Question for Graeme & others concerning DCT compression

[Graeme Nattress]

DCT processing is hardly intensive in the way that MPEG2 is, so I don’t think “extra time” matters as the codec is already faster than real time without much difficulty.

Graeme

– http://www.nattress.com – Film Effects and Standards Conversion for FCP

[Graeme Nattress]

The JVC HD100 does indeed create a 60p MPEG2 file for 24p, but the extra frames are not merely duplicates, but flagged duplicates. That means that in terms of compression efficiency it’s recording just the 24p frames (like it’s something like a 1 byte flag to say that a frame gets duped) but you see a 60p MPEG2 coming out the other end, but it only takes up as much space as a 24p MPEG2 stream. Quite clever really….

Graeme

– http://www.nattress.com – Film Effects and Standards Conversion for FCP

[Deleted User]

Thanks for your time & informative responses!

– Peter

[Barry Green]

[Peter DeCrescenzo] “Wouldn’t it be easier & faster for the cam’s electronics to simply _copy_ the redundant frames (for example, the 6 frame difference between 24 fps and 30 fps) instead of _compressing_ the redundant frames? Those 6 frames are identical, right?

No, those frames are never identical to any other frames. The pulldown frames are created from taking a field from the prior frame and a field from the next frame. So, in 2:3:3:2 pulldown (the most compression-efficient mode), you have four full frames, followed by a fifth that is made from half of one and half of the other. Sort of like AA/BB/BC/CC/DD. Since the recording format is interlaced, the frames get split into fields for recording (same as in HDCAM 24PsF). The “fifth frame” is made from one “field” of one frame and one “field” of the other. Not a duplicate of anything else.

If a cam’s in 24 fps mode, I’d expect it to allocate relatively little of its processing muscle to record 6 redundant frames onto tape, and instead spend more of its time and horsepower to better-compress each of frames 1-24 (in 24 fps mode) than it normally would to compress each of frames 1-30 (in 30 fps mode).
In a GOP-style system, that is what happens. In DV, or Digital Betacam, or M-JPG, or MPEG-IMX, that’s not how it works. Every frame is compressed separately, regardless of anything happening in any other frame. So every frame gets the exact same compression applied to it, whether it’s a duplicate or not.

That’s why the data rate is so much lower in DVCPRO-HD’s 720/24p mode when recording to P2; duplicate frames are simply dropped, they’re not compressed at all. So instead of taking up 100 megabits of space, it only occupies 40. Each frame is compressed the same, but there are fewer frames, so it takes up less space.

But are you saying that’s not how it works? Or does it depend on the cam/model?”
It depends on the recording format. DV or any of its variants, and any other frame-based codec (such as, again, Digital Betacam or MJPG etc) all encode frame-by-frame. So does MPEG-IMX, which is an MPEG codec but is intra-frame only. Long-form GOP compression, such as MPEG-2 HDV, can take advantage of frame duplication to get more efficient compression for the frames it does compress.

—————–
Get the most from your DVX camera. The DVX Book and DVX DVD are now available at https://www.dvxuser.com/articles/dvxbook/ and at Amazon (https://tinyurl.com/54u4a)

[Toke]

[Barry Green] “Sort of like AA/BB/BC/CC/DD. Since the recording format is interlaced, the frames get split into fields for recording (same as in HDCAM 24PsF). The “fifth frame” is made from one “field” of one frame and one “field” of the other. Not a duplicate of anything else.”
Well, because tape systems care only about fields (they are not recording frames, just fields one after another), there are duplicate fields.
I really don’t have a clue if those duplicate fields are compressed again…

[Barry Green]

All i-frame compression schemes always compress whatever’s in its frame. There is never a savings on duplicate fields/frames when you’re working with a frame-based compression system.

—————–
Get the most from your DVX camera. The DVX Book and DVX DVD are now available at https://www.dvxuser.com/articles/dvxbook/ and at Amazon (https://tinyurl.com/54u4a)

[Toke]

[Barry Green] “There is never a savings on duplicate fields/frames when you’re working with a frame-based compression system.”

Maybe the term “field-based compression” would be more accurate.
And original question was not about better PQ with duplicating fields, but saving CPU power…