Film-Tech Forum ARCHIVE: Dolby Digital Drop-out Adjustment

This is topic Dolby Digital Drop-out Adjustment in forum Film Handlers' Forum at Film-Tech Forum ARCHIVE.

To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=1;t=010441

Posted by Markus Lemm (Member # 3456) on 04-29-2010, 04:49 PM:

I have a DA-20 that drops out at some splices, pre-existing or our splices. Using the DRAS I see that when it drops it has more than 4 consecutive unreadable data blocks. Is there a way to change this value to 5 or 6, maybe with a firmware update?

Posted by Monte L Fullmer (Member # 2797) on 04-29-2010, 07:53 PM:

Don't use opaque tape, if you're doing this, or get the narrower width, 3/4" "Splicemark" splicing tape that won't be picked up by the reader since this narrower tape will not block out an entire, 4 sprocket block of the SRD track.

This goes for all SR-D readers - both basement and penthouse readers.

-Monte

Posted by Joe Redifer (Member # 3) on 04-29-2010, 11:17 PM:

Is it dropping out at the reel changes? Even with solid yellow opaque tape (was that redundant?) it should not drop out at a reel splice if your equipment is in a good state of repair. But splices between trailers and in the middle of the reel you need to be really careful on how you make the splice, making sure your splicer is adjusted and not chopping things up.

And no, there is no way to increase the amount of unreadable blocks because it would just revert to silence if you did.

Posted by Brad Miller (Member # 2) on 04-29-2010, 11:23 PM:

What is your typical error rating?

What brand and type of splicing tape are you using?

What kind of splicer are you using, and is it TRULY in proper alignment?

What model Dolby Digital reader do you have?

Something is definitely wrong, as I NEVER have dropouts at splices. Most often I find it is due to lousy splices, but you could have a reader in need of adjusting, or if you have a cat699 or cat700 penthouse it might be time to change the bulb, etc.

Posted by Markus Lemm (Member # 3456) on 04-30-2010, 09:16 AM:

We only use the yellow opaque tape for splicing films running in DTS, the clear stuff for anything SR or SR.D. Typical error rate is about 3 to 5 on trailers and newer films, our Dolby Reader is the DSTR-20 installed about 1 year ago.
Brad you are right about the splicer not being in perfect alignment. I find that I have to hold both sides of the film apart to get undamaged blocks. The splicer does need some work.

My real question is since Dolby Digital is already buffering the sound for my penthouse reader by nearly 3 seconds, couldn't it just compensate by doubling up known blocks so that the sound does not drop when it gets to the damaged area?

Posted by Sean McKinnon (Member # 612) on 04-30-2010, 01:22 PM:

When aligning the reader do you use a scope in addition to dras? If not you really should...

Posted by Louis Bornwasser (Member # 3063) on 04-30-2010, 01:58 PM:

My one-note mantra: dras is almost useless for alignment. Scope only (plus dras on-the-back-end.) Louis

Posted by Demetris Thoupis (Member # 833) on 05-01-2010, 04:54 AM:

The BACP penthouse usually comes factory pre-alinged and no further adjustment is needed. I would check if you have another DA20 to change the CAT 670 (video aquisition board) to check if the problem moves.
Demetris

Posted by Sean McKinnon (Member # 612) on 05-01-2010, 01:42 PM:

is the dstr-20 the penthouse or the basement reader i don't remember?

Posted by Joe Redifer (Member # 3) on 05-01-2010, 02:41 PM:

Penthouse.

Posted by Sean McKinnon (Member # 612) on 05-01-2010, 02:47 PM:

Then it should be factory adjusted perfectly however if something got loose in shipping or someone tried to "make it better" after install it may be out of alignment now. Using a scope is the only way to tell if it is in correct alignment and that you have an adequate video signal level.

Posted by Demetris Thoupis (Member # 833) on 05-01-2010, 03:15 PM:

The bacp unit actually comes with the triled indicator. Yellow for excessive , green for adequate and red for Low video level. Factory alingnent is spot on.vmaje sure you have constant tension of reader from platter. A video of the splice passing through might help us more because there is a secret adjustment. Also make sure you thread on the guidelines indicated across the chassis especially at the tensioning rollers.
Demetris

Posted by Joe Redifer (Member # 3) on 05-01-2010, 03:28 PM:

The Dolby Cat 702 has that same indicator with the same colors. I received a replacement LED for one of my DSTR-20's and it is almost always orange/yellow (though sometimes green when film is running). It can't be turned down enough. Plays fine, but I'm sure someday the LED will wear down enough to the point where it is perfect.

