SOME BUG??? automatic movement - different speed while showing on laptop vs. desktop

[Virtualguide]

Hi boys,
I am producing some another virtual tour with 37 panoramas. Like in my former presentations, also in this i am using automatic movement and showing each of panoramas one after another...

I prepared the code on my notebook, however after I tested in on desktop I noticed that the movement of panorama is about 3x faster. And this is big problem.
Look at a part of my code:
move1-0a="
timer+=1,2400,,move1-0b
pano.tilt=8,2500,smooth,
pano.zoom=0.73,2500,smooth
"
move1-0b="
timer+=1,3700,,move1-0d
pano.pan=46,4000,smooth,
pano.tilt=5,4000, smooth,
"
move1-0c="
timer+=1,2000,,move1-0d
pano.zoom=0.95,1800,smooth,
"
move1-0d="
pano.zoom=0.73,2500,,
timer+=1,9300,,move1-0e
pano.pan=253,9500,,
pano.tilt=-5,9500,,
"
move1-0e="
timer+=1,2900,,move1-0f
pano.pan-=16,2000,,
pano.tilt=-5,2000,,
pano.zoom=1.55,3000,smooth,
"
move1-0f="
loadpano(panoName=mov/ext4&pan=-144&tilt=-7&zoom=1.20,1000,fade);
onTransitionEnd=move1-0g
"

move1-0g="
timer+=1,3000,,move1-0h
pano.zoom=0.73,3500,smooth,
"
..... etc...
"

something is wrong. if I tell the player to pan somewhere in 9,5 secs, it shood be 9,5 not faster !? (I understand that the oldier computers would do this slovwlier, but not that the powefull ones should do this several times faster???)

Has somebody of U such experience?
I am going to test older pano.swf now, I think I did not have such problem in the past...

[Virtualguide]

SO... seems, it does not depends on pano.swf version. I changed the latest version for the one from 27/10/2007 and nothing has happened...

Hans, Scotty,... the others??? What do U thing?
Did somebody of U make something like this? Can U test the speed of movement on two different computers?
Thanx...

[Scott Witte]

Igor,

I've never considered timed tweens all that precise in FPP. That said, can't say I've ever noticed tweens moving noticeably faster than specified, only slower, FWIW. Stumps me.

[Trausti Hraunfjord]

Can you upload a copy of this project for testing?

[Virtualguide]

Thanx boys,
so I have uploaded some files here:
http://virtualguide.sk/test/auto/
here is first part of the working code, or check the xml:
move1-0a="
timer+=1,2400,,move1-0b
pano.tilt=8,2500,smooth,
pano.zoom=0.73,2500,smooth
"
move1-0b="
timer+=1,3700,,move1-0d
pano.pan=46,4000,smooth,
pano.tilt=5,4000, smooth,
"
move1-0c="
timer+=1,2000,,move1-0d
pano.zoom=0.95,1800,smooth,
"
move1-0d="
pano.zoom=0.73,2500,,
timer+=1,9300,,move1-0e
pano.pan=253,9500,,
pano.tilt=-5,9500,,
"
move1-0e="
timer+=1,2900,,move1-0f
pano.pan-=16,2000,,
pano.tilt=-5,2000,,
pano.zoom=1.55,3000,smooth,
"
move1-0f="
loadpano(panoName=mov/ext4&pan=-144&tilt=-7&zoom=1.20,1000,fade);
onTransitionEnd=move1-0g
"
However I noticed that the problem is not with the speed of my desktop but with slow notebook... I have tried on another desktop with dualcore intel E8400 + 3GB RAM and it works little slowlier, but only very little difference is there between two desktops.

I carefully measured the times and found that actually the desktop shows the presentation right way...
I wander how is it possible that 1-year old notebook with 2GHz dual core, 4 GB RAM of memmory has such a problem?
Graphic card? I do not know much about flash player, does it utilize GPU for accelerating our flash panos? Is there some difference between playing the same tour on two computers with different graphical cards, I mean with different support OpenGL/DirectX ?
Thanx...

[Scott Witte]

Igor,

Tween motions are heavily dependent on graphics card (GPU) speed. Most laptops and pedestrian desktops will have less powerful GPUs and tweens will be slower, sometimes painfully so.

