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: Why do ~70 of 220 CDs contain HTOA of around 32 frames? (Read 19993 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #25 –
My guess is that it is linked to the PQ offset settings used by the (pre-)mastering engineer.
It was recommended in the early days of cd-mastering to put the marker a bit before the actual start point to avoid cd players with slow fade-in or mute "features" chopping off the beginning of the audio. The mastering software I use has global settings to automatically adjust the offsets. About half a second for track 1 seems a plausible value to me, but there are no rules. This will added to the minimum 2 seconds pre-gap for track one that can't be modified since AFAIK it's required by the red book standard.

These are quite common features for pre-mastering software. I found this on the Roxio website.
Quote
Track 1 Start Offset: Track 1 Start Offset is a special start offset option that affects only Track 1.
[/size]Below is an example of the start of a PQ sheet, generated by my Sonic Solutions pre-mastering DAW. It shows the PQ offset values (all zero here since I prefer to set them by hand).
Code: [Select]
PQ Log:

Delivery Type:CD, DDP or Image - (CD Times are 75 fps)
Time Format:       30/NDF
PQ Track 1 Offset:       00:00:00:00      PQ Start Offset:       00:00:00:00
PQ Splice Offset:        00:00:00:00      PQ End Offset:         00:00:00:00
PQ MinIndex 0 Width:     00:00:00:00      PQ UPC/EAN CODE: 

PQ Track / Index Information:
————————————————————————————————————————————————————————————————————————————————
T-X  TITLE/ISRC     COPY EMPH NO OFFSET     OFFSET        OFFSET        CD          
                              TIME          TIME          DURATION      TIME        
                              hh:mm:ss:ff   hh:mm:ss:ff   hh:mm:ss:ff   mm:ss:ff 
————————————————————————————————————————————————————————————————————————————————
1                              
   0   Pause                 -00:00:02:06  -00:00:02:06   00:00:02:00   00:00:00   
   1   No.1                  -00:00:00:06  -00:00:00:06   00:02:52:20   00:02:00   
                                            Total:        00:02:54:20   
————————————————————————————————————————————————————————————————————————————————
[/size]

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #27 –
I've always started individual tracks with 0.5s silence* (because so many crappy CD players and mp3 players fade in the start of tracks when you seek to them), but that's in the track - it's not a pre-gap - if it is a pre-gap, it doesn't solve the problem.

* - 1 second for the first track. Can't remember why. I must have once owned something that took that long to settle down.

Cheers,
David.

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #28 –
Quote from: Porcus on
Quote from: mjb2006 on
The pause must be at least 2 seconds (150 frames), and it must be between 2 and 3 seconds (225 frames) for the first track and for the lead-out.
The latter was news to me. Any source?

As for the former, I think many people ignore this when writing CDs, especially when gapless transitions are in use. It’s common for burning programs to attempt to enforce a two-second-long gap that does contain silence, but that isn’t usually compulsory. Sure, one could leave in the pregap even when it does not contain silence, i.e. has no relevance to the material, but do many discs actually do this?

Good old logical Red Book.

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #29 –
The burning apps that put two seconds of silence between tracks do so starting after the first track. They don't add silence after the two-second lead-in, which is what is under debate.

Also, David is right, HTOA of silence does no good when it comes to problematic players that fade-in. What's worse is when the beginning of the first track actually starts in the ~32 frame extension of the lead-in (which I have experienced). Now there is no doubt in my mind that that is a mastering error. The problem is solved by padding the beginning with silence, not by moving the start point further into the beginning of the track.

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #30 –
Quote from: greynol on
Now there is no doubt in my mind that that is a mastering error.
It would be interesting to find out if it's an avoidable (human) error, or that the mastering software was to blame. TBH I can't see any reason why a mastering engineer would deliberately create such a short HTOA.
Goratrix mentioned earlier that mostly old masters were concerned. If that's true it might be possible to narrow down the problem to the (pre-)mastering tools. Before 1990 mostly (Sony) U-matic format was used. After 1990 cdr/pmcd, Exabyte and later also DDP image gradually became more popular, where the PQ data was created in a DAW (computer). The large variety in systems can make it difficult to trace the problem.
Did anyone else notice a link between this HTOA problem and the year of mastering ?

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #31 –
I don't think the problem is a short HTOA.  Rather, I think it was a lack of silence before the beginning of the first track which caused the beginning of the audio to fall into the HTOA.  Had the HTOA been any greater I imagine it would have been even worse.

In case I wasn't clear, Kees, I have to back up in order to play the track from the beginning.  Otherwise you only catch the tail end of a cymbal crash.

