It's good to hear that my intro page is getting used!
The DVD format was developed in the mid-90's (actual development probably goes back a fair bit farther than that), and at the time there was no way to display more than 480i in the home environment. Therefore it made perfect sense to develop along those lines.

The infrastructure for film is getting close to a century old now, and the 24fps speed still works well - moving to 30fps would be a 25% increase in the amount of film used (with costs across the board - particularly film stock, developing, and distribution) and would require the retrofitting or replacement of an insanely high number of cameras. It's just not practical to do so.

A CRT television (standard definition) can use all 480 lines, although not all sources offer that many lines. I believe that standard VHS lands around 250 or so, for example, with S-VHS going up to the full 480.

You actually touched on one of the principal points of the new Secrets 1080p article: a fixed-pixel display device like DLP or LCD is inherently de-interlaced. Plasma's the same way, for that matter, as I understand it. It can't display interlaced source material without de-interlacing it first. All of the pixels (or mirrors, in DLP) get refreshed at once every 1/60th of a second. If an LCD display with 720 lines of resolution gets a 1080i input, it has to both scale and deinterlace the input before it can display it (usually by deinterlacing first and then scaling the 1080p signal).

>The infrastructure for film is getting close to a century old now, and the 24fps speed still works well<
Ok, how about a news room camera or a network camera covering a basketball game. Is that shooting in 24fps?

Thanks again

Bob

If it's a live broadcast, they aren't using film - they'd be recording either to analog videotape or (more and more often these days) to a digital medium of some sort, which either way will be 60 cycles per second.

This actually brings up one of the long-standing challenges encountered by progressive scan DVD players when deinterlacing - some material started life at 24fps and some (such as many, but not all, TV shows on DVD) started life at 60fps. The DVD format provides flags for identifying the video's origin, but the flags aren't always right (hence problems with "flag-reading" progressive scan players).

Gonk,

>This actually brings up one of the long-standing challenges encountered by progressive scan DVD players >

That was going to be my next question. Would this be a problem just for progressive scan, what about the first generation players? A 480I player would not have to deinterlace but it would still have to do the 3:2 pull down??

Bob

A player outputting 480i doesn't run into this problem because 3:2 pulldown is only an issue when deinterlacing film material, so there's no need to worry about 3:2 pulldown with 480i output. When you play back a 480i source at 480i to a display that operates at 480i (as would be the case with a DVD player used with a standard definition TV), none of this is an issue.

This discussion has reminded me of a project I started last summer - a general-purpose guide to home theater that would fill in the blanks inevitably left by user's manuals, similar in content to my Home Theater Basics page but organized differently. The idea came from conversations with UMtiger and his wife when they were trying out the Model 990, with a bit of inspiration from Outlaw's SMS-1 Guide tossed into the mix. I got a pretty decent draft of the document pulled together by last September, but then got buried in work - plus I didn't know quite what I'd do with it once it was done anyway (some folks suggested publication, but I was even thinking of offering it as a PDF download for a small PayPal fee). It's been sitting around untouched for six months now. Anybody interested in seeing what I was thinking of?

>A player outputting 480i doesn't run into this problem because 3:2 pulldown is only>

Why is this, you still have 24fms, don't you have to deal with that since the "clock" on the TV is still working at 60Hz? I must be missing something. Bob

Keep in mind that the migration from 24fps film to 60cps video has already happened when the DVD is produced. When the DVD master is created from that 24fps film source, part of the process involves developing 10 interlaced frames of video from every four frames of film. (There's an example of this near the middle of this Secrets article .) As far as the DVD player is concerned, every disc it gets is 480i at 60 cycles per second (or 30 complete frames, more or less). When the video output is kept at 480i, that initial "trick" of video mastering performed before the DVD was created will remain unaltered and the player has nothing to do. It's not the best solution, of course, but it's the only one that works reasonably with 480i/60 from a 24fps source.

When we give the player a chance to generate a 480p signal, we are offering the player a chance to correct some of the limitations imposed by going from 24fps to 480i/60 - which is both blessing (improved picture quality) and curse (opportunity to screw up that improvement). The player is going attempt to create 60 complete video frames, using alternately three and two copies of each original film frame (10 complete frames for each 4 originals).

>When we give the player a chance to generate a 480p signal, we are offering the player a chance to correct some of the limitations imposed by going from 24fps to 480i/60<
Just so I understand you, even with a progressive scan player, when it looks at the DVD the migration from 24-30fms has already taken place?

>using alternately three and two copies of each original film frame (10 complete frames for each 4 originals).<
When you say orginal frames here, you mean the migrated frames on the DVD which are at 30Fps?

Bob

No matter what video processing the DVD player offers, the source material on the disc is already 480i/60 (30 frames per second, interlaced). That is inherent in the DVD format - there's no way around it.

I mean original film frames (of which there were 24 per second before the film was transferred to video). In other words, the player is working backwards to piece those original frames back together while taking into account the "p24"-to-"i60" conversion process that took place during the DVD mastering process.

This is exactly why deinterlacing is hard, and why "cadence" based deinterlacers often have better luck. Deinterlacing has to be done differently for sources that were originally 24fps film or 30fps interlaced video. After all, if the DVD is taken from a TV show that was shot on video tape, there was never a 3:2 pulldown performed and the deinterlacing simply needs to combine every two interlaced frames - doing a 3:2 pulldown as part of the deinterlacing process would actually mess things up. A flag-based deinterlacer trusts the person who mastered the disc to flag the video data properly ("this is taken from film" or "this is taken from video"), and if the flags are set wrong the deinterlacer is out of luck. A cadence-based deinterlacer looks at the patterns of the individual interlaced frames and tries to recognize when a 24-to-30 conversion (3:2 pulldown) has taken place in the creation of that video transfer, then behaves accordingly.