Click the link in this topic's first post.

Yes, I did it.. but my current version is 0.15.1(beta) which is a lot newer than the version 0.14.1b from the 1st post. There is so many updated version inside this forum.. I really confuse about which one is the best and updated.. Alt. I love MAD, but anyone tell me which one is the best MP3 decoder for winamp? I'm still confuse about it too...

MAD decode in 24-bit wavefrom
mpg123 decode in 16bit wavefrom with 24-bit output..
there is few of them too.. but I really dunno which I should use.

correct me if any of my information r wrong m I hope if there is anyone can really answer my question.. thank you!!!

I did really answer your question. The file linked to in the first post is always replaced by the most current version, which is dated 2007-01-02 12:04 at the moment.

About the MAD or mpg123 question: I wasn't ever able to hear any differences between their decoding capabilities, and neither of them caused noticeable issues concerning CPU usage, hence it's a matter of personal preference which one to use. MP3 decoders are long-lived ones these days, with a lot of optimizations having been done in the past. Additional plug-in features, like good tagging capabilities, streaming and Unicode support might be more crucial for most users to come to a decision.

Which is the latest mad plug-in ? Where can I download ?

As Junon said, the link on the first post will always be the latest version.

Here is the link again,

http://www.zen23652.zen.co.uk/in_mad.zip

I'm sorry for not having time to do many things with the plug-in now. I don't want to release my remade (very buggy) code yet, and I've not had time to back port some of the other changes. Please keep testing and posting your problems. Trust me, I'm not fully happy with the plug-in yet so its far from being dropped!

Thanks to all

How is it going with the final?

We haven't heard from you for months, have you forgotten about it or that

Hi,

Its coming along slowly, I've rewritten most of it from scratch into C++ integrating Unicode support from the start. Hopefully this won't impact older Winamp version support, but will certainly benefit the newer Winamp versions now the move to Unicode is completed. I'm also aiming to have the plug-in use the global Winamp 5 preferences. In terms of features, I'm not planning for anything new yet until I've got this new code working properly.

Cheers for still being interested

As a side point, do you think I should provide options to enable/disable ID3v1 UTF-8 read/write support, or just have it fixed enabled?

Hi,

I think that you should make these options configurable, mainly for compatibility with hardware and/or older software players.

When I unzip the mad file I get the in_mad.dll and a folder called in_mad with subfolders and loose files. I put the in_mad.dll file in winamp plugins folder and it works fine. I have been using it for a long time. My question is, what what do I do with the in_mad folder? Am I missing some functionality by not putting that folder somewhere? Could someone enlighten me?

in_mad folder is source code

It is nice to get such a quick reply, thanks. So, I will only have use for the source code if I want to tinker with how the in_mad.dll functions or works with various programs? That is not something I can do yet, but I am very pleased withe the dll as it is and will let the experts make it better if that's possible.

It's been more then a month since we've last heard from you, MoSPDude. How is the development going?

I think it would be good to provide an option, if it's not too much work.

Slowly....

Once I've got this work project finished (hopefully mid-June) and recovered properly from computer failure, I'll have a lot more time to work on the plug-in. I currently end up doing odd bits every week but not much unfortunately. I've got so much unfinished code in, that the last build I disabled the GUI just to ensure the thing did still play files back! I just hope this new structure is worth it and that the Unicode stuff does work properly out of the box.

Do I need to uninstall the Nullsoft MPEG Audio Decoder or any other default plug-ins from Winamp to use the MAD plug-in? I'm using Winamp 5.33.

Nope. In the Nullsoft decoder's configuration there's a line displaying all file extensions which should be handled by the plug-in. Simply remove MP3 from this line, enable MAD and restart Winamp. Having done so, Winamp should play MP3 through MAD, checking an MP3's properties will make sure it actually does. The new plug-in uses an own GUI for metadata editing.

