Film-Tech Forum ARCHIVE: Curious about DA-20 error readings

This is topic Curious about DA-20 error readings 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=007912

Posted by Jack Ondracek (Member # 1466) on 04-09-2006, 01:07 PM:

How is it that a trailer... probably printed from the tail end of some film roll, will show error rates of 0 or 1, yet some brand-new features will never drop below 5?

We're running XLs, top-mounted reader, dedicated Kelmar/Filmguard on each screen (pictures on our web site).

Having scoped everything, and verified all parameters per the book, I'm assuming the good rates on the trailer roll indicate good alignment. Bad assumption?... or what?

Posted by Louis Bornwasser (Member # 3063) on 04-09-2006, 02:19 PM:

Low readings indicate that the film involved happens to agree with your reader. It is entirely possible to be "out" slightly and have an also appropriately "out" film. It is therefore necessary to be in the center of "normal." That is why it is always necessary tol use verified good test loops instead of just any film, which could be almost out-of-tolerance. Louis

Posted by Jack Ondracek (Member # 1466) on 04-09-2006, 06:10 PM:

Makes sense to me. I got a bunch of new Dolby test film through Vern... enough to make several loops. I've seen the comments here about Dolby being less accurate than NT... did that include the digital stuff, too?

Posted by Louis Bornwasser (Member # 3063) on 04-09-2006, 06:53 PM:

Jack; the DIGITAL loops are double-checked at Dolby so they should be OK. Louis

Posted by John Hawkinson (Member # 1135) on 04-09-2006, 06:59 PM:

Brad Miller suggests (in a thread long ago I cannot find) that you should use a bunch of trailers from the major labs (Deluxe and Technicolor, multiple locations). That way, assuming that the trailers are printed on equipment calibrated the same as features (or even the same equipment), you're pretty much guaranteed to be right.

It makes sense to me. Why rely on a test film to be at some standard when you can adjust empirically for what is really correct?

Of course the labs have a harder calibration problem...

--jhawk

Posted by Sam D. Chavez (Member # 1841) on 04-09-2006, 09:12 PM:

The previews are typically made on a different production line is one possible reason. Others have noted the previews often look better than release prints as well.

As to test films, I believe the negative used to make Dolby test films used EC-9 software while the previews and features are using the current EC-11.

Posted by Richard May (Member # 2627) on 04-09-2006, 09:22 PM:

I think Sam partially answered the question. I too have seen trailers run much cleaner in regards to low error rates than features do. Whats up with that??

Posted by Monte L Fullmer (Member # 2797) on 04-10-2006, 04:06 AM:

..wondered about that as well - watching the display on a DA-20 when trailers are playing and with excellent error rate readings..then the splice to the feature and all of a sudden mid '5' to '6' then suddenly 'F' readings.....but the unit doesn't drop out of digital with this sudden flip-flopping (uh..oh .. hook up the laptop with the WinDRAS program installed and watch the jitter rate go crazy...have I got a goofy print..?)

-Monte

Posted by Michael Schaffer (Member # 1204) on 04-10-2006, 05:53 AM:

quote: Sam D. Chavez
As to test films, I believe the negative used to make Dolby test films used EC-9 software while the previews and features are using the current EC-11.

Could you elaborate on that a little? What is EC-...? Why would different encoding software (if that's what it is) cause different error rates on printed stock?

Posted by John Hawkinson (Member # 1135) on 04-10-2006, 10:04 AM:

Well, Monte, the jitter rate is really independent of calibration. So if you have a print with a high jitter rate, it's definitely a characteristic of print.

--jhawk

Posted by Gordon McLeod (Member # 33) on 04-10-2006, 10:08 AM:

Using a trailer is all well and good but a reference is a reference and I have found typically it is best to rely on standardized test films

Posted by Sam D. Chavez (Member # 1841) on 04-10-2006, 10:29 AM:

I forget what the EC actually means. The 9 and 11 refer to software revisions in the Dolby 5.1 encoder used in making the soundtrack. The higher error rate on test films does not have to do with the EC-9 or 11 specifically afaik, but the whole process of making a soundtrack has improved over time.

As Gordon says, the reference is the reference.

Posted by John Hawkinson (Member # 1135) on 04-10-2006, 11:58 AM:

I think I wasn't so clear. The point of using trailers is not to use a trailer, but that you can use a WHOLE BUNCH of data points, and calibrate to the mean (because you have multiple trailers from each of the labs).

So you're not stuck calibrating to a single reference that might be slightly off. You shoot for the setting that makes the most of the trailers come out right, thus optimizing for the average of Deluxe Toronto, Deluxe Hollywood, Technicolor North Hollywood, and Technicolor Mirabel. With perhaps DuART and Technicolor New York thrown in for kicks [Smile]

(Of course, maybe I was clear and you still disagree -- that's fine [Smile]

)

--jhawk

Posted by Jack Ondracek (Member # 1466) on 04-10-2006, 12:18 PM:

I generally agree with Gordon's notion of relying on standards for system calibration. John's post also outlines what'll probably be necessary here. Unfortunately, for me, that feeling goes back to when you'd want a single standard reference for audio tape recording. Many of the stations I worked for sent dubs out to other stations, and having my equipment GENERATING product as close to NAB standards as possible was always my goal.

