Archive for January 20th, 2008

The HD war is over

It’s amazing how fast a product can die. After the news broke on the news outlets a few days ago that HD-DVD is dead because of Warner’s decision to move exclusively to Blu-ray (although Blu-Ray was winning slowly but surely), within 10 days the 62-38 percentage of sales went at 85-15. Overwhelmingly in favor of Blu-Ray.

Now, there are those who say that Apple’s new HD renting system will also kill Blu-Ray and Netflix, but it’s not the time yet, and it won’t be for another 10 years. AppleTV’s 4mbps 720p files can’t compete with Blu-ray’s 18-40 mbps 1080p quality (plus the extras on the discs, and the feeling of physical ownership is a driving force to buy/rent real discs). The iTunes/AppleTV deal has its niche, and it’s mostly for those who can’t be bothered a lot with movies and don’t care about quality.

We discussed with JBQ last night over dinner at a local restaurant that a 480p DVD disc encoded in VBR 9mbps mpeg2, has BETTER quality than AppleTV’s 720p HD at 4mbps h.264. H.264 is twice better than mpeg2 in compression, true, but 720p is also more than twice bigger than 480p. From our point of view, it makes more sense from the quality point of view to serve a true 480p anamorphic file at 3 mbps h.264, than the current 720p deal. But unfortunately, 480p is not considered HD (it’s EDTV), and this wouldn’t fly with the consumers. So basically, it’s a marketing problem, not a technical one. Apple decided (no surprise there) to go with a decision that flies high with the clueless consumers rather than delivering the best possible quality. Another problem is bandwidth: at 5GB per movie rent, Comcast won’t be too happy… Bandwidth ain’t free.

Video editor mockup

I am continuing my (agressive) discussion (it started because of my Linux video editor rant) with a particular Linux video editor developer who doesn’t want to face the truth that his application sucks in both usability, features and looks. Of course, I told him so point blank — I never hid my opinions and I won’t start doing so now. When he asked how it should look like, I sat down for 4 hours and made a mockup for him. Here’s the result:

This mockup borrows ideas from both iMovie and Sony Vegas, and it includes features that are easily accessible: e.g. tape capturing/recording, direct editing of DV/HDV/AVCHD and any other format that ffmpeg/mencoder can support, a two-sized timeline (normal and mini), support for any frame rate, track rotation/placing, clip rotation/pan/cropping, plugins, transitions, fade in-out right from the timeline, slow/fast-motion up to 16x (both by CNTRL+resizing a clip’s edges and by editing properties), group/ungroup elements, transparency/overlay support, enough plugins for color correction and more, snapping on/off, automatic fade transitions when two clips overlap, a 3 quality preview window for speed reasons, preview sizes that are only at 1:1, 1:2 or 1:4 so quality is preserved for easier editing, reversing, full screen and secondary monitor support, on-the-fly de-interlaced preview, exporting dialogs like this one, a fork of the Lagarith lossless codec to make it faster/smaller, and other things that I won’t mention here. My design is even SVGA friendly as long as your footage is non-widescreen DV (it will preview at 1:1 size).

Now, you are probably thinking that this is way too much, and it just copies Vegas’ abilities, an application that’s been in development for years and has reached some maturity. Well, not really. Believe it or not, I have left out a lot of features that exist in the consumer version of Vegas: compositing and compositing children, masking, audio recording for narration while video plays back, custom timeline size, velocity engine, keyframing for the plugins and the timeline, a proper audio mixer, markers, regions, pre-rendering of regions, “takes”, trimmer window, image sequence import/export, DVD authoring companion application, different time formats and timecodes, and more. Also, consider that the Pro version of Vegas has a lot more features, like 32bit editing, inverse/telecine, Pro text editing, DVCPRO-HD and lots of plugins.

So, no, I am not being an asshole trying to force “out of this world” difficult features for an OSS video editor. The features I left in are must-have in this time and age, end of story. If you can’t deliver them, then you might as well stop developing your application (as long as you develop it in order to be useful to users instead of just your programming hobby, of course). Another thing I would suggest is to not cut off yourself from Windows and Mac. Download the trial versions of video editors that exist for these platforms and by all means, copy their best ideas from each.

But as I wrote the other day, this is not a one-man job, neither a buddy-created application. It requires at least 10 engineers who know what the hell they are doing, and they are sitting next to each other. Which is why this can only be delivered in any user-friendly fashion by Red Hat/Novell/Ubuntu (or Google, if they want to be more helpful to Linux users). It will take 2 years to get there, $100,000x10x2=2 million USD. That’s the minimum you need to develop such an app (without counting any license fees, we would assume we just use these codecs the way we use them now on Linux: for free). In a more practical, real-life scenario, you need more engineers and at least $5 mil.

The new state of things

A few months ago I was a bit unsatisfied with my life. I felt angry and sad most of the time. I thought I was just in mid-life crisis. But things are good now. For the last 2 months things are improving from good to best.

And then it hit me why: JBQ used to be unhappy and overworked while at Openwave, while this is not so anymore over at Google. He just has more time for me now, he is less aggravated, and more loving. I feel that JBQ is happier than ever at Google, less stressed, more optimistic. And this has had a very positive impact in our relationship.

Update: JBQ blogged about it too.