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: How much silence belongs to our music ? (Read 4600 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

How much silence belongs to our music ?

I just read several articles about the system how Andre Wiethoff set the reference for the eac offset database. I draw the conclusion (like many others), that (read)offset correction is absolutely useless when i don't intend to burn audio-cds once in a while. Just because we don't know how much silence the producer intended to add before the first and after the last track, we can't say its more accurate with offset correction. All we have to make sure is that we don't cut anything else than absolute silence from the beginning or the end. Furthermore i would even say it is legit (in some cases) to cut null samples from your ripped files. Only if you tend to listen to full albums the silence between tracks is necessary, because some producers 'take the beat/rythm to the next song'.
In that context, a question comes up (im not into the Red Book): trackmarks are relatively to the start position of a cd ? This would mean that we could get absolute offset (at least for this mastering machine) with two consecutive tracks, one with noise and the other one with silence, without a gap between them. But this just seems to be too easy that no one thought about it. So i would be totally satisfied to see silence before the first and after the last track. For burning you could even refer to a combined read/write offset not to get generation effects, but the audio cd as a backup and listening medium is dying anyways, with more and more lossless compression coming up. So why bother ?

How much silence belongs to our music ?

Reply #1
Plus, depending on the player, you can easily add "gaps" if you desire them. And most burning apps as well allow to add gaps between tracks.

Overally, i think its a matter of taste. As for me, i strip all digital silence when ripping, so that i can decide how much or how less silence i want. Plus, without silence, applying an automatic crossfade is easier if there isn't too much silence between tracks.

As for artists - i doubt most care much about about silence - and if they do, then its easy to notice (talk about a super-long fadeout, then 9secs of analogue silence and then 5secs of digital silence).

- Lyx
I am arrogant and I can afford it because I deliver.

How much silence belongs to our music ?

Reply #2
Some tracks start off at full volume on the first few samples. (Or vice versa with the end of the track being loud.) If you're running in shuffle mode then these samples can get mixed up with quieter sections of adjacent tracks which gets very annoying.

How much silence belongs to our music ?

Reply #3
Quote
trackmarks are relatively to the start position of a cd ? This would mean that we could get absolute offset (at least for this mastering machine) with two consecutive tracks, one with noise and the other one with silence, without a gap between them. [a href="index.php?act=findpost&pid=354857"][{POST_SNAPBACK}][/a]


That's the way offsets appear. Depending on the drive offset, the track marks are more or less early, or late, compared to the noise.
The problem comes from an error in the Red Book specifications. The track marks are interleaved relatively to the audio data !

When the audio runs from track 1 to track 2, the track marks actually say something like
11112111211122112211221222112221222122222
instead of
111111111111111112222222222222222222
So where does track 2 begin ? Well, somewhere in this mess.

How much silence belongs to our music ?

Reply #4
Quote
That's the way offsets appear. Depending on the drive offset, the track marks are more or less early, or late, compared to the noise.
The problem comes from an error in the Red Book specifications. The track marks are interleaved relatively to the audio data !


I don't expect you explaining the red book to me, but in this case offset correction is not only affecting the silence before the first and after the last track, but even the relative position of a certain sample to the first sample of a track ? So it would be better to use Andre's (probably incorrect or not absolute definable) reference-offset to correct the read-offset of my drive as to just ignore read offset and keep ripping with 600++ samples differing from this reference. This could lead in extreme cases to a song having samples left in the track before or after itsself ?
Any mistakes in my thoughts ?
So offset correction is not that useless as i thought, is it ?
thanks for your replies, guys.

edit: I don't really get your explaination about how trackmarks are interleaved in the audio data. I as a newbie thought up to now a track mark was a single information. How could this lead to something like
Quote
11112111211122112211221222112221222122222

? If you don't mind you could give some more information about these track marks or at least some resources.

How much silence belongs to our music ?

Reply #5
Quote
in this case offset correction is not only affecting the silence before the first and after the last track, but even the relative position of a certain sample to the first sample of a track ? [a href="index.php?act=findpost&pid=354919"][{POST_SNAPBACK}][/a]


The samples are always in the same order. It changes its position relatively to the beginning of your file, not relatively to any other sample.


Quote
This could lead in extreme cases to a song having samples left in the track before or after itsself ?[a href="index.php?act=findpost&pid=354919"][{POST_SNAPBACK}][/a]


Yes.