Artist: Iron Maiden
Album: Powerslave
Track: Aces High

Another disc in my collection with the same problem is Steve Morse Band - The Introduction.

These discs were likely manufactured in the late '80s or early '90s.

I'd love to hear an explanation as to how this could not have been an avoidable human error.

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #32 –
Quote from: greynol on
I'd love to hear an explanation as to how this could not have been an avoidable human error.
Last week I ripped an old Philips Classics CD from 1990 that I recorded. Lo and behold, track 1 has a 33 frame pregap. I know for sure that this was a U-matic (Sony 1610) master. Then I checked a few more old cd-rips which I know to be U-matic sourced and they all have this short pregap.

While I do have quite some pre-mastering experience, I've never attended a glassmastering session in the cd-factory. AFAIK U-matic tapes were transferred in real-time to the LBR (laser beam recorder), making them popular amongst the bits-are-bits sceptics who didn't trust higher-speed transfers. U-matic tapes used a time-code signal to synchronize with other equipment. The Sony U-matic based audio editing system DAE3000 was sample accurate, so the sync can't have been the problem. Perhaps there was a delay somewhere in the path that caused this offset. I'll try to ask a friend who owns a cd-pressing plant if he has an explanation.
To be continued

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #33 –
Thanks Kees.

FWIW, I'm pretty sure that I was able to reorganize the audio data in order against an alternate pressing which did not have the issue and verify it against the AR database.  In other words, the data that ended up in the HTOA exists in at least one other pressing where it was not in the HTOA.  I know this because the AR database does not include hashes from HOTA.

In any event, careful QC should have caught the issue.  Maybe it did and it was decided to go ahead and sell it anyway.

Regarding bits are bits skeptics, I fall into that category as I own a disc from a pressing that contains errors that are not present in other pressings from the same digital master (master as in the digital data prior to it being pressed).  I know it isn't just my disc because of AR data combined with the ability to extract the same data from multiple drives and programs with multiple settings without even the hint of difficulty.  This is assuming the immaculate physical appearance of the disc isn't adequate enough evidence.

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #34 –
Update: I've just found a cd with a track-1 pregap that is too recent to have been mastered on U-matic, probably on cdr. Sigh.
Next step is to compare a pre-master and a pressed cd, to check where the problem creeps in. I'll try to find and analyse some old masters (most cd-factories don't send them back though).
Quote from: greynol on
In any event, careful QC should have caught the issue. Maybe it did and it was decided to go ahead and sell it anyway.
I agree that if the problem occurs on the pre-master, it's the responsability of the pre-mastering engineer. OTOH a chopped off start of track 1 is hard to miss, so I'd be surprised if it would go unnoticed.
The pre-mastering engineer's job is to create a master that is approved by him/her and the client/artist. What happens after that is beyond his control and AFAIK he's never asked/payed to do QC of the final product (vinyl might be an exception). Just imagine productions with a surround master, stereo master, iTunes master etc. Replication is supposed to be transparent. That's what makes this thread so interesting to me.
Quote from: greynol on
Regarding bits are bits skeptics, I fall into that category as I own a disc from a pressing that contains errors that are not present in other pressings from the same digital master (master as in the digital data prior to it being pressed).
With "bits are bits sceptics" I mean ppl who "hear" differences in bit-identical signals. From what you describe those two pressings are not bit-identical (checksum), right ?

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #35 –
That's correct.  They are not bit-identical.

I do not pretend to know the reason for the errors (all of which I raised are indeed audible), so when you mentioned high-speed transfers I wasn't in any position to dismiss it as a possible source for error.

My hunch is that things weren't as easy 20+ years ago, especially as it relates to ensuring data was free from corruption, and it seems the process was less than fool-proof; far less than what some in the business may have believed to be to the point that it was maybe even taken for granted.  That or some of the people involved were sloppy and/or just didn't care.

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #36 –
Can the offset issue explain it, or are the mistakes too far off? If it sounds good in my end on my CD player, then I'm sending the master to you, and finally it ends up with a customer whose player starts into the first transient ... ?

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #37 –
It's about 3,500 samples for the Iron Maiden title.

