All my example code (FlashBuilder 4.5 project) is here:  LINK
I’m putting the project up instead of links to examples online because I dont want bandwidth issues to play a part in anyones testing, and the clients stuff is all loaded locally as well.

So I started looking at Stage Video about a week ago.  I didn’t touch it at all during the beta period.  Now, my client was interested in some baseline tests that mimicked some of their use cases.  During my tests, I immediately starting running into little issues, that really got me frustrated.  I’ve tweeted a few of the problems I’ve encountered, but at this point, I’d have a hard time recommending Stage Video to my client.

Issue #1:  Looping of Stage Video [ OSMF Looping fixed with OSMF 1.6 Sprint 5]

My first attempt with SV was extremely basic.  I just wanted to play a video and loop it.  Easy enough, right?  You’d think.  I set up my SV example pretty much just like I would when using the “old” way with NetConnection and NetStream, listening for the NetStatusEvent “NetStream.Play.Stop” and then seek(0) and have it loop.  In all my tests, I got a noticeable “blink” when the video would loop.  I never noticed any blinking of the video if I didn’t seek, or if I paused and did a seek, but if it was just a seek(0), and the NetStream was playing, “blink”.  I even tried pausing, seeking and then resuming, and still got a blink.   Two interesting tidbits:  1) If I set up a timer to watch the time, and manually looped it about .5 second before the end of the stream, it worked fine, but didn’t work if I tried to manually loop it right at the end, instead of waiting for the NetStatusEvent   2)  According to a response to my original tweet, this is fixed in OSMF.  Great.  really? I have to use OSMF to loop? Okay.. I’m game. . whatever.  I’ll be honest, I probably should look at what OSMF is doing under the covers to get around this, because the moment I moved over to using OSMF 1.6 Sprint 3, it was able to loop w/o blinking.  Problem solved, right? Just use OSMF 1.6..   Yeah.. not so fast….

[Update]

I was able to confirm this blink on my 2010 iMac, 2011 MacbookPro and also a quad-core Dell running Vista.  All the computers were running the latest FP 10.2.159.1, except the iMac which is running the latest RC of 10.3

I put an example here on line with the test showing the blink:
http://www.yapiodesign.com/test/stagevideo/StageVideo_LoopBlink.html

I also put a Quicktime capture of the blink on my iMac:
http://www.yapiodesign.com/test/stagevideo/StageVideoBlink.mov

[UPDATE - June 16th]
With the latest OSMF 1.6 Sprint 5, the issue of the video blinking when looping has been resolved.  One issue down, more to go!

Bug logged: Flash Player JIRA #6806

Issue #2:  Starting OSMF using Stage Video after a Mouse Click

Yeah..  the title should confuse you. What?  A mouse click?  seriously.  After I got over the blinking issue using OSMF, I was setting up tests and using buttons to display the various tests.  All the sudden, my videos were playing audio, but not video, until they looped around once, and then the video played on the 2nd loop.  WTF?  If I forced OSMF to not use Stage Video, things worked just great. No problems.   If I took out the button control and had the OSMF Stage Video test start after ApplicationComplete event, things worked just fine.   Add in a button and things got weird.  I’m completely stumped WTH is going on here.   I’ve narrowed down the issue ( I THINK), to autoPlay = true, and loop = true.  If I set autoPlay = false, and then mediaPlayer.play() things work.    I’m going to chalk this up to a bug in OSMF 1.6 Sprint 3 and I’ll be submitting a bug to the OSMF Jira once I post this and post the Bug report # here.   OSMF 1.6 is beta. I get that.

Bug logged: OSMF(Strobe) JIRA #1297

Issue #3: SerialElement

Ok.. This isn’t related to Stage Video, but while I have your attention….here goes.  According to Adobe, the Serial Element definition includes “The media elements that make up a SerialElement are treated as a single, unified media element.”.  So my silly thinking would say that using a SerialElement that OSMF would handle making sure all those elements are ready to go when needed.  Turns out I’m wrong.  As OSMF encounters a progressive media element as part of SerialElement, it loads it when it encounters it, and doesn’t preload it at all.  So if you have 4 child elements in your serial element, child 2 is only loaded once child 1 has completed.  I’m using local assets and its bad enough, so I imagine loading from a web server is only that much worse.  After doing a few google searches, turns out that is the intended behavior.. or that’s what I read on an Adobe forum.  I’m sure I could dig up the link if anyone is interested.  But what a silly thought.  Who would want that?  So my answer is to preload all the elements myself.  Luckily, I found a solution buried (yes, buried) in the OSMF examples.  …and guess what? It works fine when using “regular” OSMF, but failed me when I was trying to use OSMF 1.6 with Stage Video.  Great..   All I’m asking if for Adobe OSMF team to a) add the ability to tell OSMF to preload the SerialElement child elements….and maybe in the process, take care of the duration issue as well. If I have to handle preloading the child elements myself, and have to tell OSMF the default duration of each child media element, I’m not really sure why I’m using Serial Element at all.

