Author Archives: Multimedia Mike

GSoC 2008 Students

Google has announced which students have earned slots for the 2008 Summer of Code. As with previous years, I don’t know whether to congratulate or console the constituents of this collection:

  • Alexander Strange: Generic frame-level multithreading support for FFmpeg
  • Bartlomiej Wolowiec: Nellymoser Encoder
  • Jai Menon: ALAC Encoder
  • Keiji Costantini: LGPL reimplementation of GPL sws_scale parts
  • Kostya: AAC-LC Encoder
  • Ramiro Polla: MLP/TrueHD encoder
  • Sascha Sommer: WMA Pro Decoder
  • Sisir Koppaka: VP3/Theora Encoder
  • Zhentan Feng: MXF Muxer

Feast your eyes on those ambitious projects. It’s going to be quite a summer.

A hearty “thanks” and “good luck to you too” go out to the registered FFmpeg mentoring crew, including Andreas Setterlind, Andreas Öman, Aurélien Jacobs, Baptiste Coudurier, Benjamin Larsson, Jean-Baptiste Kempf, Justin Ruggles, Kristian Jerpetjoen, Luca Barbato, Reimar Döffinger, and Robert Swain (2006 GSoC Alumnus). And as always, there’s the unofficial über-mentor, Michael Niedermayer, who also has the final say in whether a project’s code is ready for inclusion into the tree.

BFI Boredom

In the nick of time, Sisir Koppaka finished the BFI playback system and qualified for FFmpeg’s 2008 Summer of Code:


BFI Boredom

The format is used in a few multimedia-heavy games from Tsunami such as Flash Traffic: City of Angels, which is pictured above. Remember that FFmpeg does not stipulate that a supported format be at all useful, or that it come from a good game. As you can see from the sample above, not even the actors could maintain any enthusiasm through the production.

Portable Movie Super Player

I still read the IMDb Studio Briefing everyday, though it gets a little discouraging. I sometimes wonder if there will ever be anymore interesting multimedia tech news. I should have more faith: New Movie Media Devices Predicted. Really, the story here is that IBM has developed a new, giant capacity yet very small storage method. This is one of those curious situations where they don’t mention how large capacities can possibly reach but instead express the capability in terms of how much media the thing might theoretically hold. It’s left as an exercise to the reader to decide what the average size of a ‘song’ or ‘movie’ might be and compute from there.

Remember the days when CD-ROM storage capacities were expressed in terms of how many printed documents it could hold? Later, the benchmark was number of pictures, then songs. Now it’s movies. This article cites that a device built around the memory could hold the 3500 movies or 1/2 million songs. Thus, the average movie is ~140 times larger than the average song.

The weirdest aspect of the articles floating around is that the hypothetical device would come with 3500 movies prepackaged and the consumer would purchase codes to activate individual movies.

Given recent media consumption trends, there’s little reason to doubt this strategy.

Sun’s Multimedia Rumblings

I’m reading fluffy press releases today about how Sun is going to work towards developing an open video codec: Sun Tackles Video Codec. The article is short on substance which is generally what earns this article a spot the Multimedia PressWatch category of this blog. Something about an Open Media Stack (OMS), perhaps correlated somehow to Open Media Commons (not to be confused with Open Media Now!).

It’s hard to find anything about this initiative that’s not a rehashed press release. But this Sun blog seems to have the most authoritative information, abstract though it may be. They present a fascinating design approach: Rather than evaluate algorithmic techniques based on their performance, evaluate them based on their legal status.

Good luck to them. Here’s a Wiki page to track it.