[Audio-Me]

I can see using vbr on electronic music, but for acoustic, the thought of js makes me cringe, how would it ever work? I don't think there would be any instances where both channels were the same.

 

[Joe Bloggs]

For frames where the sound on the two channels are similar, Joint Stereo encodes it thus:

L+R = sound that is common to both channel
L-R = sound that is different between left and right

It's just a more efficient way of encoding the two channels when the two channels are similar. It's not ignoring the differences, you know

 

[Joe Bloggs]

Remember to use ABX when you are doing your comparisons...

 

[blessingx]

Audio&Me, I'll ask again- why trust compression at all? Maybe it's just me, but accurate pycho-acoustic masking seems more strange then two channel similarities (especially in acoustic settings when there are sometimes less mics, etc.). And for what it's worth my music tastes of late have moved towards boarderline acoustic listening- Kate Rusby, Hem, Gillian Welch, etc. I use lame extreme and aac 320 and haven't had any problems.

xcalibur, there's some good stuff here and here for PC's (and for Macs). Keep in mind there are some differences between aac mpeg 2 III and acc mpeg 4 audio, depending on where you plan on playing them back, etc. As for 320 limits, I saw somewhere here that it can go to 512. In Apples implimentation, the highest setting is 320. Question is if that's 320 ABR (it's definitely not normal VBR) that can hit higher. Can't get a satisfactory answer to that (and my testing looks like CBR or extremely accurate ABR).

 

[clintf]

Quote:

argh..this is pretty confusing..anyway...looks like AAC development for the PC is moving along pretty well...maybe 320 AAC is the way to go...or wait..AAC doesn't have max out at 320 right?

No it doesn't have a limit of 320kbps frame sizes, but you have to think about what you're gaining by going higher. When you start going higher than 320kbps then you're starting to enter lossless compression territory and defeating to point of using lossy compression. I don't know why you think Joint Stereo is confusing (just use --alt-preset standard/extreme or insane, nuff said), but if that's confusing you're going to find AAC more confusing when you have to deal with different headers and trying to mux AAC streams into playable MP4 containers to play on portables or auiod software. You're also going to have to make a choice which encoder to use. If you're on windows you can purchase Nero's Plugin (if you happen to have Nero Burning Rom or would be willing to buy it, at least until Nero Digital is released), Psytel (basically the predecessor to Nero's plugin), or FAAC. If you're into Apple they have the Quicktime Encoder. From what I've heard the Psytel encoder is probably the best public AAC encoder, at least if you're using the well-tuned VBR profiles (-streaming, -normal, -extreme, -archive, -ultra ).

Quote:

Originally posted by Audio&Me I can see using vbr on electronic music, but for acoustic, the thought of js makes me cringe, how would it ever work? I don't think there would be any instances where both channels were the same.

5 minutes with WinABX and what you think would be a different to encode sample would be enough to reassure yourself that there isn't a problem with Joint Stereo, at least with --alt-preset standard/extreme/insane.

 

[Moguta]

Quote:

Originally posted by Audio&Me I can see using vbr on electronic music, but for acoustic, the thought of js makes me cringe, how would it ever work? I don't think there would be any instances where both channels were the same.

What makes you think electronic music doesn't have differing stereo channels? I can think of a lot of electronica that uses panning-from-side-to-side effects, and plenty have good stereo seperation. They work fine with LAME --alt-preset standard.

Contrary to what you may think, many of the Hydrogen Audio members (and most of the LAME developers) are audiophiles with equipment and ears on par with many of you. Why else do you think they'd be tweaking and improving lossy codecs, for no pay!, when nearly every casual listener thinks 128Kbps Xing-encoded MP3s sound just fine?

I think your confusion is arising from a misinterpretation of what joint stereo is. Like Joe Bloggs noted, joint stereo simply stores what is in common between the channel AND the differences. In the --alt-presets, joint stereo actually isn't used 100% of the time. From my understanding, if LAME determines the stereo seperation in any part of the song is too great for joint stereo to represent efficiently & accurately, it will use true stereo for that frame rather than joint.

The LAME developers know what they're doing.

 

[Joe Bloggs]

They have people that can tell apart LAME CBR 320 using best settings (--insane) and the original wav in ABX. And even mpc --xtreme, which is 10 times better than any mp3.