On the other hand, if after using a known standard on my PLAYBACK equipment, I get great readings on some media, and lousy readings on others, I can see the point for compromise. Going back to my radio days, we had a bunch of Ampex PR-10 machines that had thumbscrews so that you could fine-tune playback head azimuth. It made a big difference for us... especially when receiving dubs from other stations! [Smile]

Back to the booth here... IF such a compromise resulted in overall error rates around, say, 3... maybe 4 tops I'd be much happier with the system as a whole.

I have to admit though, knowing my system might have to be "off" in order to play everything "OK", does bother me a bit. It seems that the error rates were more consistent, back 2 or 3 years ago... but maybe I'm just imagining that.

Posted by Michael Schaffer (Member # 1204) on 04-10-2006, 05:29 PM:

quote: Sam D. Chavez
The higher error rate on test films does not have to do with the EC-9 or 11 specifically afaik, but the whole process of making a soundtrack has improved over time.

I guess that must have something to do with better error correction and data redundancy then since I don't see how the physical properties of the printed track would be improved by improving the encoding software?

Posted by Sam D. Chavez (Member # 1841) on 04-10-2006, 06:58 PM:

I don't think it has to do with the software, it's probably just a clearer image as the test films were created very early on for obvious reasons. The negative has never been updated as it works fine for the intended purpose.

Posted by Eric Robinson (Member # 2966) on 04-12-2006, 11:56 PM:

As a newbie on the block I mistakenly used Dolby Digital test film along with DRAS to set my azimuth and lateral values to "0".

I then learned at a later point in time that each piece of Dolby Digital test film has a handwritten calibration error value. The test film I am using now requires that I set azimuth to "-2" and lateral position to "+1.6" due to the printing error on the test film.

I always use a couple of brand new trailers from different vendors to run through after I have performed a digital alignment so I can view the results of my efforts. Then when the film in use gets run my numbers go in the tiolet.

When ever a theatre employee complains to me, though, that they are having digital error rate issues I always ask them to run a brand new trailer through during a break to see if there really is a problem with the alignment.

ERIC

Posted by Ron Curran (Member # 3631) on 04-13-2006, 04:25 AM:

Unfortunately, as has been pointed out many times, the digital dots are in one of the most vulnerable areas on the print. They will work without dropouts for a long time if the print is run by careful operators.

If perfect colour filters could be invented, the dots could be printed magenta where the cyan analog track is – but that is probably too hard.

Posted by Louis Bornwasser (Member # 3063) on 04-13-2006, 05:43 AM:

Remember that the computer hookup occurs after error correction and is mostly useless for full set up on Dolby Digital. There really is no substitute for the proper scope. Louis

Posted by John Pytlak (Member # 331) on 04-13-2006, 07:06 AM:

quote: Ron Curran
If perfect colour filters could be invented, the dots could be printed magenta where the cyan analog track is – but that is probably too hard.

The Kodak/ORC Cinema Digital Sound (CDS) digital sound system printed a dye track in the current analog track area. Cyan dye is used for red LED readers, magenta dye would work best with green LEDs:

http://members.aol.com/cds3570/

quote:
The first commercially available digital sound-on-film system for theatrical motion pictures, CDS is gone now. Today we have DTS-Digital, Dolby Digital, and Sony Dynamic Digital Sound. In fact SDDS was designed by some of those who pioneered CDS.

What went wrong? Well opinions here vary. Some would say that the choice to use minimal data reduction and the resulting very small bit (pixel) size made the system much more susceptible to every type of problem including print wear, film lab errors etc. Others will point out the fact that the lack of a backup analog track (In 35mm CDS sound was printed in the same area were the analog optical track would normally go) would forever make CDS a premium process, requiring dual print inventories, not to mention making the occasional breakdown catastrophic and reputation destroying. Some have said the electronics simply wasn't reliable enough for hot projection booths. But everyone agreed on this: When it worked and it really DID work quite well the vast majority of the time, IT SOUNDED GREAT!

Posted by Ron Curran (Member # 3631) on 04-13-2006, 07:25 AM:

Thanks John. The Colortech system didn't get a hearing either, though it sounded promising. I was pipe-dreaming about the analog and digital tracks sharing the same area if they could be discretely colour-seperated.

Posted by Jack Ondracek (Member # 1466) on 04-24-2006, 01:38 AM:

This week, Silent hill played with a 2-rate. Benchwarmers is running at 5. What? Set it so they both run at 4?

This sucks.

(Joe's "pointy-finger" picture goes here!)

Posted by Michael Schaffer (Member # 1204) on 04-24-2006, 02:28 AM:

It doesn't really matter if it runs stably around 5. Remember, that doesn't make it sound any different, unless it goes above 8 and drops out.

Posted by Jack Ondracek (Member # 1466) on 04-24-2006, 03:22 AM:

Yeah, I know, Michael.
It just doesn't seem like that's the way it should be... that's all. Otherwise, it does work OK, and nobody knows about it but me. You just get the feeling that some quality control issue is falling through the cracks, and it doesn't seem right that TRAILERS should come up with better readings.

Posted by Louis Bornwasser (Member # 3063) on 04-24-2006, 01:13 PM:

My guess is that trailers are printed relatively slowly compared to the actual feature. Louis