Posted by Sam D. Chavez (Member # 1841) on 05-01-2010, 06:50 PM:

There is a LED current adjustment on the DSTR inside the CCD cover.

BACP actually came up with this idea and later Dolby modified their board to do the same Tri-color LED.

Posted by Steve Guttag (Member # 268) on 05-01-2010, 07:17 PM:

All current basement readers feature the red/green LED thing too. Also, the CP650 can read out the video level on either its front panel (stupidly, it is in the "installers menu" rather than in the normal menu) or via the set up software where some engineer was thinkin'. The display is HUGE so it is easy to read at a distance.

As for Dras...I suspect, on a CP500 or DA20 system, I could get the reader alignment, including video level just about perfect using only Dras..it would certainly be harder as you are looking at results. On a CP650, I could also probably do it though it would be much harder since I wouldn't have the LEDs on the "corners found" to aid in getting things close, particularly with the LED positioning.

Steve

Posted by Joe Redifer (Member # 3) on 05-02-2010, 12:47 AM:

Steve, why do you think it is stupid that the video level is in the installer's menu? Just curious.

Posted by Steve Guttag (Member # 268) on 05-02-2010, 03:52 AM:

Because there is no reason to put it there. If one has a problem and you are trying to diagnose it (over the phone, in some cases, as a service company)...putting the CP650 in installer mode, requires passing through the Pink Noise/set up menus to get to the Video Level.

There is nothing about the video level that is "dangerous" for the end user...it isn't an adjustment, just a display. It is no more dangerous than the "error thermometer" in the DA10/20/CP500.

Steve

Posted by John Hawkinson (Member # 1135) on 05-05-2010, 08:57 AM:

Whoops! A while ago, Markus asked, "My real question is since Dolby Digital is already buffering the sound for my penthouse reader by nearly 3 seconds, couldn't it just compensate by doubling up known blocks so that the sound does not drop when it gets to the damaged area?"

That's not now it works. There's not a huge amount of redundancy in SRD. There is some, but generally not enough to compensate for four consecutive bad blocks.

Buffering doesn't help you -- when the splice goes through the reader, it fails to read the digital track. The fact that this is decoupled from the time at which that data plays does not help it get the data back.

For the special case of reel-end splices, Dolby implements "splice caching" where the data blocks for that time are included elsewhere in the reel (additional redundancy), so those splices don't cause SRD to drop out. But that level of redundancy is not generally available with SRD, so you're out luck.

--jhawk

Posted by Markus Lemm (Member # 3456) on 05-05-2010, 03:51 PM:

I should clarify my question a bit better.
I was thinking that on my reader the offset is 270 blocks or about 2.8 seconds so the processor has time to duplicate some blocks over time to compensate for unreadable blocks. Let me give an example of what I mean.
Lets says the reader is unable to read 10 blocks of data in a row. This value is sent to the card that buffers the sound.
Lets says that every 24 blocks or 6 frames the offset buffer duplicates a block of data.
So after 240 blocks or 2.5 seconds the sound is "stretched" to compensate for the loss.
I realize that for the basement reader guys this would not be effective since the offset is too short but for anyone with a penthouse reader this could really help.
Any Dolby engineers out there?

Posted by John Hawkinson (Member # 1135) on 05-05-2010, 04:28 PM:

Ah.

The operation you ask about is not implemented.

The SRD processor has plenty of time to do it, however. The SRD track leads the analog track by 6 frames, which leads the picture by 20 frames. Even in the worst case of a basement reader, there are 250ms to perform such an operation, and that kind of operation would take less than 1ms to execute.

A lot depends on what you mean by "stretching." Generally speaking, let's imagine it needs 1 frame's worth of time to process the data. Once it has seen 4 frames of missing data (16 perfs), it would need to decide whether to drop to analog or to execute such a "stretch." It can't know whether there will be 4 more missing frames, or 40 frames, or never.

Many would argue that it is much better to fallback to analog immediately (10ms for a block or 42ms for a frame) rather than try to do any "tricks."

It is true that generally speaking repeating the last block probably sounds better than silence, and that empirically, sometimes you do get silence.

I believe Dolby also does do some psychoacoustical masking in the event of bad data, but that may merely be because of high error rates and not because of entirely unreadable blocks.

Anyhow, the short answer is it doesn't really do what you're asking for, and it is not because of time limitations. Philosophically I don't think there's widespread agreement that such a thing would be Correct to do, but they could certainly do a bit better than they do at present.

--jhawk