EAC: does the quality of the drive matter?

[Psyt3k]

I've been reading different opinions on this one. Does the quality of the disc drive influence the quality of your rips? I've read that the type of drive doesn't matter for the quality of the rip, since it is digital 1 to 1 conversion. Others do promote better / more expensive drives for ripping.

 

I just tried to rip a cd with EAC on one computer, it had to error correct the first track for + 1h and it was still only at 30% for that track. I just inserted it in my other pc with another disc drive (of the same price) and it rips the track in 3 minutes. So apparently the type of drive can affect the speed of error correction, or just being able to rip a track or not. But could the first drive with the very slow error correction lead to a rip of less quality?

[Purple Sunray]

It does.

Worked for company doing CD/DVD recording software for some years and had todo with a lot of different drive manufactures, and it's like everywhere... You can't expect a drive, where the pickup hardware costs 0.2$ to work same like the one with the 5$ laser. There must be something why one is soo cheap and the other isn't.. not saying that the quality of the hardware has anything to do with price you pay for it on the store.

"Losing quality" however is not really possible. Either the drive can read the bits, can't but is able restore it via ECC, or it can't. The restore via ECC (error correction code) is lossless, if it was successful you get original value, no down-sampled or whatever modified version of the sample, so there is no quality loss, just needs more time during ripping.

If you don't get any data, the problem solving is up the ripping software and most just stop the ripping, as you can't really restore the missing pieces. You could do any kind of fancy interpolation stuff if there are only some single samples missing, but usually it has just no sense to continue if you run into reading a corrupted block unless you just skip it (than you rip will sound like playing a CD on the old car-cd-player without anti-shock while driving offroad =) )

[Psyt3k]

It does.

Worked for company doing CD/DVD recording software for some years and had todo with a lot of different drive manufactures, and it's like everywhere... You can't expect a drive, where the pickup hardware costs 0.2$ to work same like the one with the 5$ laser. There must be something why one is soo cheap and the other isn't.. not saying that the quality of the hardware has anything to do with price you pay for it on the store.

"Losing quality" however is not really possible. Either the drive can read the bits, can't but is able restore it via ECC, or it can't. The restore via ECC (error correction code) is lossless, if it was successful you get original value, no down-sampled or whatever modified version of the sample, so there is no quality loss, just needs more time during ripping.

If you don't get any data, the problem solving is up the ripping software and most just stop the ripping, as you can't really restore the missing pieces. You could do any kind of fancy interpolation stuff if there are only some single samples missing, but usually it has just no sense to continue if you run into reading a corrupted block unless you just skip it (than you rip will sound like playing a CD on the old car-cd-player without anti-shock while driving offroad =) )

Thanks for your reply! I don't know much about this stuff, but I think I sense a contradiction in your answer. As you said: "The drive can read the bits or it can't" and "you won't get a down-sampled or modified version". This means that the rip will either be correct or it will not be ripped at all. Right? So the quality of the ripped wave file with EAC should be correct than?

While using EAC you have to turn off the C2 error correction of the drive itself, so the program can do it instead.

 

So actually you don't need a more expensive drive in order to make perfect lossless rips with EAC?

 

Any other advice welcome!

 

 

 

 

 

 

[Purple Sunray]

EAC is exact audio copy, or? If so, you should know why you need to turn C2 off if I tell you that C2 is the "fancy interpolation" mentioned above

On ripping there are 3 parts involved: the CD, the drive and the software. On the CD there is your audio data and stuff like CRC checksums to verify if read data is correct and there might ECC to restore lost data.

The drive reads the disc. Good drives might just be able to correctly read everything and EAC makes an exact 1:1 copy of what's on your CD. But there might be also stuff on the disc that can't be read correctly by the drive. In this case there are a couple of things going to still get this data. First the drive will try to read sector again, and again and again (what slows down ripping of such CDs too) until it decides to give up. Next is to try to restore the data using ECC, if any.

If this was not possible, there is no more way to restore the data and EAC will tell you if it found such samples and where. At least it did in the past.

C2 now comes on top of that (runs on the drive). It looks at samples before and after the missing one and interpolates. It make a guess what this value could be and reports it up to the ripping application. That's also why you need to turn it off when using EAC. If it would be enabled, EAC would have no chance in detecting corrupt data, the drive will just do C2 and report some interpolated values instead of the real one.

So no, with EAC you should have no quality loss, as it either reads what's on the disc, or doesn't - as you wrote, C2 is off, so no guessing about what it could be. Instead it will try to read the sector multiple times, what makes ripping slow, but if one of those read operations was successful you have the original data.

[Purple Sunray]

So actually you don't need a more expensive drive in order to make perfect lossless rips with EAC?

