14 and although the seeking was fine, there were too many other niggles which made me switch back to MM default, including:

* Song now playing would skip backwards when changing any tag details;
* Evaluated a different "track volume" value for most songs I tested. (Not sure why, but if I was to continue to use this plug-in, I'd have to re-analyse many thousands of songs...)

I am not having any problems with this plugin, but I don't recall having the problems you mention with v14.
Actually come to think of it, I get a little stutter in the now playing song when I edit any tag details.... but I am pretty sure I also got that in MM default.
And I don't really use the analyze track volume feature so I can't really comment on that.

[Big_Berny]

There is a newer version of the MAD plugin. Here is the link to the forum with instructions on how to download and set it up.
http://www.mediamonkey.com/forum/viewto ... 07&start=0

The MAD v2 I tested here is from the same developer but newer: http://www.hydrogenaudio.org/forums/ind ... opic=46824

[gpzbc]

The MAD v2 I tested here is from the same developer but newer: http://www.hydrogenaudio.org/forums/ind ... opic=46824

That's good to know. I'll give it a whirl.

What is the difference between the one that you mentioned and this one? http://www.hydrogenaudio.org/forums/ind ... &hl=in_mad
This one seems to have more recent development.

[gpzbc]

Never mind. i figured it out.
Thanks for the heads up!

[Nebbin]

I am not having any problems with this plugin, but I don't recall having the problems you mention with v14.
Actually come to think of it, I get a little stutter in the now playing song when I edit any tag details.... but I am pretty sure I also got that in MM default.
And I don't really use the analyze track volume feature so I can't really comment on that.

I found the current MM default plugin to have almost no glitches at all when editing the tag of "now playing" tracks. On the other hand, the skip for the MAD plugin seems to be quite dramatic in comparison... but that could just be my setup.