Mixing framerate on same tape? ok?

Russ Motyko · February 2nd, 2008, 12:24 PM

Sony Vegas 7 seems to have no problem switching between frame rates.

Tom Roper · February 2nd, 2008, 12:57 PM

Some people get upset when you tell them 24F has 3:2 pulldown flags added putting 24 progressive frames inside a 29.97 stream. I don't understand why the anxiety. It must harken back to the DV days or something. We're only talking about added repeat flags in the header, not wasted bandwidth from unnecessary encoded frames.

But the repeat flags ARE there, to state otherwise is counter to the observations of Nate Weaver and David Newman.

http://dvinfo.net/conf/showpost.php?...5&postcount=10
http://dvinfo.net/conf/showpost.php?...9&postcount=12
http://dvinfo.net/conf/showpost.php?...4&postcount=15

Having the repeat flags is good and bad. It's good because 24F clips can be natively dropped onto the same timeline as 60i clips if your editor supports it. The played back 60i video just seamlessly switches between 29.97 and 24 fps with nary a hiccup, which makes possible creative opportunities.

But the 60i containerized 24F flags confuse Blu-ray players like the PS3 which can be forced into true 24p progressive output modes for 1080p tvs that support 1080p24 modes, i.e. 72hz, 120hz. On those, 24F containerized inside 29.97 streams takes on a horrid, irregular judder IF you force 1080p24 playback mode from the Blu-ray player. Leave the Blu-ray output at 1080i60 and all is well.

What remains to be seen, is if a hybrid Blu-ray authored disk (24p mpeg2 inside an AVCHD wrapper on DVD5/9 red laser media) can be made to play on Blu-ray players. For that, the 24F pulldown flags would need to be stripped, to get the Blu-ray player to NOT see the stream as 29.97, and thus NOT have to force 1080p24 output, NOT impart its own reverse pulldown mechanisms on the 24F repeat flags (which causes the irregular judder), but just read 24p unforced from the disk.

It's not a huge priority for me as I'm not a big fan of Canon 24F. My resolution testing using the Imatest MTF50 imaging software and the Canon XH-A1 shows a drop in the vertical resolution to about 595 lines. Since the XH-A1 is an interlace scanned sensor, I think 24F is dropping half the fields and interpolating up, i.e. bobbing. The result is generally good, although the image staircases on diagonal lines.

The 60i interlaced mode is essentially artifact free because the 1080 field lines are blended. That too limits vertical resolution to around 700 lines, but I think it could be kicked up somewhat closer to the horizontal resolution by tweaking the [Detail Horz-Vert Balance] parameter, at the expense of introducing some possible twitter/moire artifacts into the XH-A1's otherwise organic 60i image.

Richard Hunter · February 2nd, 2008, 07:56 PM

Quote:

Originally Posted by Pete Bauer

As has been discussed ad nauseum, the only thing not entirely progressive about F Mode is that the part number of the CCD is that of an interlaced chip, but Canon say they've modded it so all 1080 lines are read at the same instant.

Hi Pete. Do you have any reference for this? The only Canon white paper I could find was very brief, and said this:

"These Frame modes have the same look as progressive frame rates, but are not labeled
“progressive” because they are created with an interlaced chip. The end result is exactly the
same to the editing system (and to our eyes) as 30p and 24p, respectively."

It does not mention anything about reading all lines at the same instant, which I thought was not possible with an interlaced CCD. I would be very interested to know how they do it if this is correct.

Richard

February 2nd, 2008, 12:24 PM	#16
Russ Motyko Regular Crew Join Date: Oct 2007 Location: Portland, OR Posts: 56	Sony Vegas 7 seems to have no problem switching between frame rates.

February 2nd, 2008, 12:57 PM	#17
Tom Roper Trustee Join Date: Jun 2004 Location: Denver, Colorado Posts: 1,891	Some people get upset when you tell them 24F has 3:2 pulldown flags added putting 24 progressive frames inside a 29.97 stream. I don't understand why the anxiety. It must harken back to the DV days or something. We're only talking about added repeat flags in the header, not wasted bandwidth from unnecessary encoded frames. But the repeat flags ARE there, to state otherwise is counter to the observations of Nate Weaver and David Newman. http://dvinfo.net/conf/showpost.php?...5&postcount=10 http://dvinfo.net/conf/showpost.php?...9&postcount=12 http://dvinfo.net/conf/showpost.php?...4&postcount=15 Having the repeat flags is good and bad. It's good because 24F clips can be natively dropped onto the same timeline as 60i clips if your editor supports it. The played back 60i video just seamlessly switches between 29.97 and 24 fps with nary a hiccup, which makes possible creative opportunities. But the 60i containerized 24F flags confuse Blu-ray players like the PS3 which can be forced into true 24p progressive output modes for 1080p tvs that support 1080p24 modes, i.e. 72hz, 120hz. On those, 24F containerized inside 29.97 streams takes on a horrid, irregular judder IF you force 1080p24 playback mode from the Blu-ray player. Leave the Blu-ray output at 1080i60 and all is well. What remains to be seen, is if a hybrid Blu-ray authored disk (24p mpeg2 inside an AVCHD wrapper on DVD5/9 red laser media) can be made to play on Blu-ray players. For that, the 24F pulldown flags would need to be stripped, to get the Blu-ray player to NOT see the stream as 29.97, and thus NOT have to force 1080p24 output, NOT impart its own reverse pulldown mechanisms on the 24F repeat flags (which causes the irregular judder), but just read 24p unforced from the disk. It's not a huge priority for me as I'm not a big fan of Canon 24F. My resolution testing using the Imatest MTF50 imaging software and the Canon XH-A1 shows a drop in the vertical resolution to about 595 lines. Since the XH-A1 is an interlace scanned sensor, I think 24F is dropping half the fields and interpolating up, i.e. bobbing. The result is generally good, although the image staircases on diagonal lines. The 60i interlaced mode is essentially artifact free because the 1080 field lines are blended. That too limits vertical resolution to around 700 lines, but I think it could be kicked up somewhat closer to the horizontal resolution by tweaking the [Detail Horz-Vert Balance] parameter, at the expense of introducing some possible twitter/moire artifacts into the XH-A1's otherwise organic 60i image.