I'm now using two versions of Winamp - Version 2.95 with MAD and a newer 5.x version with the standard plugins. From reading this thread it looks as though I could now update to use the latest 5.x version with the latest MAD plugin. I have actually been fairly sure that the 2.95+MAD combination provides better sound - and I'm using a 16 bit sound card (Creative MP3+) and I often use Winamp also to play MP2 files encoded from broadcasts. Whether my perception of better sound is real, or really due to improved MP3 (MP2) decoding, or possibly due to other subtle factors, such as a slightly different frequency response curve, I can't really say - though I do think it is substantially better and makes some output from DAB radios bearable.

One other interest is using Winamp with aac+ plugins, so that I can listen to stations from sites such as http://www.tuner2.com

I'm looking for any advice about the compatibility of the various plug-ins. Can I have a configuration of Winamp (the latest perhaps?) which will (a) use the MAD decoder for preference, and (b) switch to aac+ automatically where needed, or does this all have to be done manually?

There's no need for switching between anything, because the decoders don't interfere with each other in any way. MAD handles MP3, AAC in its common MP4 container is decoded by in_mp4, raw .aac (which is extremely rare) is a case for Nullsoft's MPEG audio decoder. Removing the .mp3 extension doesn't completely disable the latter, it still works for any other format found in its configuration, including AAC.

Afaik, the in_mp4 of nullsoft is just a MPEG-4 container parser, which outputs the raw aac to the in_mp3 codec.
There existed one in_mp4 from audiocoding which decoded the audio itself, but that's not the case now.

Another question. I tried the 32 < 24 decoder setting because i couldn't find anywhere if my soundcard is 16 or 24 or what bit (even in the control őanel->sound thing), and it worked, i mean it doesn't crashed, or there was no error, it plays the music. This means that my soundcard is 24 or 32 bit or what? If it was a 16 bit card than it wouldn't play the music, right? In the directsound status menu it says '32 bits per sample, 2 channels'.

Also, what is the difference between 24 bit padded to 32 and 24 bit? Is the padded one better or no difference?

24bit padded is a 32bit signal, where only 24 of them carry information. the other 8 are just zeroed in order to maintain bit aligment.
So there is no difference in quality. The hardware might require one format or the other, and if storage matters, the 24bits is smaller than the 24bit padded one.

So, I have a 16 bit soundcard, is it better or the best quality if I set Noise Shaping and Rectangular dither?

If you don't use any additional (usually non dithered) attenuation, or EQ than it's ok. That's why it's best to use the MAD plugin with kernel streaming output and without any processing in winamp. Otherwise, they're going to truncate the samples, so there's no point in dithering the output of the decoder.

i have MAD 0.15.1 beta, is that the latest? the given links doesnt say much about that.

and another question, im using MAD with XM Play, is that plugin fully functional with it as with winamp?
tnx.

Hi Ivan,

The MAD decoding library (libmad) 0.15.1b is the latest version, the library itself hasn't been updated since early 2004 (I think!). The ID3 tag library (libid3tag) is also at version 0.15.1b, however I had to patch and fix parts which can be found in the source with the plugin.

As far as I know, XM Play should if it takes Winamp input plugins work ok with this, but I've not tested it - so if you could try it and report if it works or not, that'll be great. I had been testing it with MediaMonkey and looking at problems with compatibility.

Cheers

im using MAD with XM Play for quite a long time.
I cant tell much because I didnt make any compares with Winamp.
Plugin can be configured like it should be with winamp.
When a song is played I can see ID3, APE tags and LAME tag (if the audio file is encoded with lame) through the plugin (view plugin file info), so it's obviously doing something.
But what interests me is behavior of some applied options.
Im not sure that they are really functional.
For example, in "output" tab (in plugins config) option "Enable internal EQ".
Or "Link to Winamp EQ", I suppose that nothing will happen on XM Play because plugin cant detect XMs EQ.

Winamp handles shoutcast streams horribly.

In Winamp's options, there is a place to set the 'default extension for unknown file types', the default for this is 'MP3'. If you set this to 'OGG', when you try to play a typical shoutcast stream (http;//host:8000), in_vorbis will try to decode it, and fail. Because shoutcasters don't set the stream extension, winamp has no idea what format the stream is in. The workaround for this problem if you remove 'mp3' from in_mp3's extension list is to create a new fake default extension in its place for unknown/shoutcast streams. I used "strm" for the defaut extension and added "strm" to in_mp3's extension list. Then in_mp3 will decode all shoutcast streams. If the stream has a mountpoint (http;//host:8000/stream.mp3) then the proper mp3 decoder (MAD in this case) will be used instead of in_mp3. The problem here is that in_mad has problems with icecast2's metadata

It's a mess

Could you please add option "Ignore ReplayGain info in LAME tag"?

