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: What if offset is not corrected? (Read 7003 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

What if offset is not corrected?

Hi,

Although I've read alot about it, I didn't find the answer.

I just realised I've ripped quite alot of CDs without correcting the offset. EAC + AR do it automatically but I also ripped CDs using foobar without manually correcting the offset.

I must admit it's OK to my ears but what are the effects if the offset is not corrected?

Thank you.

What if offset is not corrected?

Reply #1 –
CD audio has a sampling rate of 44100 samples per second. If your drive's read offset correction value is 6 samples, then track boundaries will be shifted by 6/44100 or 0.000136 seconds compared to an offset corrected extraction.

Offset correction is most / mainly useful when doing bit-for-bit comparisons of files extracted from drives with differing offsets. So the effect of this would be increased difficulty if doing manual verification, or if using a tool which does not compensate for offset differences.

What if offset is not corrected?

Reply #2 –
From a practical standpoint, it is possible that if the track indices were close enough to the actual non-zero audio on a disc, you might wind up with an audible "pop" at the end of a song (since shifting forward would give you a tiny bit of the next song appended to the track), or at the beginning of a song (since shifting backward would give you a tiny bit of the previous song prepended to the track).

In actual practice, the effective silence between tracks is usually sufficient to accommodate all but the most severe offsets, and only gapless audio - live albums, or the second side of Abbey Road - will have any effective shift in the presentation... and even then the shift is hardly likely to be noticed, since hardware audio CD players may also have an effective offset that varies from model to model.

    - M.

What if offset is not corrected?

Reply #3 –
Quote
it is possible that if the track indices were close enough to the actual non-zero audio on a disc, you might wind up with an audible "pop"

Possible, but surely very unlikely, since most offsets are just a few samples?

As I understand (correct me if I'm wrong), there is no one 'correct' offset correction that guarantees extraction of all samples from all CDs; the reference value used by EAC et al. is, as trout has said, mainly to enable comparison between different drives (usually different users).

What if offset is not corrected?

Reply #4 –
Since CDs have 44100 samples per second, the likelihood that anyone can audibly tell the difference if the offset isn't corrected is minimal at best. 

+/- 441 samples is +/- 1/100th of a second ( or  +/- 0.01 seconds).

Even if your CD/DVD drive has an offset of +/- 1764 samples, that is still just 4/100ths of a second, or 0.04 seconds either way.

I'm pretty sure you wouldn't be able to reliably tell the difference between a 4:00.00 track and a 4:00.04 or 3:59.56 version of it.  I know I couldn't.

What if offset is not corrected?

Reply #5 –
Offsets do not alter tracks' durations; they shift the audio relative to its indices.

What if offset is not corrected?

Reply #6 –
Quote from: dv1989 on
Offsets do not alter tracks' durations; they shift the audio relative to its indices.


But in terms of adding or removing up to .04 seconds of silence (in my example) in the gap at the beginning or end of a track, can anyone really tell the difference audibly?

What if offset is not corrected?

Reply #7 –
I doubt it--as, like you, I said in my post.

What if offset is not corrected?

Reply #8 –
Well, that's clear! Thank you everyone!

What if offset is not corrected?

Reply #9 –
You couldn't tell the difference if the rip is otherwise fine, but isn't offset correction required for AccurateRip to function at all?  While I wouldn't worry about the offset differences being audible, I'd be concerned about not using AR to help verify the rip is accurate in the first place.

I've never tried using Foobar to rip CDs though, so I don't know what kinds of options it has for detecting errors.

What if offset is not corrected?

Reply #10 –
Quote from: WonderSlug on
But in terms of adding or removing up to .04 seconds of silence (in my example) in the gap at the beginning or end of a track, can anyone really tell the difference audibly?

While this not be the case here, I want to clear up what seems to be a common misunderstanding.  With the exception of either the first track or last track (depending on the direction of the offset correction and assuming your drive is not overreading), silence is not added to interior tracks when an offset is applied.  Instead, samples are grabbed from the adjacent track.  Now maybe you're talking about silence coming from the adjacent track, in which case I agree with you, though this wasn't exactly clear from your post.  Also, I would not use the word "gap" when explaining the process since it has a specific meaning when it comes to DAE which has nothing to do with offset correction.

Quote from: Drexl on
While I wouldn't worry about the offset differences being audible, I'd be concerned about not using AR to help verify the rip is accurate in the first place.

You can use CUETools.

If others have raised any doubts, I suggest you re-read what M's reply.  It was perfect.  While many may never hear a difference with any disc in their collection, it is possible for an offset of even just one sample to be audible, if that sample happens to belong to an adjacent track and has enough amplitude.

Hmm, I just had an interesting thought, considering that the reference used by EAC is probably not absolutely correct, maybe we should call it offset calibration rather than offset correction; or is this just a distinction without a difference?

What if offset is not corrected?

Reply #11 –
Quote from: greynol on
While this not be the case here, I want to clear up what seems to be a common misunderstanding.  With the exception of either the first track or last track (depending on the direction of the offset correction and assuming your drive is not overreading), silence is not added to interior tracks when an offset is applied.  Instead, samples are grabbed from the adjacent track.  Now maybe you're talking about silence coming from the adjacent track, in which case I agree with you, though this wasn't exactly clear from your post. Also, I would not use the word "gap" when explaining the process since it has a specific meaning when it comes to DAE which has nothing to do with offset correction.