Issue #4:  SerialElement Looping

Ok.. so after all the drama of #3 above, I said hell with preloading the SerialElement child mediaElements, and just told OSMF 1.6 to loop the SerialElement using Stage Video.  I got pauses between the child media elements, and then it looped, and then things got very weird.   I set up  4 child media elements as part of my Serial Element.  The first time around, all the videos played. Once the loop started, it played child 1, but nothing displays for child 2, 3, 4.. loops around plays child 1, skips 2, 3, 4..and repeat.    Ok.. Another OSMF Jira bug going to get logged.

Ok.. appears this one may already be logged:  JIRA # 1226

Issue #5:  Stage Video Hardware Decoding and Rendering

What is the secret sauce to get Stage Video to use hardware decoding?  I know the hardware requirements on Apple seem to be limited to nVidia GPUs.  So that rules out nearly every top of the line Apple system since Apple has switched to ATI.  Doesn’t matter who we place blame on for this, but not having hardware decoding sucks.  Ok.. on to Windows.  I tried my wife’s fairly new Dell quad core with Windows Vista (we didn’t’ get Win7 b/c her office hasn’t upgraded).  No dice.  Hardware rendering.. no hardware decoding.  Okay.. I’ll try one of the Linux boxes at my clients office.  yeah. dont laugh…I know.. but for kicks, right? You guessed it, didn’t work.   So I really have no idea what magic secret hardware sauce is required for users to get hardware decoding.

Where I’m left at this point:

Even after all the issues above, I’ve managed to find workarounds to at least get my tests working.  Stage Video, either when using it as-is, or with OSMF is definitely not a drop-in replacement.  I put part of that on OSMF 1.6 still in beta. I’m sure it will get better (hopefully).  I have no idea what hardware is required to get hardware decoding, and I’m not even sure if it would solve the video tearing I was seeing when I was using it. Right now, I really can’t see myself recommending Stage Video over my clients home-grown Linux Flash & video solution.  It would sure be nice to use Flash for everything, but that’s not a reality at this point.  So I’m going to log those bugs in OSMF Jira, keep watching twitter and the Adobe blogs for updates, and maybe we’ll get there eventually.

Side note (and a little more ranting):  We really need better examples out there for OSMF and Stage Video.  When OSMF first came out, the examples were overly simplistic, or buried deep in the OSMF svn.  Documentation was lacking.  Now we have StageVideo , and we have some great getting started examples, but I believe if there were anything beyond the extreme basics, some of these issues would have already been spotted.

Thanks for reading and providing any insight into workarounds.  I’m really interested to see what sticks I need to rub together to make hardware decoding fire.  Who knows…. maybe I’m holding it wrong.

To re-purpose a line from Steve Jobs, the Android UI is a “bag of hurt”.  Where Apple iOS is very controlled (I agree, too much at times), the Android OS seems to be all over the place.  You can hand an iOS device to a 3 yr old and they’ll be moving around and finding things very quickly.  I probably spent two minutes with my niece over the holiday and she proved that point.  There are still times when I’m using the Droid 2 where I’m asking myself “WTF?” and I can’t find something.

Case in point, open the Android camera app.  You have some settings controlled via on-screen slide out (scenes, effects, flash, switch to …), but then other settings are only accessed by hitting the hardware settings button (picture mode, tags, settings – yes settings within settings). And you can actually open both at the same time so they overlap. Pure UI Craptasic Awesomeness. This isn’t some 3rd party app either. Its the built-in experience on the Droid 2.  Then on top of that, just the basic UI feedback mechanisms like screen animations like slides feel clunky. Maybe its better on other Android phones.. not sure..  But this isn’t the only examples.   iOS is far from perfect, and Android is definitely far from perfect.  Android is free, open source, easy to develop for, and I’m sure will lead to more work, particularly for the Flash developers using AIR, but from a consumer perspective, its not a long way to go.  I hope 2011 brings big improvements.

