Forget about WebM

For the past 6 or 7 months, I’ve attempted to provide video to all types of platforms by encoding videos in both H.264 and WebM. WebM is the encoder of choice for Firefox, Chrome etc in other words “not-Apple” (though Microsoft has thrown in with Apple on H.264). I haven’t had much success. Most of the time, I was not able to play WebM encoded video in Firefox nor Chrome and it was an additional step in my workstream.

Over the weekend, I did two things; I re-encoded all of the videos using the x264 encoder and I deleted all of the WebM video. Hopefully, this hasn’t caused too many issues with my viewing public. Please contact me via the Contact page if so.

A coupe of quick notes:
1) I used x264 in Compressor by adding the x264 component to the Quicktime components in Library. Larry Jordan has a nice explanation of how to do this.
2) For most of the HD videos, I created half-size videos then upscaled them on playback. This results in a moderately small file size with nice quality.
3) The trick on using Compressor is to follow Larry’s advice of creating .mov files which is a Quicktime container. Most browsers will play a Quicktime container but my Nexus 7 had a problem. The key is to use Quicktime 7 to convert it from .mov to .mp4. Larry states one can do this without re-encoding but fails to advise how.

There is a “pass through” option when using “File -> Export -> Convert Movie to MPEG-4”. This will simply copy the file from .mov format to .mp4 format without re-encoding.  Click on Options on the Save window and select MP4 from the drop down menu at the top of the box and click on the Video button right below the dropdown. Below the Video button is another dropdown and on this one, select Pass through. Click OK and Bob’s your Uncle, you will have a file that has broader compatibility than the Quicktime container, .mov.

Hype – a new approach

Thanks to Jon Gruber, I learned about Hype.

Hype is a OS X application that makes it easy to create dynamic websites using HTML5 or more accurately, HTML5, Javascript and CSS3. Hype works. It is also version 1.0 so it is basic but it works, doesn’t crash and holds great promise.

I’ve been attempting to change my WELLINGTON ROAD Pictures site from a database-driven, somewhat static site to a bit more dynamic, HTML5-based site for sometime. See here, here, here, here and here. Sigh, note the last time I attempted to make changes was almost 9 months ago.

I read about Hype, paid my $29.99 and dove in. The first thing I did was try to create an identical version of my current site. Yeah, that was easy but it was also clear that my database approach to my current layout was far superior. With my current layout, when I create a new film, I upload a poster jpeg, my film and add another line to the database. No coding required. Also somewhat boring from a design perspective.

With Hype, I needed to duplicate a new scene (think web page) for each film. And I would have to embed the Notes and Time text on each page. The management overhead is considerably more using the Hype approach. The extra steps make sense for I am duplicating a static site with a content creator for dynamic sites.

What would be required if I went to a more dynamic design? My mental image of a dynamic WRP site would be to use a film metaphor. I wanted to represent each film by a frame in a film strip. I would have a limited number of films showing and the user would move the film forward or backward to see additional films. Click on a frame would show the film using HTML5 video playback with its most important attribute, play full-screen.

So I started anew with Hype. Using a film strip image, I created 4 scenes, each with a filmstrip of four films and I added very simple text buttons to indicate forward, backward and home. It was quite simple to add transitions to make the filmstrip look like it was going from left to right and right to left for forward and backward, respectively. The crossfade transition also looked nice for showing the film. Hype did advise me that the transitions were only available in Webkit browsers (Safari, Chrome) as Opera and Firefox had yet to implement them. Hype also advised me that only Safari and IE would show the H.264 forms of my films and I would have to use WebM or Theora encoding to have them play in Opera, Firefox or Chrome. [Ed. note: The 1990’s are calling and they want their browser incompatibilities back.] It worked and it took about 2 hours worth of work. Not bad!

My review of Hype:

Ease of Use: Excellent! I did read over the manual quite quickly to gain an overview of how Hype worked. I also watched the first tutorial. So maybe 10 minutes of preparation. I quickly created a basic website than went to creating a dynamic website in 30 minutes. The design guides work well and the inspector mode continues to be an excellent method of managing metadata. With my site design, I had a significant amount of duplication of scenes. Copying elements from one scene across to the other scenes worked well and accelerated my development. Speed of  development and ease of use is where Hype shines.