MoSPDude, you're a damn genius, I love your work.
It is indeed the best decoder of all (until now I used the old one though O.O)!!
You changed the way I listen to music, so thanks, really!

Waiting for new updates \ news.

Cheers ^^

I can't take all credit, I only decided something had to be done with it being heavily out of date, and combined a lot of best bits from other sources with fixes. The main credit goes with the author of libmad and libid3tag. As for a "Ignore ReplayGain info in LAME tag" option, it'll be in the new plugin.

I'm currently working on the GUI section of the new plug-in, and in the interest of internationalisation I'm thinking of making the plugin read strings in UTF-8 format for the GUI from a "in_mad.lng" or "in_mad.ini" file for each language if its present. Is this a good idea? or is there a better way, rather than having to rebuild the plugin with growing resources in the DLL?

Other than that, the new C++ code is using some blocks from the old C (especially the playback) and decodes files and streams quite happily. New sections have be written to handle Unicode only from scratch and libid3tag seems ok with it at the moment. I'm hoping to have more time now to push things along.

Opinions and comments are always welcome

I'm having problems with Last.fm scrobbling system, is the actual MAD plugin fully compatible with it and it's my problem or it's not (and maybe you can manage to add compatibility with it)?

Greetz ^^

The plug-in should be compatible, though I've not tested if Unicode characters work with it. I use last.fm now regularly and seem to have no problem with scrobbling. There was a problem with it a while ago, but that was solved. What kind of problem is occuring, and what winamp and last.fm version are you using?

Solved, I was using the last vv of both Last.FM and MAD, but there was something very wrong with my hardware, actually, guess my motherboard is ** up.

Don't know exactly how it managed to compromise scrobbling, but, e.g. I notice d that "something" was checking the Floppy each time a song was played.
And somehow this + other random things [read: the ghost in the machine] prevented from scrobbling -.-

I know that with this post your consideration of me may be nuked, but I swear I have no way for explaining how my PC escapes from my control, lol

MoSPDude wouldt there be any way of fixing that annoying OK promth that comes up when winamp drops connection to a stream, so i dont have to click ok.

I just want it to skip to next number without me having to click ok, since there is really no need for me to click ok to it :-)

The OK box just says the following:

Error Reading Data
An unknown error occurred.

and then there is an clickbox with OK in it thats all

Great work you are doing on this plugin, keep it up ;-)

is there a way to synchronize bitrate display with current position in file? on my system bitrate display is delayed by the length of output buffer (i'm using dsound output plugin).

for example if you set buffer length to 20 seconds (max) then winamp will show bitrate of current position + 20 sec in file. it's very annoying to me.

I'll remove that OK prompt completely, and leave a notice on the Statistics page of the stream info box with a reason for the stream error. As for the bitrate display at current output position I'll have a look into it.

I'm making a change to the advanced ID3v2 tag editor. Rather than have the fields for the frame appear in one Edit box, I'm hoping for it to open a dialog box with a list of the fields and one Edit box for changes to be made per field. I'm also testing having multiple file info boxes open at once, and allowing Winamp Playlist Editor to continue when one is open.

Its getting near to being complete, its just time consuming 'wiring' up these bits of code to do individual things that the old plug-in code did.

Thanks for your interest still.

Your the best

I've just recently stumbled on this plugin, so far I really like it.

I have tried to look through this entire post and I don't see any reference to someone requesting the ReplayGain options below.

I use Rockbox on my IAudio X5L mp3 player. Rockbox has the awesome option below for applying ReplayGain settings:

Source -> Rockbox Manual (see Section 7.1 Playback -> ReplayGain.)

Is that a possible addition to the in_mad plugin? I don't think it is possible with the Winamp ReplayGain plugin either. I've looked in foobar2000 as well, couldn't figure it out there either.

