I asked this question in April, but it received only one useless response. Nothing has happened since then, while newer versions of Firefox have been released.
I first thought the problem was combined with Mac OS (mine is 10.6.8), but the same problem of blank frames is noted in Firefox 22 on Windows 7.
You can create a bug report to see if there is a solution possible as this is apparently a regression.從原來的回覆中察看解決方案 👍 2
What issue do you have while running it with Firefox? I am able to play it fine using Firefox. To better troubleshoot, please make sure version 22 is the version you are using.
Try disabling graphics hardware acceleration. Since this feature was added to Firefox, it has gradually improved, but there still are a few glitches.
You might need to restart Firefox in order for this to take effect, so save all work first (e.g., mail you are composing, online documents you're editing, etc.).
Then perform these steps:
- Click the orange Firefox button at the top left, then select the "Options" button, or, if there is no Firefox button at the top, go to Tools > Options.
- In the Firefox options window click the Advanced tab, then select "General".
- In the settings list, you should find the Use hardware acceleration when available checkbox. Uncheck this checkbox.
- Now, restart Firefox and see if the problems persist.
Additionally, please check for updates for your graphics driver by following the steps mentioned in the following Knowledge base articles:
Did this fix your problems? Please report back to us!
I watched it for a while. My wireless is too slow to wait for all of it. It seems that many images may be displayed before they are fully downloaded, so they display as partial. Is that what you see?
Windows 7 x64.
I don't think you understand how the movie should look like. There are too many blank frames if I run it using Firefox 17 onward. Please run it using Chrome or IE, and make sure you the movie runs in the same way.
I think you have started to notice the problem. That seems to be the explanation, i.e., images are displayed before they are fully downloaded. But note that this does not depend on the network speed. According to http://www.speedtest.net/, my download speed is 26.66 Mbps, and yet this annoying problem (only with Firefox, not IE or Chrome) consistently happens. I think Firefox is broken since V 17 in this regard.
Hi halocme, I didn't study the script in detail, but if the img.src is simply changed on a timer without knowing whether the image had downloaded fully, you run the risk of displaying partial images. I'm not aware of any changes that would affect that, but I haven't researched it.
What this script seems to do is start 369 image loads. When a load completes, it increments a counter of completed image loads, and then uses the counter as an index into the list of images to set the image source in the page. In order to work as intended, the images must finish downloading in the order specified in the array, i.e., in the order requested. I hypothesize that Firefox is not hewing precisely to that order, or for some other reason the images are not completing loading in that order, and therefore the script is setting src=an image that has NOT completed loading rather than the one that actually just completed loading.
Edit: I should clarify that I did not wait for all 369 images to download, so I don't know whether the animation runs smoothly once all the images are in hand. I was only watching the loading behavior.
由 jscher2000 於 修改
No matter how the images are loaded now, the movie runs perfectly well under Firefox V15. So describing possible defects of the script does not help. Please analyze what made Firefox inferior to Chrome and IE since V17.
Current Firefox 25 Nightly builds have a nice Network monitor with a graphic representation that allow to follow exactly how long it takes to load the images.
It shows for me: 370 requests, 335,970,79 KB, 97.02 s
I need to slow down to about 11 frames (24 or faster is really bad and 16 shows some flicker) to make it display smoothly.
I'm on Linux, so there is probably not much hardware acceleration.
A short range of about 80 frames works well with over 180 frames/s in swing mode.
Hi halocme, you can search MDN and Bugzilla as well as I can to see whether you can figure out what changed.
For the reasons I mentioned, I'm not convinced there is a problem with Firefox rather than a problem with the script.
Does that mean Firefox V25 is much better than V17-22? Honestly, it does not sound as encouraging to me as I will even upgrade to it from V15.
I am still surprised that Firefox V15 has no problem running the entire movie. Same is true for Chrome and IE. I still wonder what went bugful in Firefox V17. I hope Firefox is at least as good as Chrome and IE.
I am a user, not a developer, so those documents do not sound a bell. Please explain what you find there.
Again, your attack of the script, no matter how unprofessionally it may be written, does not help at all. I just want Firefox to be at least as good as Chrome and IE, or even Firefox V15 that I have downgraded from V17 because of this problem.
It works for me in Firefox 22 if I set these both prefs to false (a restart is necessary):
So this looks like a problem with decoding the images that Firefox isn't doing fast enough.
Firefox 23 and later crash for me if I set image.mem.discardable to false.
Current Firefox version seems to keep less images decoded and free up memory too soon.
If I use the Swing mode then I can clearly see this: for a short while after the reverse start the animation is smooth until Firefox needs to decode images once again.
With the current cache setting it work to about 80 frames.
What me worries more is that flipping the prefs to false makes later Firefox versions crash.
Crashing usually indicates either memory corruption running out of memory and overwriting other memory or just not being able to allocate more memory.
There are also a lot more image prefs present in those later versions, indicating that there have been a lot of changes.
You can create a bug report to see if there is a solution possible as this is apparently a regression.
Thank you, but could you show me how to submit a bug report?
I downloaded 23.0.1 this morning, and found that this version resolved the issue that I had raised. I am very pleased that the resolution came without a bug report. After reading the instructions on how to submit a bug report, I felt I was not qualified to submit one.