Here's a CT log:
Code: [Select]
[CUETools log; Date: 6/4/2013 12:56:02 PM; Version: 2.0.9]
Pregap length 00:00:33.
Offset applied: 4100
[AccurateRip ID: 000d3835-005a3db7-630bf308] found.
Track   [ CRC    ] Status
 01     [74e319cb] (25/25) Accurately ripped
 02     [411b3a99] (25/25) Accurately ripped
 03     [9a65e404] (25/25) Accurately ripped
 04     [36435385] (25/25) Accurately ripped
 05     [4987bf36] (25/25) Accurately ripped
 06     [67b63155] (24/24) Accurately ripped
 07     [c3803e0e] (24/24) Accurately ripped
 08     [39750ba0] (24/24) No match but offset

Track Peak [ CRC32  ] [W/O NULL] [  LOG   ]
 --  100.0 [6158DF5D] [21D3F87B]           
 01  100.0 [5DA607FB] [6C8C1421] [612AA18E]
 02  100.0 [6B15AA1C] [DF03A376] [BDC19FBC]
 03   87.7 [958E7133] [8EB90C18] [2566AD66]
 04  100.0 [A5371478] [18D2C817] [5A1F6A22]
 05   99.5 [6602A575] [A5E40345] [BA6821B0]
 06   98.4 [F7751F7B] [BC1B6592] [312EDC82]
 07   96.4 [33221F0B] [92B2E742] [F75B87EE]
 08   93.1 [A2A699D8] [565BCF6F] [0A5F3435]
I ripped these tracks with a configured offset correction of -4070 samples (the remaining 30 samples are attributed to the drive based on the reference used by AR) to get rid of the issue with the first track.  The beginnings of the rest of the tracks all play fine. I don't recall if they did before the offset was applied.  I definitely had to scan backwards into the HTOA to play the beginning of the first track on the various hardware CD players I've used.

The Steve Morse Band title needed an offset correction of about -2000 samples.  It also has a 33-frame HTOA index.

I have a double-disc title which requires a similar offset in order to correct the start points for all tracks on both discs.  Neither have an HTOA index.  I ripped these with with an offset correction of -3500.

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #38 –
That's a lot. And certainly not the most common Powerslave AR-ID – could they have caught the mistake and silently pressed a new batch?

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #39 –
Quote from: greynol on
These discs were likely manufactured in the late '80s or early '90s.
I've been trying to find a pattern and all of the discs with a track-1 pregap I could find were made in the 80's and 90's. It doesn't seem to be linked to pre-mastering equipment/software or Record Company. My best guess at the moment is that it's a (historical) glassmastering issue.
Does anyone have a problem-cd which was made after the year 2000 ?

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #40 –
Quote from: Kees de Visser on
Does anyone have a problem-cd which was made after the year 2000 ?


Discovery by Daft Punk (2001)
Drukqs by Aphex Twin (2001)

Those are the only ones from this century that I could check. I no longer have cue sheets from my other albums.

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #41 –
Are we now qualifying discs with a small HTOA that doesn't affect playback sound quality as problematic, calling it an "issue"?

That would be silly!

I see no reason to give any credence to what has so far only been unsubstantiated FUD.  That it may inconvenience people attempting to manipulate ripped audio for whatever purpose hardly qualifies.

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #42 –
An easy fix would be for ripping software to rip those pregaps as track 0, systematically (or giving the option to do it).

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #43 –
It's not like there's a big button with HTOA on it but with EAC you can rip it in a couple of ways, none of which are all that difficult.  This has been discussed before on multiple occasions.

The thing is that most of the time there is really no real rational reason to rip HTOA when it is small like this.  When size of the HTOA exceeds a certain threshold that, IIRC, was completely reasonable, the GUI in EAC will display the first track in red.

Looking at the dBpoweramp website, HTOA extraction has been available since v13.  From what I've read, the functionality appears to be user-friendly and direct, unlike that in EAC.

Aren't there some open-source ripping programs to which the functionality can be added?  Have at it.

Sorry about the edits, sometimes things occur to me after I've posted.  I know that annoys at least a few people.

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #44 –
CUERipper also can rip HTOA to separate track and it is opensource
http://www.cuetools.net/wiki/CUERipper
 

Why do ~70 of 220 CDs contain HTOA of around 32 frames?

Reply #45 –
Quote from: greynol on
Looking at the dBpoweramp website, HTOA extraction has been available since v13.  From what I've read, the functionality appears to be user-friendly and direct


Yes. Insert the CD and it pops up in the track listing as track 0, above track 1, with title Hidden First Track (ability to rip depends on CD drive). Will be ripped as such too. Not unlike the way you explain that EAC works (what is wrong with that one?) Ordinary 2-sec pregaps do not, so I assume that is length-based like EAC too. I have one that is as short as 10 seconds.

dBpoweramp will add 1 to the totaltracks field, opinions will differ on whether that is a good thing, and otherwise tag the file with those tags which are common through the album.