I've been using Mp3Gain for years to manage my ReplayGain values, but I would really like to fully switch over to LAME ReplayGain tags and undo/then remove the Mp3Gain APE tags from my files.

Any thoughts?

IIRC, the Lame header does not provide for album gain.

True. Adding ReplayGain values to ID3v2 tags is a much better choice using foobar's scanner. They can hold both track and album gain infos and are much more widely supported than LAME's internal solution. Besides, this course of action wouldn't implicitly require undoing the MP3Gain modifications, given the fact that album gain had been applied to the files before. If all of them were equally loud, due to having applied track gain to the data, then scanning the files without undoing the changes first would be pointless, of course.

Maybe I'm just confused about the "other" ReplayGain implementations. I know that Mp3Gain writes APE tags to the mp3's. I thought Foobar/Winamp used the LAME ID3v2 tags to store the ReplayGain analysis info.

When I view the tag details via in_mad of a Foobar/Winamp processed file the ReplayGain info shows up in the LAME ID3v2 Tab under the fields below:
Type: TXXX
1: ISO 8859
2: replaygain_album_gain / replaygain_album_peek / mp3gain_undo

Viewing the same files in Mp3Gain shows nothing, as expected.


Either way, I would still like to see if MoSPDude can comment on whether or not "Track Gain if Shuffling" option is possible.

The Lame header is a different animal from ID3v2 tags.

Ok, . Thanks for straightening me out on the terminology.

It should be possible to add an option so that Track ReplayGain is applied when shuffle is on, and Album ReplayGain when shuffle is off. I think I'll put it in the drop down list next to the enable checkbox. I'm also adding a "Ignore ReplayGain info in LAME tag" option from a previous request by nemoW.

I'm sidetracking myself from getting tag writing running, and looking into improving the ID3v2.3 tag support by playing with the Winamp 5.3 global configuration. I think I might omit the graphical EQU2 editor and work on improving it as well. Unicode is both a blessing and a curse .

I'd avoid the term LAME tag. It's more properly described as a header.

This might help stem future misconceptions.

awesome news, glad to hear your considering the option.

While you're at it, you could also have a look at a possible bug in the ID3v2.3 handling of the "Year" tag. TYER is completely ignored by Winamp 5.35's Advanced Title Formatting. I just tested this line...

... in the settings, which should look like this:

I noticed that the [2006] is only displayed if a ID3v1 "Year" or APEv2 "Date" tag exists. With nothing but ID3v2.3 tags found in a file, Winamp leaves an empty space instead.

Besides, some more extensive research revealed that the corresponding "Recording time" TDRC information in ID3v2.4 works well, this issue obviously only relates to ID3v2.3's TYER type.

Update: Adding TDRC to ID3v2.3 works too. Problem is, any application I use writes TYER tags instead. The whole lot of different ID3 standards is a serious mess.

Yes, there are some issues with TYER frame handiling.
I've tagged a file with mp3tag 2.39 and built-in WinAmp 5.35 tag editor using UTF-16 ID3v2.3, writing only "Year" info.
In both cases, mad plugin showed 2 frames in that file: ZOBS (obsolete frame) with values TYER and the actual year a wrote in the tag and TDRC frame with the ISO encoding and year I entered.

BTW, TDRC is not defined in ID3v2.3. So it looks like a bug.

EDIT: Ah, it seems that everything is fine. There is one option in the plugin's Title/Tags tab that I've missed! "Prevent ID3v2.3 to v2.4 conversion" prevents libid3tag from upgrading v2.3 to v2.4. When checked, the plugin shows only TYER tag.

This happens with ID3v2.3 tags, when "Enable ID3v2.3 tag writing" and "Prevent ID3v2.3 to v2.4 conversion" are unckecked. Checking them causes the TYER frame to be displayed correctly inside the advanced ID3v2 editor, but Winamp still fails at reading it.

I'm not sure, but maybe Advanced Title Formating works only with built-in MPEG decoder?

Have you tried to use the title formating in the mad plugin? Although it has only if-then-else.


EDIT: Yes, I've just reproduced all the issues you mentioned. There's something wrong with v2.3 tag handling...