FlashAid – screen reader detection

Over at OSFlash they’ve released a tool which aids in the detection of a screen reader (before throwing them tons of AJAX or incompatible Flash). After a quick squint at the code it seems they’ve accounted for Flash not returning correct values for screen reader presence until a few seconds have elapsed, so it looks like good stuff.

Check it out here.

Vista and the Flash plug-in

It sounds like Flash may not work 100% as expected in Windows Vista unless a user’s using the latest version of the plug-in (9,0,28):

Senior Product Manager for Adobe Flash Player

Aral Balkan

It seemingly relates to a subset of specific functions which conflict with Vista’s security model:

  • Shared objects (Flash cookies)
  • Local Connection (Flash movies talking to each other)
  • File reference (saving and opening files within Flash)
  • Express Install (auto-updating a user’s version of the plug-in)

I haven’t a copy of Vista to check any of this unfortunately, although I’ve no reason to doubt it. It sounds like something that might only affect users upgrading from XP and retaining older versions of the plug-in, or developers, developers, developers…

Flash video best when divisible by 16?

I’ve been looking into whether better-quality Flash video is produced when sticking to dimensions that are divisible by 16. This seemingly has been the rule with older codecs, but I’ve had to investigate the issue for Flash 8 .flv video. Adobe doesn’t seem to think so, stating:

Although these ratios are standard, and should be used to avoid distorting the video, the size of the encoded video is not set in stone. The original web video sizes used heights and widths that were evenly divisible by 16. This was mandatory for many early codecs. Although this is not necessary for modern codecs, you should stick to even heights and widths.

This doesn’t quite tally with the advice on On2’s site (producers of the Flash 8 codec):

If you reduce your video size (image dimensions), your picture will be sharper. Another small factor you may consider is that optimal frame sizes are divisible by 8. That means that if your image dimensions (width and height) are divisible by 8, you will be wasting as few bits as possible on encoding data for portions of the video which do not actually appear. This is a complex issue relating to the way image compression uses 16×16 and 8×8 blocks to form an image.

A forum post by Zappu here enforces the 16/8 rule, as does this entry on MultimediaWiki. Conversations with Dave Curtis (hello Dave!) have been very helpful and also seem pretty conclusive on the “16 issue”.

It’s probably best to stick to dimensions divisible by 16, but not set in stone. Complex clips with lots of motion and/or panning would probably benefit from the 16-rule to a greater extent. I hope to produce some encoded tests myself, but if anyone has any in the meantime the links would be much appreciated.

Why is there so little clear information on this on the web?!?

wiiminder.com

While the Opera browser on the Nintendo Wii’s still in Beta, this looks like providing a few additional widgets to make surfing a better experience.

In case Nintendo are reading this (as if!); please, please, please enable as much Flash/JavaScript as possible in the final version of the browser. It makes sense if your games release schedule is as slow as it is currently, as you’ll get a lot of homebrew games and activities developed by the masses that’ll tide people over in the meantime.

Being able to capture the full range of wiimote movement and button presses in Flash would lead to some amazing content being developed, and it wouldn’t matter that the PS3 browser features Flash 7 too if you open up the API. Then the developer community wouldn’t have to resort to workarounds like Mario’s great efforts.