[ edit on Jan 18th - a friend questioned my judgment on why I came to my conclusion on Android. Figured I'd post it here]

Here is a list of things I found clunky in Android:  Btw — iOS isn’t perfect, but it’s the little things that it really excels at.

Unlocking the phone:  The combination of haptic feedback and the sliding animation was jittery when sliding in a normal fashion.  The haptic feedback actually amplified this feedback experience.

Mute icon on locked screen:  This is a sliding metaphor, yet the response to set audio on or off.

Overall:  the combination of hard and soft UI is confusing. Many times, there are onscreen controls and controls accessed via the settings button.  A great example of this is the camera app on the Droid 2. You have a slide out menu for some settings onscreen on the right to access some camera features, including switching to camcorder, and yet, you have other important camera features, like camera modes, hidden away in the settings button.  (fyi I tried to find out if this is the stock camera app, and I really dont’ know but I did check that Omar has the same on his phone).

Settings button:  When does any app/screen use this or not.  I found myself routinely pressing this to see what is hidden away.  It surprised me how often things were hidden away w/o any feedback to the user that things existed there.

Back button:  I like the back button, but its not consistent.  Some popups will use OK, Cancel, and some won’t, expecting you to use the back button.
Portrait to Landscape (& vice-versa) transitions.  This is probably more of a personal preference, but the immediate snap in orientation is jarring.  This phone is far faster than my 2-1/2 year old iPhone, so there isnt’ a technical reason that it can’t change orientations with visual feedback to the user.  Transitions serve a very important function of user context.

Overall:  The phone doesn’t come with any music or video organization tool.  Some people hate iTunes, but it serves as a great tool for organizing your music, videos, and apps.  These aren’t always going to exist only on the phone.  This is why I pointed out iTunes+iOS.   It’s a package deal. And the home user wants a way to manage these items.  Instead, at least from what I see, Google & the carriers leave that up to the customer to figure out.
Notifications: I like the idea of having one place to get to your notifications.  I think Apple should adopt something like this.  However, getting to your notifications on Android is horrible.  That tiny top bar is about 1/2 size of the normal height needed.  Numerous times, I have fat-fingered it, and missed and had to try again.

List items that go somewhere, List items that just display:  Go to the About Phone screen.  This screen is a big list.  Some of the list items only display info, and some are triggers to new screens w/o any visual feedback that clicking on them will go somewhere.  iOS is very consistent with using “>” on list items that will go to more detail.  In the sound settings list, they actually use the “dropdown caret” icon.  Although, when in Contacts,  there is a similar caret icon for linked profiles, that doesn’t go anywhere, but instead drops down more information.

Android Market online:  why is this only usable when I’m on the phone?  Go to the market, find an app you are interested in and there’s no price, no Google desktop installer, no QR code even.  AppBrain does this better than Google.

Input Textfields: This may be something that Verizon or Motorola did, but why the red highlight around the active text field when I’m inputting text.  Red means bad.  Either pick another color to draw visual interest or only use the red when its meant to be used.

And I could probably go on.  I doubt most of these are only Droid 2 issues.  Your response seems to be that because it has better features, that it makes up for a lesser experience.  Maybe thats the case with some people, but I would seriously doubt that most people feel that way.  Here’s a good quote from Alan Coopers “The Inmates are running the Asylum” –> “Power User is a code name for an apologist. Regardless of how hard an interaction is, or how uselessly obscure a feature is, the apologist will unerringly point to the power and functionality of the gadget, blithely ignoring the difficulty of actually using it”   Apple “gets” this concept.  Google does not get it yet. I think voice commands is one step in the right direction, but that’s a good solution to a narrow circumstance (driving while using your phone).

Ok.. some quick thoughts on web and mobile for the year:

H264 vs WebM: Yesterday, Google announced they were dropping H264 support from Google Chrome.  More power to them.  They paid for and released WebM as open source, and they have a vested interest in seeing it taken up as the new web video format.  Google likes to make money (and they have lots to be made on YouTube and other areas) and not spend it.  There is a lot of distrust for the MPEG LA even though they said there would not be a license fee for end-user and only for manufacturers building encoders/decoders.  Apple and MS are firmly behind H264, and its being used everywhere.  WebM on the other hand is free, free, free.  I have no idea how this will pan out, or if it will ever.  So if you are targeting video playback everywhere, be prepared to encode and host multiple copies of your video source. And you are already probably doing that or using  Flash Player for clients that dont’ support H264.

