Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: CD Offset Questions and Confusion (Read 3997 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

CD Offset Questions and Confusion

Hello,

I am having trouble fully understanding some issues regarding CD offsets.  As a "casual" EAC user, I initially believed offsets were a mere hardware flaw that, once resolved, would allow bit-for-bit identical rips of commercial "silver disc" audio CD's (provided other aspects of EAC were properly configured).  The foundation of this belief, at least on my part, was grounded in the assumption that all "silver disc" audio CD's were produced identically.  Meaning, I believed CD audio specifications ensured that all audio CD's were manufactured according to the same exact physical guidelines.  Hence, for instance, the pregap of any two "silver disc" audio CD's would be 100% identical (in position, length, etc.).  If this were true, it would logically follow that the only offset variable is that of the CD hardware.

However, after reviewing the HA FAQ threads on the subject, it is now my understanding that there are in fact two offset variables: the CD hardware and the CD itself.  Now, if I am not mistaken, the CD audio specification does not define, to within a degree of accuracy of a single bit of data, the end position of the pregap and start position of the audio.  Therefore, it seems highly unlikely that any two "silver disc" audio CD's will have an identical "internal" offset (unless, perhaps, they were manufactured in the same batch).  So then, this leads me to ask the question... if an "exact" pregap specification does not exist for "silver disc" audio CD's, how can these CD's be appropriate for determining an external, "reference" hardware offset?  In order to determine an external, "reference" hardware offset, I believe Andre (the designer of EAC) first had to choose an "internal" CD offset.  Can someone explain how Andre determined what "internal" CD offset should be used in order to calculate the external, "reference" offset that all CD hardware would be calibrated from (and why, if at all, his choice was a good one)?

Also, please correct me if I made any mistakes in my above analysis, or contribute if there is anything one can perhaps explain in a better manner. Thanks!

*edit*

One other somewhat unrelated question: does all "accurate stream" capable hardware produce the same offset 100% of the time?

CD Offset Questions and Confusion

Reply #1 –
Yes, accurate stream capability means that the reader has a constant offset, i.e. it will always return to the exact same sector whenever you issue the same request.

The reference "offset 0" was chosen to be where apparently quite a few cd's have their beginning. That's the only reason for that particular choice iirc.

CD Offset Questions and Confusion

Reply #2 –
I'm not sure I understand your question (and, as always, Pio2001 will later post something on this thread that will be the exact opposite to the way I believe things are*  ), but here goes: I think you mean the "detect offset" function that EAC has on it's settings screen. Well, in the included EAC documentation there is a list of pressed CDs with their Batch numbers, for which offset is known to Andre, in relation to his Plextor test drive. So, if you insert one of those discs (theorethically identical to the one Andre used), EAC will do calculations and produce a correct offset value for your drive.

In my experience, none of the listed CDs match any of my CDs, so I determined my offsets "the old-fashioned way". I burnt a CD with some MP3s I converted to wav. Then I extracted the burnt tracks and compared them with the decoded tracks on my hard drive. As I know the writing offset (yup, that's another one) for my drive (drives do seem to be created equally) I did the math and got my read offset.

Now, look, I just did this because I have been using EAC since 1999 and back then there was no place like HA to help. So, I used the info on one of the sites linked from the official EAC site. It was called "Dick's Elite DAE", IIRC. This guy Dick made a big deal on how yo should calibrate your drive's offsets to get the best copy and blah, blah...
So I did. And know, everytime I reinstall EAC on my machine after an OS upgrade I still set the offsets (I know them by heart), as an old habit. But, as it has been said countless times, offsets mean sh*t. They are so small that you would probably have to extract and burn the same track hundreds of times to actually hear a difference.

So, don't worry!

*And you should listen to Pio2001 instead. He seems to be HA's resident EAC expert. I'm just filling in 
I'm the one in the picture, sitting on a giant cabbage in Mexico, circa 1978.
Reseñas de Rock en Español: www.estadogeneral.com

CD Offset Questions and Confusion

Reply #3 –
Quote
I think you mean the "detect offset" function that EAC has on it's settings screen.


Yes, I am just trying to figure out what reasoning Andre followed in order to determine the reference "internal" CD offset that all "external" CD hardware offsets would be calibrated from (the hardware offset is what we adjust in EAC). As best I can figure, to quote Patsoe's post, "The reference "offset 0" was chosen to be where apparently quite a few cd's have their beginning."  But this makes me wonder... exactly how many "silver disc" audio CD's did Andre use in order to determine the "best" or "most common" internal CD offset?

Quote
offsets mean sh*t. They are so small that you would probably have to extract and burn the same track hundreds of times to actually hear a difference.


Perhaps for most users, but I am interested in starting a live music trading circle, and want to ensure, even after hundreds of burns, the audio will be perfectly entact. But, you also can't leave out the ultra-purists that demand total perfection. 

CD Offset Questions and Confusion

Reply #4 –
Alright, so you want to calibrate your drive...who am I to tell you otherwise?

Do what I did "all those years ago", go here: Dick's CD-R Page and follow his instructions (man, what a trip! he even has those pics of the way EAC used to look like! dude, I'm getting old).