As said.. more expensive doesn't always mean better quality, worst case you get the exactly same hardware again, but more expensive because re-labled by a more popular brand^^ but can't really give you a tip, I'm out of that business since quite some time

 

But you only need a better drive if either EAC reports read errors, or you don't have the time to wait for hours because EAC needs to read every sector 100 times until it succeeded^^

[Psyt3k]

EAC is indeed Exact Audio Copy. Alright, that's a clear answer and actually also what I wanted to hear! Good to hear it coming from someone who worked for a company for recording cd/dvd software. For me it doesn't matter if it takes longer, or if it just doesn't work on severely damaged cd's. As long as the rip is perfect once it's done!

 

Other experiences / advice still welcome!

[Purple Sunray]

> Good to hear it coming from someone who worked for a company for recording cd/dvd software

 

Ey man.. i coded that shit for years.. still having nightmares if I hear about C2 n'stuff

[Psyt3k]

Another question: Does the read capacity of a drive go down after long usage?

 

I am planning to rip my 800 cd's with 2 pc's. So each drive will have to read 400 cd's to rip, whilst being in "normal use" since 2 - 3 years. Can the read quality of a drive go down after long usage in a way that it would affect the rip?

 

If it only affects speed, I wouldn't mind, I just don't want to realize my rips aren't 100% exact after having ripped hundreds of them...

[Padremac]

I'm not sure if EAC rips are always 100% accurate. I remember I once was ripping some old, in poor condition CD and EAC was hanging for couple of hours without any progress. I found there was an option to enable "burst mode" in which CD was ripped in a minute, but I guess it was not accurate rip (though I couldn't hear any discrepancies).

[needle ninja]

What the above post meant by "Either the drive can read the bits, can't but is able restore it via ECC, or it can't." is that the drive will either read the wave or it will fail to read the wave and you will get a section of blank audio. I used to have this happen. I would have read errors and open the wave in Sound Forge to see that the wave had breaks in it. It is a problem you cannot fix, you have to re-read the file off the disk.

ECC is something you don't want to have kick in. You want the wave that is on the disk read perfectly the first time.

[Psyt3k]

What the above post meant by "Either the drive can read the bits, can't but is able restore it via ECC, or it can't." is that the drive will either read the wave or it will fail to read the wave and you will get a section of blank audio. I used to have this happen. I would have read errors and open the wave in Sound Forge to see that the wave had breaks in it. It is a problem you cannot fix, you have to re-read the file off the disk.

ECC is something you don't want to have kick in. You want the wave that is on the disk read perfectly the first time.

I let EAC disable the C2 error correction of my drive, so the program can do the correction itself. I guess this C2 error correction is the same sort as the ECC that you mention? So if that is disabled, the program can see for itself if it can rip it or not. If it encounters a read or sync error, it is mentioned immediately + in the log file. So I assume that once the program says "no errors occured" and "all tracks are accurately ripped", it will be 100% so? It also compairs the tracks you have ripped with the tracks in the Accuraterip database.

 

Are you sure you were using EAC as I described here? Thanks about the Soundforge tip btw, I will test my tracks with that program too. Was using Foobar until now.

[needle ninja]

I always just used the original sound to compare the ripped one to. Sounds like you have a fancy system.

[Purple Sunray]

I guess this C2 error correction is the same sort as the ECC that you mention?

No.

ECC relies on additional information to restore lost data, C2 does a guess what lost data could be. Not going into mathematics here, so very simple example: you have 3 signal values: 10,20 and 30.

Value 2 can't be read anymore from the CD as it is broken. With ECC there are some more bits on the disc than just this 3 values - using this bits and some doing some math on with will allow you to find out that value 2 was 20. If it was possible to run this math, the result will be correct. C2 works without this additional information. It looks at 10 and 30 and does a guess what that value in between could be. It will most likely report 20 too, so both were able to correct the error, but if the second value was 29 instead of 20, C2 will most likely present you a wrong value compare to the original. That's the reason why you turn off C2 when using EAC - you can't say for sure if this value is correct or not -- while you don't care much about ECC or other correction algorithms that are able to correctly restore your data (as you know that is correct )

[Paul Eye]

So C2 kicks in only when there's information the drive can't read?

[Purple Sunray]

In theory yes - in practice nobody really knows..

Bad/Unreadable sectors are not that uncommon on hardware level. Means the laser tries to read a sector on the disc and delivers some data, the firmware compares it with the CRC (checksum) and detects that the data just read is not correct. So the firmware will move the laser to same sector as before and try to read again. The number of how often the drive will try to re-read is more or less left to the manufacturer and that's the important point C2 kicks in as soon as the laser gives up.

So drive A might do 100 tries on failed reads and does C2 afterwards. Your CD is pretty bad and so it needs 24h to rip as it need reads every sector 99 times (but at the end the result is correct).

Now let's switch from geek to to the marketing guru role.. "Gogo!! buy that new drive B!! It's incredible fast! Does your 24h rips in less than 1h!!".. that's the marketing slogan. "Go do C2 right after first failed read and don't lose time with re-reads! C2 interpolated WAV's are way enough for Justin bieber" is the development task ;D (hope you get what i mean *g*)