Android vs iOS: Android will grow and grow and eventually be everywhere that iOS is not.  Does that mean Androids win?  Far from it.  iOS isn’t going anywhere, and its not like Apple is going to sit still.  Both iOS and Android will continue to “battle”, but ultimately the consumer and your clients will pick one or the other based on personal bias, whats cool, whats cheap, what has the game they want, whatever.. and you’ll be designing or developing for one of them, or both of them.  This is the new Mac vs PC.  Both have benefits, both have disadvantages, but both are viable solutions for both end user and developer.  I look at Android and iOS like having dinner at your relatives house.  Android is like eating at your cool uncles house where everything goes, but its always a bit messy. Yeah.. he’s got the cool toys, playstation, xbox, but you keep uncovering little tidbits of dirt and unpolished gunk in the corners.  Dinner with iOS is like being at your aunt’s house who is always prim and proper.  Full of rules.  You feel confined at times, always afraid to do something wrong, but when you look around, you had an awesome meal, got some culture, and still had a good time.

RIM Playbook and other tablets: Lots of cool tablets were shown off at CES this year. Most of them were running Android Honeycomb, and then there was RIM.  The RIM Playbook is very interesting to me.  Its hardware specs are very good, the price will supposedly be sub-$500, and the OS (QNX) is running Adobe AIR.  As a Flash Platform developer, that makes the Playbook a near perfect device, right?  Potentially.  However, I don’t have faith in RIM yet.  The Playbook will be marketed towards business. It has the strong suit of RIM and they have a good foothold with their Blackberry.  But I believe to be very successful, you need to embrace the consumer market, and I just don’t know if RIM has the experience or desire to do that.  Unless they are devoted to this, I don’t see it growing organically. I could be wrong, wouldn’t be the first time (or the last).  As for all the Honeycomb-based tablets shown at CES, like the Motorola Xoom, very impressive hardware and they should sell well. The total success of the tablet market will come down to apps. Phones will always sell because people need to make calls. Tablets, on the other hand, needs apps. They are mini-computers and will be used as such. If we start to see really great apps for Honeycomb that aren’t just re-sized phone apps, and start to see some big names get behind app development, then I think Android-based tablets will take off.

Apple:  I expect Apple to continue doing what its always done. Take a bold hard line approach to its design and development of new hardware and software.  You either love them or hate them. Seems most of that split seems to fall around what you think of Flash and HTML5.  Apple isn’t going anywhere, and they’ll continue to sell their products and provide a very good consumer experience.  I would have think that 2011 will bring us a new iPhone, some advancements with iOS (4.5? 5.0?), and of course, OS X 10.7 Lion.  The more I look at what Apple is doing with Lion and iOS, the more I think we’ll see a convergence soon enough, at least with software.

I love to bitch and complain about Adobe and Flash.  I do this because I’m passionate about what I do and Adobe and Flash are a big part of my professional life and I always want them to be better.  I also love my Apple products.  None of them are perfect, but I have a room for each in my career.  I see 2011 being the year that I get off my butt and start doing some real iOS development, instead of turning the work away like I had done in the past.  I also see myself continue to do Flash work for both the desktop and mobile.  All this “fragmentation” is opportunity.   Enjoy 2011.

Flash Catalyst isn’t even in public beta yet, but based on the potential I see with the preview release handed out at Adobe MAX, I could see Flash Catalyst replace the Flash IDE eventually.

Now currently, FC is being billed as a tool for interation designers for creating UI mockups to use for wireframing, adding interactivity to your AI, PSD or FW files, getting client sign-off and handing off to your developer. The big goal is that behind the scenes, FC is built on Eclipse and actually is building a Flex Project file (FXP actually) to take into Flex Builder and wire up to live data. Talking this idea over with others, I’ve heard some same that isn’t the focus of Flash Catalyst or its not intended for that, yada yada.  But when FutureSplash hit the market, who thought we’d be building web applications with it eventually?

Looking very briefly at both Flash CS4 and Flash Catalyst while wearing the “designer hat” , in both tools I can draw, import from existing Photoshop, Illustrator, and Fireworks files, animate using the more intuitive After Effects-like timeline , add interactivity via built-in features/behaviours and even go “behind the scenes” and add more advanced interactivity via ActionScript. I can easily see users looking at FC to build what they currently use Flash IDE for. Even Adobe’s own marketing on the FC site states “Flash Catalyst can output a finished Flash SWF or AIR application that’s ready to publish on the web.”

