my profile | my password | search | faq & rules | forum home

» Film-Tech Forum ARCHIVE   » Operations   » Film Handlers' Forum   » Dolby Digital WOW at reel change? (Page 4)

This topic comprises 7 pages: 1  2  3  4  5  6  7

Author Topic: Dolby Digital WOW at reel change? John Hawkinson Film God Posts: 2273 From: Cambridge, MA, USA Registered: Feb 2002 posted 01-07-2006 02:48 PM                     We seem to have come full circle. Dominic, yes, of course the problem is because the film is speed is changing. The question Mark is asking is why the CP650 does not smooth out rapid changes (a speed decrease followed by a frequency increase) to reduce the apparent visbility of the problem. It is not a direct "shortening the wavelengths" phenomenon, though -- that would apply in analog. Instead, the Dolby Digital decoder is deliberately changing the frequency in order to speed up the material to catch up (or slowing down, as appropriate). Rick: no, this is not about data loss. The "wow" is not about a resynchronization failure. The Dolby Digital decoder is continually watching the speed of incoming data blocks (to monitor the speed of the film) and adjusting the output frequency/speed such that they match. Because if your projector is running at 24.1 or 23.5, Dolby Digital still wants to work, and lip sync still needs to match. More than just a CRC, there is actually a Reed-Solomon error correcting code. See the "Theory of Operation" section of the Dolby DA20 manual, pages 102-104 of the PDF (aka pages 5-1 through 5-3). But it's not of relevance here. If splices are involved, they are pushing the film around and changing the speed physically, not necessarily causing a reversion to analog sound. Mark did tests to confirm there were no analog reversions. --jhawk  |  IP: Logged Lonny Jennings Film Handler Posts: 10 From: Boston, MA USA Registered: Feb 2000 posted 01-07-2006 11:15 PM                     We need to look at the source of the problem, not the symptoms. The problem is NOT what happens when a splice goes through the reader. That is irrelevant; the system design of Dolby Digital allows for these speed variations and SHOULD be inaudible. The real problem is why the CP650 can’t deal with it. I believe that Mark Marshall was absolutely correct in the beginning: it’s a CP650/Cat. No. 773 software problem and it does need to be looked at by Dolby. Here’s why: The front end of the Dolby Digital system tracks the speed of the projector by looking at the perfs going by. That’s right, the perfs, not the blocks of data. It doesn’t use the blocks to correctly lock onto the speed of the projector. The front end circuitry simply looks for the brighter light coming through the perf and a PLL (Phase Locked Loop) locks onto this as a reference clock for the entire front end processing. Whatever speed the projector is running at (from +7% to -11% of 24 fps) determines the clock for all of the digital processing in the front end DSP. You can see this by watching the right hand “goalpost” on the oscilloscope (you are using an oscilloscope, right?) and see how it tracks the projector speed even if there are no data blocks printed on the film. However, it does really track the speed of the projector quite closely and responds to speed changes instantly. That’s why you can see the right hand goalpost jerk back and forth following rapid speed variations (flutter) on a soundhead like a Christie. As bad as the speed variations may be, the front end can still lock on to the perf rate even if the rest of the system can’t cope with the speed irregularity enough to read the blocks. O.K., so the system is really good at following the speed (and its variations) of the projector but, that would make a pretty bad sounding system because any speed fluctuations (and we know there are lots during a splice) would be tracked by the front end and show up in the audio playback as audible wow & flutter (see where this is headed?). So….. after the system reads each block the data is sent into a FIFO. It’s not a cute little dog; it stands for First In, First Out. It’s a clocked memory buffer and it works to smooth out the data flow so that there is no wow and/or flutter in the data before it is converted to analog audio. How’s it work? Imagine a large barrel. Picture it having a faucet on the bottom like one on the side of your house for your garden hose. Imagine pouring buckets of water into the barrel one at a time. The flow of water into the barrel is anything but constant or smooth and, of course, if you don’t do anything else you will fill up the barrel and then overflow it. So now imagine