Quality of copied CD's?
May 24, 2004 at 3:02 PM Post #31 of 32

Mr.PD

Lives to Take It Outside.
Joined
Mar 14, 2002
Posts
6,581
Likes
13
Quote:

Originally Posted by jefemeister
It's better to not read and write at the same time. By seperating the process, you allow for a better read and you're also having the computer do only one thing instead of two. You get a better read because, with a program like EAC, you read the discs as many times as is needed until the data is right. If you were writing at the same time, you could run into problems. by recording to the drive first, you also get to keep the files for future use or can better arrange the songs, etc. I have a Tascam component burner being fed by a Wadia 16, and I still prefer using my computer for both quality and convenience.


That makes sense.
What if I wanted to copy the "enhanced" portion of a disc?
I burn discs for the car mostly. I sometimes use EAC, with offset correction to rip WAV files so I can make compilation CDs. BTW, my TDK CDRW drive has an offset of +6 also.

I've been using Nero for a long time, and have had no problems other than Nero's ego, which insists on putting 2 second gaps on audio CDs, and placing MP3 files in alphabetical order. It's a pain to go in and change the 2 second gap every time I want to make a disc. I quit using Nero for MP3 discs.

edit, I do have a problem with Nero. It won't write a disc that will start in my Panasonic portable. They will play, but I have to skip the first track every time to get it to start. Sometimes I can go back to the first track, sometimes I can't ever get track one to play.
confused.gif
 
May 24, 2004 at 5:41 PM Post #32 of 32

halcyon

Headphoneus Supremus
Joined
Oct 21, 2002
Posts
1,848
Likes
255
Quote:

Originally Posted by Geek
Data stored on a redbook disc is in the form of zeroes and ones. Assuming you have a program with good error checking (EAC, for instance), there should be absolutely no loss in data; just as with copying a piece of software (both are zeroes and ones, period).


Actually, data is stored on the disc as pit land transitions (optical encoding of analog waveforms) which represent bits. It's still analog data (waveform variations), that get read in and turned into bits for the cd-drive internal logic.

As for error free reading, this is quite true as long as the disc/burn is good (reflectivity, data to data jitter, hf tracking, etc).

BTW, EAC only has multiple reading (or alternatively C2 checking) for audio reading. It can try to reread data, but it cannot replace a good drive with excellent low level data retrieval and error correction capability.

Also, Plextools Pro has additional accuracy that allows it to retrieve best bytes / sector , which give it an edge over EAC on many seriously scratched discs.

Quote:

Originally Posted by Geek
There's no logical or possible reason for "small errors" to occur that affect the sound; with a digital recording, either the errors are automatically corrected by error correction firmware on the CDROM, or they cause a "skip" in which the data is considered unreadable.


That is only half the truth.

First of all, Red book discs have only two levels of error detection and error correction using various realt-time implementations for the detection of the ECC data.

It is totally possible to get random errors that are not detected by the ECC and thus not corrected.

It is also possible to get errors that exceed the error correction capability of the drive and cause changes in the audio bitstream.

Random uncorrectable errors, if sparsely populated enough usually become noise into the resulting audio stream.

However, in practise this is quite uncommon, as most errors are not purely random, but caused by tangential scratches on the disc surface.

These longer scratches may cause a sufficient enough portion of the disc to be rended unreadable (i.e. data cannot be retrieved off the disc) or uncorrectable (data is retrievable, but cannot be corrected for).

This is called a burst error.

In case like these, a situation like what you describe as a "gap" results.

However, players have various methods of dealing with unretrievable/uncorrectable bits: some use sample and hold, some mute, some use interpolation (linear) while the best resort to higher order polynomial estimation interpolation.

In effect, the drives fill in the missing samples with silence, previous samples or by guessing.

With really short burst errors the audio data can still remain audibly indistinguishable from the original if the drive has a good error concealment algorithm (conceal = disguide, not to correct).

With long/frequent enough burst errors, the sound can be audible degraded.

Quote:

It works the same as with software: WITH SOFTWARE, EVERY SINGLE BIT (0 or 1) MUST BE EITHER READABLE OR CORRECTED FOR BY THE FIRMWARE, OR SOFTWARE ON A DISC IS UNREADABLE.

This is NOT like vinyl, which has bumps, ticks, etc., and all are read because the read device is in the form of a needle.


I'm sorry, but that's understanding the issue wrong.

Errors on audio disc can cause audible degration, without the disc turning unreadable.

Also, scratches do affect audio playback of CDs, it's just that the system (due to error detection/correction) is more robust and handling them.

Quote:

Originally Posted by Geek
If all of these things occur like people say they do, then why do copies of software work perfectly or not at all?


Because Audio = Red book and software = Yellow book (Orange book for CD-R/W).

They are different standards with different data formats, error correction data and synchronisation bits.

Audio (red book) is meant to be fault tolerant AUDIBLY. It recovers (in theory) gracefully from even bad data loss that losts bits irrecoverably.

This is not acceptable in software applications where all bits must remain the same at all times. It's either detect all bytes accurately or don't detect and flag an error. That's why Yellow book has more error correction data, the discs are authored in CD-R drives in different mode and are read with different error tolerance margins.

You can't feed "almost there" data to a cpu, while you can feed "almost there" audio to an amplifier.

A compact audio discs doesn't have to be 100% error free in reading - this is defined in the standards.

Quote:

Originally Posted by Geek
Simply put, I can't see a way for it to happen.


But it does
smily_headphones1.gif


I recommend the following: Principle of Digital Audio (Pohlman) and Digital Audio Technology (Sony). Both have a detailed presentation about why.

If you just want a quick look on the issue, the following web pages illustrates the issue as well (to a lesser degree):

http://www.samsungodd.com/eng/Inform...no=8&type_no=3

cheers,
halcyon
 

Users who are viewing this thread

Top