mp3: Different codecs/params -> Quality

^ even the most basic computer CD-ROM drive has error correction/detection. You just need software which utilizes it for audio extraction (see the first post for an example).

^ there is none (for CDs that aren't really badly scratched). I assume you're comparing the WAV files?

Edited by MikeEnRegalia - December 05 2006 at 09:43

^ I know what you mean. Those issues are only important for badly scratched CDs - and I agree that EAC has a reputation for being able to correct them much better than CDex, although it also heavily depends on the CD drive.

But we're talking about a level of physical damage which normally doesn't occur on your CDs ... have a look at this example:

http://wiki.hydrogenaudio.org/index.php?title=EAC_Vs_CDex_SecureMode

I must say that I don't follow the conclusion that he makes - I can't see much difference between the CDex results and the EAC results. It is possible however that for these badly scratched CDs the error correction mechanisms fail, which means that CDex says "no errors" but in fact there were some. The test was conducted more than a year ago though, and it's quite possible that the bug has been fixed already.

andu wrote: i'm the MASTER...  i rip audio cds to .wav files with EAC (Exact Audio Copy), which, because of it's corrector, makes perfect copies of the audio information (beating the hell out of any other audio ripper). then i encode the tracks to loseless formats (my favourite is .flac) and backup them on dvds which are kept safe, and for my portable mp3player i make an .mp3 version of the album; usually at variable bitrate or at 190kbps. these proceedings keep my original audio cd out of danger because of little usage but i am also covered with the loseless backup for "just in case".

I agree, EAC is the way to go. I started using this setup:

EAC v0.9 beta4
Plextor PX-760A
LAME 3.97 FINAL
V0 VBR new encoding


I like EAC because it gives outputs like this for each track:


Track  3
     Filename F:\EAC Music\Amon Düül II - 03 Jail-House-Frog.wav

     Peak level 98.8 %
     Track quality 100.0 %
     Copy CRC 5F33C1C4
     Copy OK

If the program finds errors, it outputs something like this:

Track  4
     Filename F:\EAC Music\In Flames - 04 Graveland.wav

     Suspicious position 0:01:52
     Suspicious position 0:01:54
     Suspicious position 0:02:00
     Suspicious position 0:02:21 - 0:02:22

     Peak level 97.9 %
     Track quality 94.9 %
     Copy CRC 7232B399
     Copy finished

I've gone back and listened specifically at those spots and I can't hear any pops or crackles.

Also it has error correction, so if it finds a scratch, it re-scans an area multiple times and takes some kind of statisitcs to determine what the bit is. The downside of EAC is that it takes it's sweet time sometimes to get everything just perfect; however I will say that I have never had a bad rip. Everything is completely transparent and I cannot notice a difference between the cd and a V0 mp3. I also think that it's the only ripper that uses the secure mode so that you can be confident that the data being ripped is exactly what's on your cd. All said, it might be a little overkill, but it does a hell of a job.

For anyone reading this, it's also a free program. :D





Edited by SolariS - January 08 2007 at 01:00

Well, they claim there's a secure mode. Whether or not there acutally is, I don't know. I do know that EAC is not a single pass extracter. EAC has clear advantages over audio extraction programs that only make a single pass on scratched surfaces. I agree, it's probably overkill, but at some point making mp3's ceases to be about getting noticeably better sound. I still use v0 vbr when I record even after numerous audiophiles have  independently concluded that a difference cannot be heard with the human ear between v2 and v0....and I honestly don't know why I do it. I think it's just important to record with a program and compress to a quality that you feel confident with. That's just my 2 cents. But I will agree. VBR is the best quality for the bitrate and anything below 160kbps is kinda iffy.








Edited by SolariS - January 08 2007 at 09:12

MikeEnRegalia wrote: ^ so what do you mean by "v0 VBR"? I suppose that "v0" stands for the VBR quality ... it is displayed as "q=2, q=0 etc." in CDex. But for VBR you also need to specify: Min/Max bitrate (I use 64kbps up to 320kbps) Joint stereo (great feature - saves disk space and provides a little more bandwidth in extreme situations)

yeah. v0 is essentially the same thing as q=0. i think that corresponds to a min of 128 and a max of 320, but i'm not entirely sure. i've never seen it go below 128 on q=0. i'm kind of being overkill in my encoding since i plan to buy a 320GB hard drive just for backing up my cds.

i've never used cdex. is it a multipass extracter?








Edited by SolariS - January 08 2007 at 23:32

Another good tip is using software like Adobe Audition to stitch back together two part tracks that get split by the ripping process because they have different track numbers on the CD. Then convert to mp3.

SolariS wrote: Here's an interesting article on the "secure mode" for ripping. From what I gather, they conclude that EAC and CDex do a comparable job on ripping a badly scratched cd. The only difference is that when you rip in secure mode, the program will tell you about it. http://wiki.hydrogenaudio.org/index.php?title=EAC_Vs_CDex_SecureMode

I don't agree with the conclusion he makes. This CD is so badly damaged that neither program could extract the exact data ... and if you look at the charts you don't see a big difference.

SolariS wrote: I'd still like to see results for cds that are scratched, but not quite so badly. Either way, I don't think it matters that much. I think it's just important to use an extracter that you're confident in. Unless you take poor care of your cds, your mp3's are going to sound great as long as you use an acceptable bitrate.

Agreed. I don't think that for CDs in good condition the results of EAC and CDex would differ ... maybe I'll conduct my own "experiment" to prove it.

VBR in a range from 64kbps to 320kbps means that during encoding for each frame (small part, like a fraction of a second) of the signal the encoder analyzes the audio and chooses an appropriate bitrate. If the frame only contains silence the encoder can choose the minimum bitrate and thus save a lot of space, if the frame contains a very complex signal the encoder can choose the maximum bitrate. When encoding with VBR you can usually specify a "quality factor" - high quality means that the encoder will tend to choose high bitrates, low quality means it will tend to choose lower bitrates.

Joint Stereo: For stereo signals the encoder usually processes both channels (left/right) separately. If they both contain the same audio (called "quasi-mono") or at least parts of the signals are identical, a lot of space can be saved when the encoder encodes the identical parts in a separate stream and the differences (3 streams).

To each stream (L/R or (Joint/L-Diff/R-Diff) the VBR algorithm is applied separately (so they may all be in different bitrates), and in some cases files encoded with VBR(64-320)+Joint Stereo can even sound better than files encoded in 320kbps+No-Joint-Stereo. That's because with VBR+Joint Stereo the encoder has more flexibility. in 320kbps CBR each channel is encoded in 160kbps, and if for example the left channel contains a more complex signal than the right, with VBR some bandwidth of the right channel can be used for the left channel.

You asked if choosing a low minimum bitrate affects the quality ... I would say not, at least not with the high quality setting. The encoder will only choose these low bitrates for audio that doesn't contain much data. The files that I create with 64-320,q=0,Joint-Stereo usually have an average bitrate of ~200kbps, which is not only a bit higher than 192kbps, but it also means that since usually the tracks contain many short phases of silence which are encoded with low bitrates, there are some complex parts in the file which must have been encoded with bitrates much higher than than the average ... and thus the files might even sound much better (and to my ears they do sound great).

    
i followed your instructions as you recommended but kept bitrate at 128KBPS due to limited space and now it takes more time to rip a CD and to be honest i don't quite "hear" any difference... maybe it's my PC, i don't know....