slowly opening the faucet to let the water out (you have three really long arms). As you keep pouring buckets of water (data) into the barrel (memory buffer) you s l o w l y, gradually, adjust the faucet to maintain the level of water in the barrel (buffer) so that the level is somewhere in the middle. If you happen to speed up a little pouring in the water and the level starts rising in the barrel you slowly open the faucet more to let the water out a little more quickly and get the level back down to the middle. Conversely, if you slow down pouring in buckets you slowly close the faucet to slow down the flow to maintain a smooth, relatively constant flow of water coming out of the barrel. This is how the system is designed to work. Any changes to the speed of the data coming in are smoothed out by the FIFO so that it is inaudible to our ears once it is converted to audio. Things like splices in the reader should have no audible effect on the sound. That’s what is should do anyway. It sounds to me like it isn’t working right in the CP650. I don’t ever remember hearing of this with a CP650 before and it seems (correct me if I’m wrong) that it hasn’t been a problem for anyone with a DA20/CP500. Also, remember that the software for the Dolby Digital subsystem in both a CP500 and a CP650 is separate from the main operation system software like on the Cat. No. 684 or the Cat. No. 774. It is separate code and can be changed apart from the operating system, sometimes without your knowledge. Most of you are somewhat familiar with what is called the Dynamic Loader feature of Dolby Digital. With it, Dolby can put new software for the Dolby Digital subsystem on the film within the digital data on the film. When you play the film, it gets automatically loaded into the subsystem and then starts running that code from now on. In fact, you can’t stop it from happening. If it’s on the film it will get into your system. It’s possible that Dolby has done a software upgrade for Dolby Digital using Dynamic Loader and the new software has this bug causing the FIFO not to work. If that’s the case, they need to fix the bug and put it on all subsequent films to reprogram all units. A couple of other thoughts: When the Dolby Digital system is playing a feature film and it gets to the splice between reels you usually aren’t really hearing the sound coming from the area of the splice. That’s because of another feature of the system called splice caching. Using some of the extra bits in the data blocks of the digital track they record “spare” audio from the area around the reel change when the recording is made on the dubbing stage. If my memory is half here I believe it is about 3 feet on either side of the splice. For example, the digital track on reel 3 would have all of the audio for the last 3 feet of reel 3 and the first 3 feet of reel 4 recorded throughout reel 3. When it is played back through the processor, this spare audio is retrieved from the track and temporarily stored in RAM. Each data block is numbered in every film so the processor “knows” where these spare blocks are supposed to go. When the reel change splice goes through the reader and inevitably some blocks are missed because of the disturbance the processor looks to see which blocks were lost and looks up the spare one in RAM and plays that one out instead; the result? No missing blocks and no reversions. Of course, due to the way the spare blocks are recorded and then obtained this only works with missing data around the reel changes. Any splices between trailers or elsewhere in any reel are not covered and there may be a reversion. BTW. Splice caching was something that was developed years after the system was designed and after many processors where already installed and working. All Dolby Digital processors; CP650, DA20/CP500 etc. old and new are able to utilize this feature because all processors were upgraded with this new software using the Dynamic Loader feature! Adjusting the dashpot in a Dolby Digital reader: DON”T. The purpose of the dashpot is to work with the flywheel and damper arms to create a mechanical filter in a Davis loop system like the Dolby Cat. Nos. 699-702. It is adjusted for a certain value at time of manufacture to produce the lowest wow & flutter in the reader and should not be touched. It doesn’t “drift” out of adjustment and you can’t make it work better by changing the setting of the valve. You can make it different and you may think you’re helping but, you can’t make it better. Leave it alone and find out how Dolby is going to fix the CP650! I hope this all helps. I hope I didn’t take up a lot of your time with this long explanation. Lonny J.  |  IP: Logged David Stambaugh Film God Posts: 4021 From: Eugene, Oregon Registered: Jan 2002 posted 01-07-2006 11:24 PM                        quote: BTW. Splice caching was something that was developed years after the system was designed and after many processors where already installed and working. All Dolby Digital processors; CP650, DA20/CP500 etc. old and new are able to utilize this feature because all processors were upgraded with this new software using the Dynamic Loader feature! I posted something about that a long time ago, about how early Dolby Digital always dropped out at reel change splices, then at some point that problem went away. Nice to know now why it went away. Thanks for all the good info.  |  IP: Logged Rick Raskin Phenomenal Film Handler Posts: 1100 From: Manassas Virginia Registered: Jan 2003 posted 01-08-2006 08:47 AM                     quote: Lonny Jennings Imagine a large barrel. Agreed! This is analogous to the Generic Cell Rate Algorithm (GCRA) used in Asynchronous Transfer Mode (ATM) data systems. I have seen similar problems with MPEG video transmission over ATM where the video would freeze with one manufacturer's equipment but not another's. The problem was the FIFO buffer in the faulty equipment was undersized and unable to keep up with the incoming data rate; think overflowing the barrel. There were also other factors such prioritization of traffic classes but in a nutshell, the buffer was to small.  |  IP: Logged Scott Norwood Film God Posts: 8146 From: Boston, MA. USA (1774.21 miles northeast of Dallas) Registered: Jun 99 posted 01-08-2006 09:47 AM                        Wow. (sorry..had to)  |  IP: Logged Louis Bornwasser Film God Posts: 4441 From: prospect ky usa Registered: Mar 2005 posted 01-08-2006 09:56 AM                        Always good to hear from Lonny. (Info from the master!) Louis  |  IP: Logged Mark J. Marshall Film God Posts: 3188 From: New Castle, DE, USA Registered: Aug 2002 posted 01-08-2006 11:36 AM                    Well well well. Thank you very much, Lonny for that excellent explaination. I'm sure lots of folks who read it learned a thing or two. For what it's worth, we've been plagued with this problem for over a year. I originally mentioned it way back here, and then again more recently here. Does anyone have any engineering contacts inside Dolby?  |  IP: Logged Sam D. Chavez Film God Posts: 2153 From: Martinez, CA USA Registered: Aug 2003 posted 01-08-2006 12:53 PM                       Thank you indeed Lonny. It's about time you showed up. Lonny is probably right, he usually is. There is another possibility. What we think we are hearing and/or seeing is not always as it seems. There is an engineer at Dolby who is fond of saying, "if it can't be happening, it's not". When I first heard him say this, I thought it was denial, but what he was really saying was to stop and re-examine your premise entirely. As to contacts at Dolby, start with Norman Cerda, it's his job to deal with this and follow through to a satisfactory conclusion.  |  IP: Logged Fred Schoenfeld Film Handler Posts: 49 From: PORTSMOUTH, VA Registered: Aug 2004 posted 01-08-2006 11:10 PM                        For those who are not aware, Lonnie and Sam were with Dolby for many years. I have recently 'picked' Norman's brain at Dolby due to the problems I had with the CP 650 remote control lock up problem. He is very good at getting to bottom of problems and will return phone calls and E mails. Dolby is probably one of the best companies that suppports its end users.  |  IP: Logged Frank Angel Film God Posts: 5305 From: Brooklyn NY USA Registered: Dec 1999 posted 01-09-2006 02:22 AM                        Hey Lonny....shout out from Brooklyn to ya! Yes indeed, good to hear from you, and as always, it's like sitting in class with a Master. Thanks for uncovering that really elegant piece of engineering solution to what happens at reel changes when the splice jocky clowns are let loose to have their way with the prints. I was totally unaware of that 3 feet of data protection was built in on either side of the splice. I guess Dolby took one look at some of the returned test slugs and figured they needed to protect six feet of film at the reel changes. When Dolby demo-ed the digital system at the New York SMPTE section meeting, one of the things they explained was how they determined what film geography would be suitable to record the digital stream. They sent out 25ft lengths of black film to a few dozen theatres which was spliced into the trailer packs. After a few months, the