http://mp3decoders.mp3-tech.org

I was reading that website and saw that they reccomend Winamp 2.7+ or mpg123 as the mp3 decoder. However i believe it was from here or the other board that MAD was a better decoder? What's the deal actually?

the test does use older version of MAD 12.2 instead of 14.2 ....

Just want opinions I guess ... I mean i dont have the hardware to really tell the difference, just a TBSC connected to a z560 set. With Sony MDR-V6 for heaphone use, not really high end stuff, but i am just curious

I think the only real difference between the MAD decoder and the Nullsoft decoder is that MAD decodes at 24 bits and dithers down to 16, so you get a couple more bits of resolution. It can also generate 24-bit output, but that's not useful for your TBSC. (BTW, I'm also running TBSC and MDR-V6)

Not many people can hear the effects of dither vs. truncation, though (and even if they can, sometimes they prefer truncation), so that feature is a moot point for me.

I use MAD over the default Nullsoft decoder in Winamp because of MAD's auto clipping attenuation feature, and I wouldn't even need this if I bothered to MP3gain my files.

EDIT: Here's the MAD changelog that comes with v0.14.2b; it looks like many of the negative comments from the Decoder Comparison page (describing v0.12.2b) have been fixed.

huh? I dont understand that. Sorry I am quite a noob when it comes to this.

I kinda understand mp3gain ... and am thinkin about trying it but i do not knoe what and if i am gonna do.

What does MAD's auto clipping attenuation do? makes it so the sound doesnt clip when volume is high?
at least that what i get...

aiya so lost again

Please can anyone provide a url where I can find the MAD plugin, I'd like to try it.

http://google.com/search?q=mad+plugin&btnI...m+Feeling+Lucky

Hmmmm, I've used the newest MAD decoder, but the only differences are that the MAD decoder lacks ID3V2 support, decodes at 24 bits (if you have the hardware to support it) and has auto clipping attenuation. Personally, I prefer using the regular 2.81 FHG decoder since its fully compatible with streaming and other plugins, I have a 16 bit card, and I prefer using ID3V2 tags. I'm not that much of an audiophile, so I really dont mind the subtle sonic differences which most people can't really hear in the first place.

Yeah, and try to stream something with it, on my system it sends Winamp into a brutal crash that locks the whole system. MAD also uses some extra resources.

I've been using MAD 0.14.2 for about almost a year (since March) and have had no problems with it whatsoever. So I am not sure.

What decoding engine do you guys recommend for people who have a Soundblaster Audigy??

I heard sound from mpg123 is good like violin sound..
It's sharp.
MAD is heard warmly generally.

24 bit decoding is good MAD Decoder better than mpg123.

The Audigy 1 isn't true 24-bit, and will downsample MAD's output to 16-bit anyway. So stick with the default Winamp decoder.

mp3gain is the same as ReplayGain: it analyzes each MP3 and changes the volume so that all your MP3's play back at about the same loudness (it's more advanced than normalization).

I'm not sure if this is a correct description, but it makes a good analogy nonetheless: MP3's store their amplitude information as floating point numbers, multiplied by a constant. mp3Gain changes that constant, so none of the audio information is modified, but the output volume is changed.

The only reason I haven't used mp3gain is because, even though it's a lossless procedure (you can do it as many times as you want without quality degradation), unless you write down the gain adjustments made to every file, or you carefully log mp3gain, you won't be able to set the files back to their original volume after mp3gain-ing them. In contrast, Vorbis and MPC store their ReplayGain information in tags/header information, so you can decide whether or not to apply the gain simply by checking a switch in the playback program.


Yes, that's what MAD does. While playing back the MP3, if it detects clipping, it then lowers the gain that it will use for the rest of the song. So let's say you're playing back Metallica, and during a loud part the song goes over by, say, 0.8dB. Then MAD will lower the gain by 0.8dB for the rest of the song. If you're not listening carefully, you won't notice that your song got quieter, and in any case it's better than clipped output. But since you say you've had MAD since March, you've probably watched the MP3 properties window and seen MAD in action.

ahhh .... i get it. =) thanks

There should be no difference in quality between mpg123 and default in_mp3 (in Winamp 2.666 and newer). Default in_mp3 is based on Fraunhover decoder, and it's more optimized than mpg123 plugin. Please take a look to MPEG Audio Decoder Compliance test.

It's good that Nullsoft switched to the Fraunhofer decoder (again) and are sticking to it (hopefully no BMW crashes again ) Their previous Nitrane decoder wasn't all that good when compared with mpg123.

MAD sounds better to me

I have an Audigy 2 and I use out_asio_021 with winamp 2.81, and I've just had this setup for a couple of weeks, but so far I think that the Shibatch in_mpg123_118ot33 sounds better than the mad-plugin, it produces a clearer sound and is not as muffled as the mad-plugins sound.

BTW, I have Logitech Z-560 speakers.