AccurateRip & Ultra Secure

[sredmyer]

Ok I am trying to speed up my ripping process. What I see happening is that any disc that is not in AccurateRip database (or is but as a different pressing) will get ripped a minimum of 4 times given the following ultra secure settings.

Minimum Ultra Passes 2
Maximum Ultra Passes 4
End After Clean Passes 2

With these settings a disc which is not in AR will get ripped twice for AR checking then at least twice more for ultra secure rip.

My question are these;
1) what is the difference between the AR rips and an ultra secure rip?
2) can the AR rips be skipped when it is known that the disc is not in AR?

Thanks

[bhoar]

Re: AccurateRip & Ultra Secure

I take it you can't enable C2 because you're using the xl1b or powerfile. That'd be the simple solution if it were available.

If I were in your shoes, I would set try setting minimum ultra passes to 1. I have a hunch it defines clean as "matching a previous pass exactly" and may include the first (pre-ultra) pass in that definition and leaves you with three passes if the disc has no errors (and if my hunch is correct, which might not be the case).

Alternately, disable ultra entirely, and you're left with two passes.

In both cases, it will move on to frame re-rip if there were any non-matching frames across passes.

-brendan

[sredmyer]

Re: AccurateRip & Ultra Secure

I take it you can't enable C2 because you're using the xl1b or powerfile. That'd be the simple solution if it were available.

Yes that is correct. I think remember seeing somewhere that these drives do support C2 but Spoon was having difficulties with the Firewire interface. Do you know it this memory is accurate? If so any idea if he plans to continue to search for a solution?

So do you know what is the difference between a AR passes and an ultra pass?

Thanks for the suggestions.
Steve

[bhoar]

Re: AccurateRip & Ultra Secure

The sony/firewire problem is not fixed yet, though some other firewire drives now work. He said he may reexamine it for R13.2.

I always refer to this chart: http://www.dbpoweramp.com/images/mis...-method-db.png

I'm not sure there is a difference between the first pass and each ultra pass per se, other than when the order in which the passes occur, how the track audio data is compared/analyzed after the track is read and the decisions made after each pass. That is to say, I think the read commands used are the same.

But I could be wrong

I think the ultra passes are most useful when dealing with drives that don't support C2 at all (or don't support it very well) and aren't that hot at correcting errors themselves. These drives tend to need three or more reads of a track to discover the frames that are most likely problematic which are then handed to the frame re-rip routines.

-brendan

[sredmyer]

Re: AccurateRip & Ultra Secure

The sony/firewire problem is not fixed yet, though some other firewire drives now work. He said he may reexamine it for R13.2.

I always refer to this chart: http://www.dbpoweramp.com/images/mis...-method-db.png

I'm not sure there is a difference between the first pass and each ultra pass per se, other than when the order in which the passes occur, how the track audio data is compared/analyzed after the track is read and the decisions made after each pass. That is to say, I think the read commands used are the same.

But I could be wrong

I think the ultra passes are most useful when dealing with drives that don't support C2 at all (or don't support it very well) and aren't that hot at correcting errors themselves. These drives tend to need three or more reads of a track to discover the frames that are most likely problematic which are then handed to the frame re-rip routines.

-brendan

I have tried to understand that chart but I am unfortunately behind the curve a bit on this stuff. Can you explain the third step in the no C2 flow. I don't understand that one. If I read it correctly data is checked against AR after each pass and the process bails as soon as one matches. But if niether the first pass nor the second pass matched what is meant by the cross checking to see if they match?

Thanks

[bhoar]

Re: AccurateRip & Ultra Secure

My guess is what it does is first assume that all frames that match exactly between pass 1 and pass 2 are probably good. Then, it assumes that, for non-matching frames, one of the two is likely to be the correct frame, so then it it permutes all combinations of the non-matching frames and recalculates the accuraterip ID each time.

e.g. if your "track" was 10 frames long and each frame was one byte long (for arguments sake) and these were your two passes:

Frame num.:   00 01 02 03 04 05 06 07 08 09
Orig Disc:    D5 AA 96 FE FE D5 AA AD FF FF <- what is stored on the scratched disc

Read Pass 1:  D5 AA 96 FE FE D5 AA AD 04 FF (byte 08 is wrong)
Read Pass 2:  D5 BD 96 FE FE D5 AA AD FF FF (byte 01 is wrong)

Let's assume there was an Accuraterip entry found for the disc ID, after the second pass failed to match it would calculate accuraterip numbers using *two* additional sets of data:

Calc Perm 1:  D5 BD 96 FE FE D5 AA AD 04 FF (byte 08 and byte 01 are both wrong)
Calc Perm 2:  D5 AA 96 FE FE D5 AA AD FF FF (yes, we got lucky and matched on the last permutation!)

In total, it checked four possibilities: the first pass, the second pass, the first pass swapping the first non-matching byte with the second pass's version, and the first pass swapping the second pass's different byte with the second pass.

Now, as you know, in reality each frame is not one byte but is over 2KB in size. And the number of non-matching frames may be larger, and the number of permutations may grow pretty fast (with two passes I think it may be exponentially, 2^n total accuraterip computations, including the first two passes as is) if there are a lot of non-matches.

So, spoon's approach is certainly more complicated that the above example: perhaps he can swap portions of a frame as well, not just the entire framel; perhaps the accuraterip calculation method allows you to retry frames or smaller blocks of data without having to recalculate the entire track data; etc...

But that's my layman's understanding of what is implied by the secure-ripping page description.

-brendan

[sredmyer]

Re: AccurateRip & Ultra Secure

But that's my layman's understanding of what is implied by the secure-ripping page description.

-brendan

Thank you that was an extremely well stated explanation.

[sredmyer]

Re: AccurateRip & Ultra Secure

I'm not sure there is a difference between the first pass and each ultra pass per se, other than when the order in which the passes occur, how the track audio data is compared/analyzed after the track is read and the decisions made after each pass. That is to say, I think the read commands used are the same.

But I could be wrong
-brendan

I think the Ultra Secure passes are done at about 1/2 speed. At least they show that in the CD Ripper's rip speed indicator. I assumed this was due to some exhaustive error checking done while ripping that was not done when doing AR rips.

For example I am no watching a track rip one of the AR rips at x9. When it switches to the Ultra pass it will slow down to baout x4.5

But this is only an observation.

EDIT:
Oh well so much for that observation. Now it is doing an AR rip at x4.5. Now I am not sure what to make of the ripping speed. It seems as though it is all over the map. Anywhere from as low as x2.5 to as high as x10 on the same drive (even on the same track). No idea why that is.

[bhoar]

Re: AccurateRip & Ultra Secure

Contemporary drives will control their own speeds. e.g. they may slow themselves down to deal with hard to read areas of the disc.

On a related note, the speed setting option for Pass 2 and the ultra passes allows you to throttle the speed down. Normally dbpa requests (asks) the drive to use the maximum speeds available before it starts each track (if you turn on the log, you see SetDriveSpeed called before each track, even if you haven't change the speed values from the defaults). That doesn't mean the drive will use the maximum speed, it's just being told to use the maximum speed of the drive *if it can*. If you set these options to more restrictive settings, dbpa requests (asks) the drive to use lower speed values before each track and that becomes the maximum speed it is allowed to use for that track. And yet, the drive can and will slow down further if necessary.

-brendan

PS - there's no such thing as the "AR pass". There's just normal Pass 1 and, if C2 is not being used, there's also normal Pass 2. Yes, they are checked against AR, but I wouldn't call them AR passes.