Yes, that's what I meant in my post above, but didn't clarify properly.  Thank you for that. 

What if offset is not corrected?

Reply #12 –
Quote
it is possible for an offset of even just one sample to be audible, if that sample happens to belong to an adjacent track and has enough amplitude.

I presume you mean that if an affected track was queued alongside another (not the next) track, and the amplitude of its last sample (a.k.a. the first sample of the actual next track) was different enough from the forthcoming track's first one sample, this could produce a click?

Whoa, I can find no easy way to articulate that question.

What if offset is not corrected?

Reply #13 –
...or just played by itself, that's correct.

If you're playing the two tracks surrounding the offending sample in question together on a gapless player, you won't hear anything out of the ordinary, of course.

EDIT: I'm not trying to scare anyone here.  The circumstances need to be either extreme or contrived for this to happen.  This is merely a possibility.

What if offset is not corrected?

Reply #14 –
But is it not also possible to get an offending 'click' when offset correction IS used? I don't have a specific example in mind, but it seems like I've encountered this before. Anyhow, maybe the result is better more often than not when the EAC established offset is applied. But I don't think the point should be misconstrued as meaning that offset correction guarantees perfect track separation.

Quote
The circumstances need to be either extreme or contrived for this to happen. This is merely a possibility.

Thank you for adding that. I will also add a reminder that most of the world rips without offset correction. And although 'they' might not be as discerning / discriminating as 'us', I think offset correction would be more standardly used if it were causing a significant problem rather than a rare one.

edit: grammar

What if offset is not corrected?

Reply #15 –
Quote from: trout on
But is it not also possible to get an offending 'click' when offset correction IS used?

It is indeed possible, absolutely!

I have a couple of discs that require an offset of a few thousand samples so this doesn't happen.  Having tracks that pass AR verification aren't anywhere near as important to me in this situation.

What if offset is not corrected?

Reply #16 –
Quote
But is it not also possible to get an offending 'click' when offset correction IS used?

I suppose it's possible at any boundary between sufficiently polar samples, as is likely to occur between non-consecutive tracks.

Edit: I'm daft and didn't consider the possibility of CDs with misplaced indices; that's a more important scenario!

What if offset is not corrected?

Reply #17 –
It's worth noting, in this discussion, that the track indices on many CDs are not carefully placed. This is true with many typical studio albums (that have some silence at the beginning of a track as well as at the end of the prior track) and in live albums or continuous-play albums (where there may be a half-second of applause or whatever before the point at which the track "actually" begins.

For many of my rips, I will pre-process the individual tracks before encoding, removing silence at the beginning (but not the end) of a file. For live or continuous-mix cd's (or songs that are this way, on an album that mostly has gaps in between tracks), I will frequently adjust the track cut-point position.
And by adjustments, I am often going further than the likely range of of CD player or drive offsets. But to be clear, I'm not talking about misplaced indices (where the track index is more than a second off from the obvious or likely point). I've seen such CDs, and those are a different issue.
Some people will say that I am going against the "artist's intent" but I strongly suspect, for the vast majority of CDs, track indices being "off" is not a result of artist's intent, but simply sloppy final mastering for CD production.

The good news is that more recent releases are more likely to have carefully-placed indexes. But I'm still surprised that it's not more of a point of careful-ness. Especially in the digital age where listening to separate tracks is much more common than in the CD age, let alone prior.
God kills a kitten every time you encode with CBR 320

What if offset is not corrected?

Reply #18 –
Yeah, I just wasn't satisfied with the track starts on my Marvin Gaye What's Going On CD (the twofer with Let's Get It On).  That album has several tracks that run together, but with very abrupt song starts and finishes.  I ripped each track using the range copy of EAC, adjusting a little bit before or after the CD's actual track boundaries, to get it just right.

For The Beatles' Abbey Road, I just decided to rip the medley as one track.  Am I really going to ever want to hear "Mean Mr. Mustard" by itself? 

What if offset is not corrected?

Reply #20 –
When a new pressing run is done, new offsets are introduced which are way larger than drive offsets (normally).

What if offset is not corrected?

Reply #21 –
Quote from: timcupery on
The good news is that more recent releases are more likely to have carefully-placed indexes. But I'm still surprised that it's not more of a point of careful-ness. Especially in the digital age where listening to separate tracks is much more common than in the CD age, let alone prior.


There's one fairly recent horrible example of careless editing, the deluxe version of Non-Stop Erotic Cabaret by Soft Cell... some of the tracks were taken directly from a previous remaster (of Non-Stop Ecstatic Dancing), where the songs Sex Dwarf and Tainted Love (Extended) are gapless, with a index between them. For the new remaster, someone cut them in the wrong place, and to make it even worse, they are not sequenced next to each other on the new album, so everytime you want to listen to Tainted Love, you get the last few seconds of Sex Dwarf at the beginning of it 

What if offset is not corrected?

Reply #22 –
all reasons why AccurateRip is good for making sure you have a bit-perfect rip, but doesn't ensure "correctness" or "perfection" of the actual track indexing
God kills a kitten every time you encode with CBR 320