Power: How much can one do with Hype? Quite a bit and not enough. I am an intermediate when it comes to web design, a neophyte in HTML5 and a non-starter in Javascript. So having the ability to quickly mockup a dynamic site with user actions moving from scene to scene was quite liberating. Hype  can and will be a good replacement for Flash beginners. Instead of learning Flash and creating proprietary websites that break the internet, one can create HTML5/Javascript/CSS3 sites that break the internet. Yes, this is the problem with Hype.

Hype creates a single webpage that then using dynamic properties to display new content. It works pretty well but it does break the concept of using a specific address for specific content. I thought that my previous design had this fatal flaw as well but one can right click on any movie to obtain the link to showing the specific film. What is also nice is that the specific page does show the film using the HTML5 Video tag, which means it will allow playback in fullscreen. Hype can not provide this level of specificity. Its as bad as Flash. Bummer. While this is a show-stopper for my site, I’ll continue to comment on Hype.

An issue I had with Hype was global changes. I didn’t see a method of making a global change to all selected scenes. I wasn’t able to select more than one scene then applying a change to all. There is a complete lack of contextual changes. Quite often, I would right click to see if there was something I could do to an element. Nope. No response.

Website: Not so much. Due to Hype’s use of creating one “page” which changes dynamically, ALL content resides in Hype’s folder. For my site, this created a 675MB upload and a duplication of all of my movie and image content. What I could have done was hand edited the Javascript file and changed all of the references to those inside the folder to the pre-existing content on my site. This isn’t a deal-killer. I think a simple property that allows one to enter the address of the content would make this problem go away fast. Version 1.1 perhaps?

Summary: I like Hype. It is easy to use, provides extremely fast prototyping and can be quite powerful using web standards. [Ed. note: I’m not even going to check if it provides validates. I mean technology that is standardized for the web as compared to Flash which is proprietary.] It does break the web so I’m not going to use it for WELLINGTON ROAD Pictures but I will use it to test out concepts. I think it would be fantastic for a class in learning about dynamic websites. I look forward to version 3.


Step 4: Updated to be dynamic

No significant changes here. Added the required code for the site to be served by a database instead of hand-coding. A couple of small formatting changes but I’m ready to go from my old but lovely way of serving videos, Shadowbox to spanking, new HTML5.

Step 4

One update and one non-update:

Update; looks like H.264 is getting to be the standard for video.  Based on the MPEG LA announcement, H.264 will be forever free if the video being served is free. Due to this announcement, I’m going to focus on serving H.264 only and forget about…

The non-update: Jillion is running out of August. Its August 28th and no sign of their sublime video code. When it shows up, I’ll take a look until then I’m off to do my own thing.

Update: I went ahead and implemented Shadowbox again. I like the way it works, I don’t like that it doesn’t use the new HTML video selector. The new one in Safari has a much nicer controller and the ability to go full size. Jilion, Ferris Jilion?

FCP 7 and ProRes

FCP 7 introduces a 3 new HD codecs for production work, ProRes 4444, ProRes 422 (LT) and ProRes 422 (Proxy). With these additions, this brings to five versions of HD codecs  available for production work. I did a simple comparison of Flip H.264 transcoding to all 5 versions, knowing full well that I wasn’t starting with the best source material.

The only version where I could see a difference was with ProRes 422 (Proxy). There were artifacts and a loss of detail. LT looked great, with no perceptible differences on up to 4444. Which means given the source, the best I could use is LT. That said, I did have striking differences in gamma between the new codecs looking identical to the source material and the “older” codecs, 422 and 422 (HQ) being much darker and losing details in the shadows.

I’ve tried to understand the gamma shift issue regarding H.264 and QT… this looks to add more confusion to the mix.

For me, going forward,  I’ll transcode the Flip to ProRes 422 (LT) then use it in FCP. I still need to test using ProRes 422 (LT) in iMovie.

Thanks, Norman!

I was thinking this, and was going to write about it but I didn’t have the depth of experience to make the point as clear as Norman does.

In his blog, Hollyn-wood (Norman, that is), Norman makes the point that the price cuts and the focus on ease of use sets this upgrade to Final Cut Pro/Studio as a significant push by Apple towards the low-end side of video editing. He doesn’t mean that Apple took features away or that Final Cut Pro isn’t capable of editing high-end feature films, what Apple did do was to introduce ways of helping the not-so-good be a lot better or sometimes, good-enough. AND, they dropped the price.

The point? A big push by Apple to become the perferred platform for all of the independent, YouTube, Vimeo, film-makers…

Thanks, Norman. My thoughts exactly!