2 new shows and timing

[Lloegyr]

I have for the past few weeks been working on a show and it seems appropriate to give some background because it is all about timing.

The show like my Fields of Fire is in fact quite old, but the discovery of PTE has started me going all over again. (There will be a new show about Prague after I go in May)

I started programming the show on PTE V3.8 on a P4, 1.7, then more and more betas starting arriving. I did not want to confuse matters so I installed V4 beta 5 on a different machine, a PII, 333 Celeron, and continued the programming on that machine. Programming turned out to be far easier than before, due to two factors, 1 preview from any point, and 2 being able to zoom in on the timeline (excellent upgrades). When finished and the show was in perfect synch I was very pleased until I played it back on the first P4 machine, it was 1.1 seconds adrift. I then tried it out on another machine, a PII, 333 normal Intel, it was also 1.1 seconds out. I took the easy way out and added the 1.1 secs on the celeron machine which then meant that that machine was out, but both the others were spot on. I then started to experiment.

I then redid the show on the normal 333 Intel machine, and it then plays ok on all 3 machines. The only difference between the 2 slower machines being that one is a celeron and thus has less cache memory, CPU speed is the same and so is RAM. (any thoughts on this aspect?)

I would be interested to hear if the show runs correctly on other machines, the ending will give it away if its out.

The 2 shows are slightly different in length and have different synch problems, also the French soundtrack was originally 19 seconds longer, I cut 6 seconds from the introduction only then changed the speed by 5% without changing the pitch (Cool Edit 2000) to give me a similar time to work with.

By the way I had heard the English words first and thought I have to do a show to that, unfortunately 22 years ago it was impossible to get the record in England. However when I was in Berlin and saw this carousel I had to take the pictures and keep them ready for the day I got the music. I did 8 days later on my final day in Berlin! Since then I have bought Jacques Brel’s records wherever I could and my biggest source was La Reunion in the Indian Ocean. This year is the 25th anniversary of his death and I would like to go to some of the celebrations that are going on. These 2 shows are my bit of celebration. Thanks to Igor’s excellent program

Mike

Mersea Island

[HaroldB]

I installed V4 beta 5 on a different machine, a PII, 333 Celeron, and continued the programming on that machine. Programming turned out to be far easier than before, due to two factors, 1 preview from any point, and 2 being able to zoom in on the timeline (excellent upgrades). When finished and the show was in perfect synch I was very pleased until I played it back on the first P4 machine, it was 1.1 seconds adrift. I then tried it out on another machine, a PII, 333 normal Intel, it was also 1.1 seconds out. I took the easy way out and added the 1.1 secs on the celeron machine which then meant that that machine was out, but both the others were spot on. I then started to experiment.

I then redid the show on the normal 333 Intel machine, and it then plays ok on all 3 machines.

I'm very curious about this.

When you say that it was 1.1 seconds "adrift" when you initially played it on the P4 machine, was the slide show itself playing faster or slower (ie, which *direction* was it adrift in)? It sounds from what you wrote that the music was too short. Did you try moving PTE and .pte file without the extra 1.1 second of soundtrack back to the P4 and preview it in the timeline and see if it drifted slowly out of alignment as the slide show progressed or there was some problem with a slide or two? (This is so weird that maybe it is a problem in the music players on one of the machines? You're not using the internal player, I presume.)

Also, when you say that you "then redid the show on the normal 333 Intel machine", do you mean that you simply moved the .pte files and created the EXE on the normal 333 Intel machine, or did you have to retune the synch?

For synched shows, the only problems I have had is that when it executes on a slower machine and I am using music that has sharply defined beats that I use as "transition points", sometimes the transitions are a tiny bit "off", presumably because it takes longer to load the slides. OTOH, that doesn't sound at all like your problem. (I've also had the stuttering problems when the slower machine can't load the slide fast enough to honor the specified transition time, but that seems like a different matter.)

(FWIW, I think that the only way this might be controlled was if PTE loaded the "next" slide from disk while it was displaying the current slide -- not when it was "ready" to display the next slide, but as soon as it was finished with the transition of the current slide. For all I know, it perhaps does it that way now.)

Harold

[Lloegyr]

Hi Harold

Thanks for your reply.

The slides ended 1.1 secs before the music. It did not actually drift slowly out, all my check points are still spot on, it is just during the final section, which is complex, but did run fine on the slow creation machine. No not using internal player.

Just moved the pte files and remade the exe. I had just added 1.1 secs to the end of show on the old slow machine, so it then ran correct on the other machines, but overran on the origination machine. It just seems to be if I create on this slow machine, actually it’s the one I connect to the net too, so it is separate from my major machine that this timing variation appears. If I create on my fastest machine, then its fine on any machine. My Fields of Fire which is 16 mins long and has 5 precise and specific transitions in the last 27 secs runs perfectly ok with no drift at all on all machines but was made on the fastest machine.

It’s quite bewildering, obviously it’s not a problem for me as I have the facility to make shows on several different machines, but it does intrigue me. I thought it an interesting discussion point. I still think that the celeron with less cache memory on the cpu seems to be the major cause.

Many thanks

Mike

[HaroldB]

Mike,

I would love to hear what Igor has to say about this.

It SOUNDS to me like this might be a manifestation of the transition time/slide duration time problem within PTE, but your facts don't jive with the way the problem typically presents.

From what you say, it sounds like the timing in the last few slides must be off. If it doesn't drift "gradually", then something must be going off *somewhere* to cause the slides to end early. I'd try figuring out exactly where that happens and then try fiddling with the setttings there to see if you can figure out what might be causing it.

Harold

[LumenLux]

Mike I just started my morning by running your fun Carousel shows. English or French, both are effective. I was caught up in the fun of the music and pictures and did not think to worry about the technical timing issues. It certainly had no obvious problems to me. Fun work. Thanks. Ending was simply with black screen remaining after sound - until I pressed escape.