ISDCF held another Plugfest this month. The primary targets for study were audio routing capability, and on-screen rendering of subtitles and captions. The results were underwhelming.
Plugfest testing has long revealed a problem in on-screen rendering of subtitles. The problem is historical, dating back to the introduction of the TI Series 2 class of projector in 2010. Prior to that, the Series 1 design produced acceptable results, for a few reasons that no longer exist. Series 1 designs were driven solely by Texas Instruments, which required its licensees to adopt their design in order to assure the industry of interoperability among competing designs, and in particular, color compatibility among competitors. When the Series 2 class of design was introduced, the licensing model matured to that of a reference design, where the licensee is given a design upon which it can innovate. It was economic for TI to do so, as it could not afford to maintain the production design in-house, and its licensees needed more flexibility to compete.
TI made a substantial mistake, however, when introducing the Series 2. For misguided reasons, it chose to leave out on-screen subtitle rendering from its reference design. Following strong criticism, it later fixed that problem, but not very well. To complicate things, TI needed to upgrade its on-screen rendering engine to support SMPTE subtitles, which have a different format to that used with Interop-packaged content. Things have never been the same since.
One of the reasons TI chose to leave out subtitling from the original Series 2 design was the belief that all servers were capable of internally performing the rendering of on-screen subtitles, without need of a separate rendering engine in the projector. Rendering subtitles in the server or IMB led to a host of other interoperability issues that ISDCF has long been addressing.
But the subtitle issue that ISDCF is dealing with today is that of synchronization. Rendered subtitles appear on-screen seconds late in some products, which is simply unacceptable. To complicate matters, there is speculation that the problem is worse when the SMPTE subtitle format is used. The overall problem of synchronization has been noted before, but not highlighted as strongly within ISDCF as it was this month. No one has stepped up to say they have solved the problem. In particular, this is true for TI, which tends to be shy to assign resources to digital cinema issues. Hopefully, more will be learned about solutions to this problem in the coming months.
3D subtitles were also studied in this Plugfest. SMPTE recently passed its specification for 3D subtitles, which makes possible the release of 3D subtitles in text form, rather than having to produce unique image files with “burned in” 3D subtitles. The general report was that 3D subtitling appears to work in the products tested, but with problems that manufacturers can easily fix. Even when such problems are fixed, however, distributors will be left with a substantial footprint of Series 1 projection systems, which have no ability to render 3D subtitles. Leaving distributors with little choice but to continue to release 3D subtitles in the most expensive manner possible – burned into unique picture files.
Audio routing got a workout at this Plugfest. SMPTE DCP requires that the server and/or IMB route the packaged audio channels to the correct output. Unfortunately, the industry has yet to produce any documents that say “route it this way.” In the absence of direct instruction by SMPTE or DCI, some manufacturers have simply ignore the need for audio routing. It can be hard for engineering departments to justify the expense, when it’s not needed to pass DCI compliance testing. The inability of some products to route audio correctly, however, caused the need to extend and upgrade the Interop DCP format, as the HFR 3D version of Hobbit failed to play audio correctly on some systems in SMPTE DCP format. Following the results of this month’s Plugfest, the conclusion of ISDCF this month was that the industry is not ready for SMTPE DCP. (Which is a leader to our next article in this report.)