What’s in a codec?

By Paul Sweeting -- Video Business, 12/2/2005

DEC. 2 | Sony Pictures caused a bit of a stir in the high-def world the other day when it announced the first full-length film to be authored and encoded on a Blu-ray Disc, Charlie’s Angels: Full Throttle, was compressed using MPEG-2, the same codec used for standard DVDs.

Moreover, in comments to reporters following the announcement, Sony officials indicated that all of the initial batch of titles at launch, from Sony as well as other studios, would also likely be encoded in MPEG-2. Although the format is designed to support two newer, more efficient codecs, Microsoft’s VC-1 and a version of MPEG-4, Sony said it is likely to rely on the tried-and-true MPEG-2 “for the foreseeable future.”

The decision surprised people, because it could be seen to negate one of Blu-ray’s most heavily promoted advantages over the rival HD DVD format: its greater storage capacity. That’s because though no codec is inherently best for producing high-quality pictures, codecs can be optimized for various applications.

For instance, VC-1 and MPEG-4 were originally created for transferring video over the Internet, so they were designed to produce high-quality high-def images at relatively low bit rates. MPEG-2 also produces high-quality images but generally requires higher bit rates.

Charlie’s Angels was encoded at an average bit rate of 18MB per second, compared to the 7MBs to 8MBs that VC-1 and MPEG-4 are designed to run at. Though the images might be the same, higher bit rates mean you have to store more bits to start with, which takes up space on the disc.

At 18MBs, Charlie’s Angels eats up the better part of a 25GB Blu-ray Disc, leaving room for no more than the ordinary compliment of extras—hardly the sort of over-the-top package many Blu-ray supporters gave as the reason to adopt the higher-capacity format.

Sony officials counter that high-quality MPEG-2 encoders have been around for years and the format is well understood, while encoding tools for the newer codecs are still being developed and the formats are less “mature.”

But the fact remains, you could fit both Charlie’s Angels I and II on a 30GB dual-layer HD DVD disc using VC-1 and still have room to spare.

Anyone who has followed the format war from the beginning, however, should not have been surprised by Sony’s decision. In fact, much of the critical differences between the formats are the result of a disagreement over codecs.

From the start, Sony set out to create a hardware format based on MPEG-2 and for a long time resisted calls by other Blu-ray companies to include support for advanced codecs in the format.

MPEG-2 is the compression format used in the U.S. HDTV standard, and Sony wanted Blu-ray recorders to be compatible with U.S. broadcasts. But Sony also has a deep institutional investment in MPEG-2—it has nearly 150 patents in the MPEG-2 licensing pool—and was keen to extend the codec’s reach.

Because of the high bit rates needed to produce high-def images using MPEG-2, however, Sony had to figure out a way to pack more data onto a 12cm disc.

To do that, it discovered it needed to change the structure of the disc by moving the data layer closer to the surface and using a more tightly focused laser. The result was a physical disc structure incompatible with DVD.

Toshiba took precisely the opposite approach in developing HD DVD. It started with the existing disc structure—based largely on its own intellectual property—and turned to advanced codecs to reduce the amount of data it needed to store to get high-def images.

The result was a nice, easy-to-manufacture disc but one that required the studios to cede a measure of control to IT companies like Microsoft, which provided VC-1 (and much of AACS) but is not particularly committed to the future of optical disc formats.

No wonder the studios can’t seem to decide which way to go.