samples were retreived and inspected for what areas sustained the least damage by the entire film transport system. They saw that one of the areas they originally were looking at had the worst damage -- the space outside the sprockets, (sorry Sony, you should have done your homework). I guess they also saw what kind of damage could be caused by lousy splices putting those slugs on and off the trailer packs and figured they'd better nip that potential problem right in the bud. Sweet. I am guessing 25 foot lengths of slugs when out and 20 ft lengths came back (bet some thought they needed to lop off two "ID frames" of the black slug every time they moved it), hence Dolby wisely added SIX feet of protection.  |  IP: Logged Sam D. Chavez Film God Posts: 2153 From: Martinez, CA USA Registered: Aug 2003 posted 01-09-2006 12:11 PM                       So what do you call the guy that hired the master? I never could see the point of hiring someone dumber than myself to be an assistant just for job security. Hiring Lonny allowed me to concentrate on other engineering and marketing issues.  |  IP: Logged Bruce Hansen Jedi Master Film Handler Posts: 847 From: Stone Mountain, GA, USA Registered: Dec 1999 posted 01-09-2006 06:53 PM                     Here is a question that comes to my mind: is there a test mode that allowes you to "hear" how stable or unstable the film is running? Could this 650 have been some how put into this mode? BTW, a CD player does the same sort of FIFO thing, but the FIFO has control of the speed of the CD, and will vary the speed of the CD to keep (that part of) the memory an average of half full.  |  IP: Logged Norman Cerda Film Handler Posts: 2 From: San Francisco, CA, USA Registered: Jan 2002 posted 01-09-2006 07:38 PM                     Aside from what Robert initially pointed out at the beginning of this thread, there has been one other case that reported this "Wow" effect during a reel(S)) splice 3 weeks ago or so. A few days ago we received and tested the Cat. No. 773 board and confirmed the issue at hand here (remember things move slowly during the holiday season). The Cat. No. 773 was replaced and all is working well. We are awaiting on another board so that we can do some diagnostics, but it is believed the problem might be a bad batch of SMT film caps on the 773, but is only speculation having only checked out one board thus far. We'll let you know more soon hopefully.  |  IP: Logged Mark J. Marshall Film God Posts: 3188 From: New Castle, DE, USA Registered: Aug 2002 posted 01-09-2006 07:58 PM                    Thanks for that, Norman. We have our regional tech coming out tomorrow to look at a few things including our 650 warble. I'll let him know that you guys are investigating something that might have something to do with our problem. If you like, I can email you his contact info so you can speak to him directly. Let me know. Thanks!  |  IP: Logged Sam D. Chavez Film God Posts: 2153 From: Martinez, CA USA Registered: Aug 2003 posted 01-10-2006 05:04 PM                       I received a call from Doug of Cinematronix in Calgary reporting a "wow" problem with the DSTR-20 reader and the 4 650's in the same theatre. I left a message regarding this with Norm at Dolby earlier today. This is getting really interesting.  |  IP: Logged

All times are Central (GMT -6:00)

This topic comprises 7 pages: 1  2  3  4  5  6  7

Printer-friendly view of this topic

Hop To: Select a Forum: Category: Operations -------------------- Digital Cinema Forum Film Handlers' Forum Large Format Forum Ground Level Feature Info, Trailer Attachments & REAL Credit Offsets Category: Community -------------------- Film-Yak Film Handlers' Movie Reviews The Afterlife Bob Maar's Joke-A-Thon Category: Classified -------------------- Equipment and Supplies Wanted/For Sale Help Wanted/Seeking Employment Category: System -------------------- Software Beta Testing Forum

The Film-Tech Forums are designed for various members related to the cinema industry to express their opinions, viewpoints and testimonials on various products, services and events based upon speculation, personal knowledge and factual information through use, therefore all views represented here allow no liability upon the publishers of this web site and the owners of said views assume no liability for any ill will resulting from these postings. The posts made here are for educational as well as entertainment purposes and as such anyone viewing this portion of the website must accept these views as statements of the author of that opinion and agrees to release the authors from any and all liability.