In your case the problem may be compounded by your qualityStatic setting. Tween motion leaves the stage quality at the qualityStatic(2) setting. You have qualityStatic=medium and qualityStatic2=high. That puts a huge load on the GPU during tweens and is bound to slow all but the most powerful.

For that reason, with rare exception, I always change qualityStatic to low during tweens and back to a higher level when I stop tweening. Unless the tween is really slow, you rarely notice the change in quality, anyway. Try that and most of your problems should be solved.

Now, to complicate matter more:

1) You are using FPP2.2 which is helped little if at all by Flash10. Switch to the current FPP2.3.1

2) In FPP2.3.1 there is effectively little or no difference between stage qualites of medium or higher. Denis says it has something to do with the way Flash10 handles AA and he wasn't sure how to change that. I'm hoping it all gets sorted by FPP3.

3) Unless you are using hotspots2.3b2 (available from the Google Group files section) qualityStatic2 has no bearing except when initializing FPP. Changing qaulityStatic changes stage qaulity whether you are running Flash 9 or 10.

If you use hotspots2.3b2 changing qualityStatic only changes stage quality under Flash9 while changing qualityStatic2 only applies if running Flash10. So, if you want to change stage quality you have to specify both a qualityStatic and qualityStatic2 setting in you function. They can be different to accommodate the different abilities of Flash 9 and 10. I thought that was a good idea when I asked Denis to make this change. After testing I changed my mind. The extra GPU strain on most machines just wasn't worth the barely noticeable quality improvement in Flash10 after all.

My recommendation is don't use hotspots2.3b2 should you be tempted. Programming will be simpler.

[Astroman]

Yep. Scott is absolutly right.
Flash is a hardware resource bitch and it also hates MAC otherwise check out a tour on MAC with lensflares vs. tour on PC with lenseflares. Btw. the lenseflare plugin cld need an update as we all know ;).

[Scott Witte]

Igor,

BTW, fascinating tour!

[HansNyberg]

The most important factors are display size and processor

I did some tests on my MacBook Pro 2.16 dualcore.
Time measured from 100% loading completed to the next pano started loading.
Initial size = 32 sec
Fullscreen 1680x1050 = 74 sec

Also browsers can be different.
On my PPC 2.0 dualcore G5 the inital size took 114 sec In Safari and 60 sec in Opera.
That was without anything disturbing it.

With another process taking 45% of the maximum 200% processors Safari took 180 seconds to run it.

Selfrunning tours like this has to be adjusted to the computer which you want to run it on.
I am sure the 32 seconds is not the fastest timing you can get but I am not sure I would want it to run faster.

Hans

[Scott Witte]

Hans,

You are certainly right about the difference between windowed and full screen. I've always taken that as load on the GPU. The more pixels it has to push the bigger the load. And the more AA it has to do (aka stage quality or the qualityStatic setting) the bigger the load.

Other than dropping the stage quality to low during tweens, I don't know of an easy solution. One that could be programmed would be a periodic test for FPS and a conditional adjustment of stage quality based on that.

TOMMY? We need another plugin ;)

[Virtualguide]

Thank U very much boys!

Scotty: I had older pano.swf only for testing there, I forgot to change for new one. But seems that I found the problem with my notebook. In power options in Control Panel was checked "balanced" and seems that this affects the performance of CPU (maybe graphic card also?). after checking "high performance" the speed is almost the same like on the desktop.

However I tried to play the tour on some other machines, and it is too slow on older ones... therefore I found your advices very useful. Thanks a lot. I am going to play with quality settings with motion twens...

Hans: thanx for testing, it will be offline CD presentation for everyone so probably I will need produce 2 versions :-) As for the speed - it too fast now because I coded it on my laptop before I found the difference... :-)

I am going to play with it.

[Virtualguide]

Scotty, as for hotspots, I do not remember now and am lazy to test it now, but I think I started to use this last version because of whirling hotspots? I do not remember now:-) Can somebody help me to recall what is better with it? thanks :-)

[Scott Witte]

Scotty, as for hotspots, I do not remember now and am lazy to test it now, but I think I started to use this last version because of whirling hotspots? I do not remember now:-) Can somebody help me to recall what is better with it? thanks :-)
Hotspots2.2.1, the latest distributed version, is the one to use, IMO. It takes care of whirling hotspots, etc. V2.3b2 only has the qualityStatic2 modification I described earlier. If you don't have it already v2.2.1 is in the files section of the Google Group.

Scott