Ok. so Flash Catalyst 1.0 won’t kill off Flash CS4, but given a bit of work and some tools we haven’t seen yet, and the desire by Adobe, I could easily see FC 1.5 or 2.0 replace the Flash IDE and for the record, I see that as a good thing.  Here are the PRO’s for killing off the Flash IDE in favor of Flash Catalyst 2.0

1. The Flash IDE has a horrible ActionScript editor. We complain about it all the time, and I’m not convinced that Adobe wants to fix it. Seems like a lot of Flashers jump into Flex for the code editor first, and then play with the Flex SDK later on.
2. No longer held back by the Flash .FLA file type.  The .FLA blows. Every new version of Flash creates a new FLA that isn’t backwards compatible. Its not efficient when working with a team who uses SVN (can’t be diffed).
3. Finally on a level playing field with the “Flex guys”.  Under the hood, we’d be using the same tool, same code base, same components and would have a shared project file. No longer feeling slighted because one or the other got tools that the other didn’t.

So for this to happen, a lot of work would need to happen to Flash Catalyst to make it compete with the Flash IDE. First off, we’d need more drawing tools. The preview release of Flash Catalyst from Adobe MAX only had the bare minimum of tools, and I would image Adobe is hard at work on a lot more drawing tools.  We’d also need a Motion Editor for Flash Catalyst, along with the 3d and IK tools that the Flash IDE has.  There are more items that would be required but those are the essentials to bring Flash Catalyst on par with the current Flash IDE. Its something that I think is very achievable by Adobe.  And maybe… its something they are planning as well.

I have to admit I’m a bit puzzled about the Z axis in Flash CS4 and how Flash chooses to rotate an object. I was toying around with some animations just visually and then tried to set up some objects there were pointing in a particular direction and was running into issues.  The Z coordinates were changing while I was rotating the object. It took a few minutes for me to figure out what was going on (I wasn’t my normal sharp self – haa haa). The 3d rotation tool was always picking the center of my object to rotate around, instead of the registration point.  I guess I never noticed that its that same way for 2d transforms as well.  Of course, my limited experiments using just code of course using rotationY, rotationX, and rotationZ only rotate around the registration point.  I spent all these years forcing myself to work with 0,0 as a registation point because it was easier when working when working with dynamicaly sized objects and loading of assets, and it seeems if you want to play with 3d, at least in the IDE, you’re better off only using a center registration point.    It would be nice in a CS4.0.1 update if there was a way to set this as default in the system prefs.  Maybe a nice little toggle in 3d Center Point? (Hint Hint Adobe)

The other part that is just sort of an annoyance is that the 3d position and view information only shows X,Y and Z coordinates but you have to get X,Y,Z rotation from the transform panel. I know the Transform panel is where rotation has always been, but if you are going through the hassle of redoing the interface, it seems fairly logical to maybe cluster these pieces together finally.

I don’t want this to sound like I’m only complaining. I really love Flash CS4. Years ago, I dreamed of an After Effects timeline in Flash. Seems like only yesterday that I was dreaming of what Flash CS4 was to become (Dec 2005: Return of LiveMotion). Just need that motion vector Blur and I’ll be 3 for 3.

Is this a bug in the Flash player? or just how it is on Mac?  I can’t believe I never noticed it before.  I haven’t looked for a workaround yet. I imagine I’ll need some javascript, but odd that I couldnt’ find much online about it either.

Thoughts? Comments? Suggestions?

[Update 11PM 12/22/08 - I didn't find it in the Adobe Flash Player bugbase, so I logged it --> LINK ]

To reproduce: Set up a Flash CS3/AS3 file. Create a FLVPlayback component on one frame (only one frame), set up a source (via AS3 or via component parameters) and create a few blank frames. Add navigation buttons that exist on all frames and add event listeners for the buttons so you can jump between the frames. Run the movie and navigate to the frame with the FLVPlayback on it, and then navigate away. The audio continues but the component instance is gone.

What I found very interesting is that I could trace the instane name of the removed component afterwards, but if I tried to gain a reference to it via getChildByName, it came back null.   I also found that after adding stop (within a try catch), that it would work when going backwards on the timeline, but not forwards. And then if I jumped ahead and came back, that I could create overlapping audio. Very very weird.

Download Sample file. Run the sample file and navigate around — you’ll see some very odd behavior.

Now, the easy answer is “just don’t build it that way” or numerous other workarounds & fixes, but reality is that a number of beginning users are used to using frames/frame labels and it is would make sense to only have the component on the frame you needed it. My question is –> Whats really going on behind the scenes? (I have a few ideas).

Kudos to my Flash Level 2 students. They are great at finding bugs in documentation and components.

http://theflashblog.com/?p=331