Quote
So it would be better to use Andre's (probably incorrect or not absolute definable) reference-offset to correct the read-offset of my drive as to just ignore read offset and keep ripping with 600++ samples differing from this reference. [a href="index.php?act=findpost&pid=354919"][{POST_SNAPBACK}][/a]


Only if your CD was written using the same offset reference. With a commercial CD, you will never know if correcting the offset will lead you closer or farther from the original layout.

Quote
I don't really get your explaination about how trackmarks are interleaved in the audio data. I as a newbie thought up to now a track mark was a single information. [a href="index.php?act=findpost&pid=354919"][{POST_SNAPBACK}][/a]


If someone wants to play track 5, it would take a lot of time for the drive to find it ! Having to read about 5 minutes of audio in order to be sure not to miss the track mark...
So in the CD, the data containing the information about the track is repeated everywhere. The drive just have to go to a given place, and the subcodes channel immediately tell him "You are in track 5, time 0:03:25:45".
The problem is that between tracks, in a small zone (about 600 samples) this information goes forth and back between the two tracks. So the drive can't be sure to be in the right one. Designers have found their own way to deal with this... which lead to different offsets.

How much silence belongs to our music ?

Reply #6
So is there any point having read offset correction enabled? Does it really offer a benefit? I mean, my drive's is only +6 but I thought I may as well use it since EAC supports it. Will it really make much difference?

How much silence belongs to our music ?

Reply #7
If you wish to compare ripped results across different drives you need offsets.

How much silence belongs to our music ?

Reply #8
So it's good to use correction as something of a 'reference value'?

How much silence belongs to our music ?

Reply #9
Quote
So it's good to use correction as something of a 'reference value'?

Yes and that seems to be the one and only use for read offset correction. And a reference value is only useful when comparing your results with results of other users, in case of eac with AccurateRip. I'm not quite shure if there are other approaches to comparing ripped audio, but I don't think so. This doesn't mean offset correction is useless as long as you don't compare, in fact when copying audio-cds you _should_ use a combined read/write offset (to avoid generation effects as I already said).
edit: In case you want to use different drives for reading and writing audio data, you should stick to separate read/write offsets as well, but imho copying already copied cds does lead to more generation loss than only wrong offsets, so i would avoid it anyways (and i see no use for this).

Quote
The samples are always in the same order. It changes its position relatively to the beginning of your file, not relatively to any other sample.

Sure, my expression was a bit ambiguous. I just wanted to say a certain sample 'jumps' to different positions in a song when ripped with different offset values. Thats clear by now anyways.

Quote
Only if your CD was written using the same offset reference. With a commercial CD, you will never know if correcting the offset will lead you closer or farther from the original layout.

If I'm remembering right you agreed at some point with andre that his 'absolute' offset went in the right direction. Not absolutely right but more or less okay. Did you change your opinion or did i get things wrong when reading your discussion ?

How much silence belongs to our music ?

Reply #10
Quote
Yes and that seems to be the one and only use for read offset correction. And a reference value is only useful when comparing your results with results of other users, in case of eac with AccurateRip. I'm not quite shure if there are other approaches to comparing ripped audio, but I don't think so.

Offset correction is usefull for :

1: Using AccurateRip
2: Comparing CRCs of rips done with different drives
3: Avoiding generation-loss when making several generations of copies of copies.

Quote
This doesn't mean offset correction is useless as long as you don't compare, in fact when copying audio-cds you _should_ use a combined read/write offset (to avoid generation effects as I already said).

No, only when making several copies of copies, to avoid generation-loss...


There are some very informative posts by Pio2001 on the subject here :

Arguments against offset correction : http://www.digital-inn.de/showthread.php?threadid=4193

How to find a reference offset : http://pageperso.aol.fr/Lyonpio2001/offset.htm

How much silence belongs to our music ?

Reply #11
Quote
If I'm remembering right you agreed at some point with andre that his 'absolute' offset went in the right direction. Not absolutely right but more or less okay. Did you change your opinion or did i get things wrong when reading your discussion ?
[a href="index.php?act=findpost&pid=355047"][{POST_SNAPBACK}][/a]


By now, EAC's offset reference is used by EAC, by AccurateRip, and by Plextools. If you want to use a reference, this is definitely the one to use.