That worked for me...
I'm the one in the picture, sitting on a giant cabbage in Mexico, circa 1978.
Reseñas de Rock en Español: www.estadogeneral.com

CD Offset Questions and Confusion

Reply #5 –
Thank you for the link, but it isn't the process for determining the "external" CD hardware offset in EAC that I'm uncertain about.  It is the question of how and why the "internal" CD offset (as I have called it) or the reference "offset 0" (as Patsoe called it) was chosen for EAC (the adoption of which for Plextools, as Pio2001 noted, is indicative of its acceptance as the standard).  Furthermore, was Andre's choice for the "internal" CD offset the best choice?  I think it is important to have these questions clarified for all current and future users of the EAC offset feature (or any program that allows CD hardware offset calibration).

CD Offset Questions and Confusion

Reply #6 –
Well there would be better options imaginable I guess...

Since it seems quite a few CDs are deviating from the EAC reference, it would be much better if the ripping always included a couple 1000s of frames into lead-in and lead-out. In that way, no matter how many differently offsetted CDs you have, you'll never throw away any parts of the music.

I mean, who cares if copies are offsetted from the original? I only start to care if audio is cut-off, because in some cases you really hear it when some samples are gone. A good copy for me isn't one with all the cue indices in the same places, but one that has also the music that was in the lead-in of the original disc.



About the choice of the reference offset: I think that the firm that sells the most glass-cutting machines has determined this choice. All labels that use their glass-cutters probably have the same offset on their CDs...?

CD Offset Questions and Confusion

Reply #7 –
Is there not anyone else that can provide more insight on the matter? Andre? Pio2001?

CD Offset Questions and Confusion

Reply #8 –
How was the reference offset defined ?

As shown in http://pageperso.aol.fr/Lyonpio2001/offset.htm , Andre supposed that the audio recording would start at the beginning of track 1 and stop at the end of the last track. Looking for CDs with a constant background noise, we can see, overreading, that some of them have noise recorded into the pre gap, and some of them also have noise running on far into the lead-out. But in some of them the audio data suddenly stops near the end of the last track and is followed by pure silence.
The place where the audio data becomes null has been supposed to be the actual end of last track, and used as a reference for offset correction.
Andre found 6 CDs with the same offset, and some others with different offsets. I don't know if some offsets were matching among these other ones.

Running the same experiment, I found 3 CDs with a different matching offset (two from the same label/catatlog/factory, and one other), and none with the same. It's 12 samples after EAC's reference. So glass masters cutters, if we suppose that it's them that define the offset, can have different offsets. The CDs with audio far into the lead-out can even come from different cutters with still another offset, that we can't measure.

What about adding 1000 of "frames" (or rather samples) and both ends ?

CD engineers are doing this from the beginning of CD recording. It is even possible that it is specified in the Red Book. Eric Charton in his book "la gravure des CD" states that usually, the first track begins with 588 null samples.
Track one always starts a bit after the track marker, as well as the other tracks, and there is usually 1 or 2 seconds of silence at the end of last track. This way, the read offset being usually positive, and the write offset usually near zero, each copy has an offset a bit superior to the previous, thus track markers are a bit ahead of the music start, and at each copy generation, we loose a tiny bit of the 2 seconds silence at the end. If the average combined offset is +600 and the silence 1 second, there can be 73 generations of copies until any audio data are lost (but each track will play one second of the previous).

Is it the best choice ?

Not necessarily. The fact that tracks might need to be recorded one sector (588 samples) after the marker might mean that the last track ends actually 588 samples after the real reference, what would explain why so much drives have an offset near 588. In fact they would have a read offset near zero, and burners would have a write offset near +588 in order to insert the wanted silence before track one.
From this point of view, we should rather choose a reference 588 samples behind. The advantages would be to get readings similar to real CD players playback (if they have the same kind  of offsets as computer drives), in accordance to what the mastering engineer expected, with no risk of cutting the few first samples of the first track, if the CD is a bit offsetted, especially reading home made CDs (from analog sources), on which people don't care about inserting a safety silence at the beginning of track one. Reading these CDs with the current offset corection, you have one chance out of two to get audio losses, while without offset correction, or with a reference 588 samples behind, you usually loose something inaudible at the end.
The problem would then be to get identical copies : the write offset would need to be 588 samples behind too, which would cancel the mentioned advantages.

In conclusion, the reference should be 588 samples behind, but in this case, writing soft/hardware should take it into account for burning wavs from analog sources, but not for burning wavs from CDs, that would already include a write offset.
That's why my advice for an accurate result has always be to manually check for the start/end of the CD, and correct it if necessary (adding silence if needed).

Bit-exact copies

I think it's not unuseful to recall here that 1:1 copy doesn't make much sense under these conditions, as I stated here : http://www.digital-inn.de/showthread.php?&...27006#post27006

Do you want to keep the noise running into the lead out ?
If the CD is offsetted, do you want to let it as it is or correct it ?
If you don't want to correct it, do you want to overread in order to keep all the original data, I mean really overread, what no software allows you, unless you untune your offsets on purpose to get the drive reading on the wrong offset of the CD ?
If you want to correct it, do you want to add the advised silences if they are missing ?...