MP3 Codec help

[Chewie]

I'm sorry if I shouldn't post this here but I've posted elsewhere without sucess.
I'm not sure if here is where I should post this question but here goes.

I've been using Cdex to rip my CDs and I was wondering about the MP3 Codec it uses. Which is the best MP3 codec and which program uses it?

Also in the settings of Cdex there are some terms with which I am unfamiliar. Can anyone explain the terms and the difference:
MPEG Version (I, II, II.5)
Mode (Stereo, J-Stero, Forced Stereo, Mono)
VBR Method (VBR Default, VBR Old, VBR New, VBR MTRH, VBR-ABR)
VBR Quaity (0 = highest? or not?)

Now this is for the LAME encoder.
I'm sure you guys will just say google it but I have and no site really explains all these. At least none I've found. Can anyone help?

[chrisharris]

I'll take a shot. The MPEG I'm not sure about...my files are encoded with MPEG I, Level 3, but that's not a setting that I remember ever setting. Seems to work fine for everybody trying to listen to them though, but I admittedly have no idea about this.

As for stereo vs. joint stereo, I use joint stereo despite a huge heaping pile of misinformation out there that says it sucks, lol. This site gives a partial explanation. http://harmsy.freeuk.com/mostync

As far as the VBR method, you don't have to use VBR at all. You should be able to use CBR. "VBR" means "variable bit rate," which means that the encoder will take parts of the file and encode it higher and it will take other parts of the tune and encode them at a lower bitrate, depending on how much sonic stuff is going on at any particular moment.

I use VBR on its highest setting for things like transferring individual tracks to people for online collaborations when using HUGE wave files is just not an option. It's great for vocal tracks, for example, where there are large sections of no audio...it'll encode those silent areas very low because it just doesn't matter how well silence is encoded, lol.

In short, VBR can save on file size, but if you use it to encode a full stereo mp3, the file's probably going to be bigger than if you just encoded using a CBR (constant bit rate) of 192kbps, which is the setting I recommend. You can get away with 160, but it's a little grainy. If you use the internet standard of 128, the cymbals get all phasey and the top end in general sounds like ass.

LAME is a good encoder. It's not my favorite, but my favorite conversion program uses it (www.dbpoweramp.com if you want to try it out...2nd link on that page). I've used CDEX in the past, and the converter from dbpoweramp is a lot faster, a lot cooler, and a lot more user friendly. One of the things it'll do is add an option to your right click menu where you can encode or tag wave files just by right clicking on them, which is nice.

At 192kbps, Fraunhofer (as opposed to LAME) wins in my book. If you go up to higher bitrates (256 or 320), I think Xing is the best encoder, even though it's probably the worst for everything below 256.

Just my opinions. Best of luck.

[sile2001]

By default CDex uses LAME, which is currently the best encoder there is. Xing is fast, but according to the majority of the opinions out there, it is the worst sounding (sorry Chris ) As to the terms, I can help with a few of them.

Mode (Stereo, J-Stero, Forced Stereo, Mono)
Courtesy of a message board:
Stereo :
In this mode, the encoder makes no use of potential similarity between the two input channels. It can, however, negotiate the bit demand between both channels, i.e. give one channel more bits if the other contains silence.

Joint Stereo :
In this mode, the encoder will make use of a correlation between both channels. The signal will be matrixed into a sum ("mid") and difference ("side") signal. For quasi-mono signals, this will give a significant gain in encoding quality. This mode does not destroy phase information like IS stereo that may be used by other encoders. This setting can be used to encode DOLBY ProLogic surround signals.

Forced Stereo :
This mode will force MS joint stereo on all frames. It's faster and it uses some special mid and side masking threshold.

Mono: Should be obvious

VBR Quaity (0 = highest? or not?)
VBR encoding uses varying bitrates from 32 up to 320, evaluating each segment of the audio for how many bits it effectively needs. The Quality setting just specifies whether it should weight it towards higher bit rate or lower bit rate. The scale is backwards, so 0 is the highest, and 10 is the lowest.

[chrisharris]

Originally Posted by sile2001

By default CDex uses LAME, which is currently the best encoder there is. Xing is fast, but according to the majority of the opinions out there, it is the worst sounding (sorry Chris )

Lol...no reason to be sorry at all. I totally agree on every setting except 256 or 320. Let your ears be your guide. I can't hear a difference between a straight binary rip to wave and an Xing at 256, but I'm kinda' thickheaded, lol.

[chrisharris]

Originally Posted by chrisharris

If you go up to higher bitrates (256 or 320), I think Xing is the best encoder, even though it's probably the worst for everything below 256.

I don't think we even disagree. Most people don't encode that high, so I can see where the vast majority of people would think Xing is a worse encoder...it usually is.

[sile2001]

Originally Posted by chrisharris

but I'm kinda' thickheaded, lol.

Hmm....must be a Texan thing, because I run in that same vein...

[Chewie]

Thanks chrisharris. That was an excellent link. Still leaves question bout the VBR and ABR unanswered. I hope someon will answer those soon. But as for Joint being better. The article admits that whie technically sound the implimentation is sometimes flawed. What are your comments on that as per each encoder?