Josh passed this one along and I was going to e-mail Josh about it but I thought I’d post about it instead (don’t worry, they’re different Joshes).
At any rate, the MAD Plugin purports to offer improved audio decoding over Winamp’s default decoder. It works with Winamp 2.x at the moment, and they add that it doesn’t work with Winamp 3 yet. Their method is described like this:
Internally, MAD represents the decoded audio signal with high precision. Combined with a form of dither, this allows output samples to be calculated with less objectionable quantization noise and more audible dynamic range than most decoders produce today. The result is unusually high-quality output that is a more faithful reconstruction of the signal encoded in MPEG audio than you may have realized possible. […]
Hmm, that sounds a bit like upsampling to me (though I’m not really sure). Still, I don’t use Winamp when I need MP3/Ogg playback — I prefer Quintessential Player instead (its interface doesn’t become super-tiny at 1600x1200).
The best one I have found is Dee 2. It’s a DSP plugin instead of straight output, but is really good if your audio is already encoded highly.
Seems that MAD is available for QCD, as well.
I don’t use winamp much, though, myself. I prefer something that runs on a real operating system
*grin*
OK, there’s actually a FreeBSD ‘port’ for this, so I installed it and gave it a listen.
Then I gave it another listen.
Then I switched back to the stock decoder and gave IT a listen.
Then I scratched my head.
…
Then, I fired up the diskwriter plugin and dumped both decodings to ‘wav’ files and burned them to a CD so I can listen on my home system (a slightly tweaked Denon DCD-3300, matching Adcom pre- and power-amp, Boston Accoustics A60 speakers, Monster interconnects and speaker cable).
I’ll get back to you when I’ve made my decision. Anyway, they both sure sound good these days.
In case others are interested, I’m posting some analysis here.
Yeah, sounds like upsampling to me also. So if the upsampling algorithm is good, then when it is downsampled for output then it should sound better (if it is like the similar process with video)
Oh, and I was afraid that the reference to a real OS was actually a reference to this other OS but was plesantly surprised. Way to go Josh!
Why thank you, George. :)
After all, Power Users use The Power To Serve.
It is upsampling noiseshaping. This plugin will only benifit you fully if you have a soundcard that can decode 24 or 32 bit audio natively. This allows more accurate decoding of the file. Remember that sampling rate determines frequency responce and bit depth determines how many steps in volume there in the 0-16,24,32 sampling scale.The more bits there are, the more accuratly the amplited can be reproduce. Just my 2cents worth….