DV Info Net

DV Info Net (https://www.dvinfo.net/forum/)
-   Canon XH Series HDV Camcorders (https://www.dvinfo.net/forum/canon-xh-series-hdv-camcorders/)
-   -   24F vs 30F in A1 vs HV20 (https://www.dvinfo.net/forum/canon-xh-series-hdv-camcorders/90359-24f-vs-30f-a1-vs-hv20.html)

Tom Roper March 31st, 2007 06:36 PM

Quote:

Originally Posted by Pete Bauer (Post 652051)
Dino's got it right.

Quote:

Originally Posted by Dino Leone
On the A1, on the other hand, having 24F selected will indeed output 24F (23.976 in fact) over firewire, correct?

Doesn't have it right. It's 29.97 over firewire. Pulldown is added.

Bill Pryor March 31st, 2007 06:56 PM

Sorry Tom, that is incorrect. When recording HDV, the XH A1 records and outputs true 24fps progressive. When recording in SD at 24F you have the option to add pulldown, just as the DVX100 does. HDV, however, at 24F, is always true 24p. No pulldown. If you take that into a 60i timeline in FCP (and presumably in other NLEs as well, the pulldown will be added by the NLE (and it will have to, therefore, render). You keep it in a 24p timeline, it stays 24P, no rendering.

Tom Roper March 31st, 2007 08:17 PM

Quote:

Originally Posted by Bill Pryor
...If you take that into a 60i timeline in FCP (and presumably in other NLEs as well, the pulldown will be added by the NLE (and it will have to, therefore, render). You keep it in a 24p timeline, it stays 24P, no rendering.

http://vsdrives.com/graphics/XH-A1/24F.png

Bill, respectfully but I think you have it backwards. My earlier post was directed at you when I said you need to take the NLE out of the equation and just consider the capture stream. Please click on the link above. The capture is a 24F stream from an XH-A1, but the frame rate is 29.97. The capture already has the pulldown in it even before it goes to the NLE.

Pete Bauer March 31st, 2007 09:49 PM

I just checked the XH User Manual (see page 40 of the English language manual). The way I read it, SD 24F gets sent out to 1394 as either 2:3 or 2:3:3:2 (shooter's choice), otherwise F-Mode over 1394 is progressive. Oddly enough, it looks like SD 30F get recorded to tape as 60i PsF, but then sent out as progressive frames. Both 24F and 30F (and I assume 25F in PAL-land) are progressive in HDV.

Tom, I use PPro so am not familiar with the dialog box you linked. Not sure why the software reports HDV F-Mode as 29.97 unless it was 30F? Canon 24F should be 23.976. I do know that PPro reports HDV F-Mode footage correctly based on progressive frame rates.

Tom Roper March 31st, 2007 11:39 PM

Pete, I used Vegas to capture from tape the same 24F clip that I captured earlier with the CapDVHS utility. Vegas reports the clip as HDV 23.976 fps progressive. But the same clip captured by Vegas is reported with clip properties 29.97 fps mpeg2 bottom field first in several other apps.

I'm certain it's 29.97 fps mpeg Psf. I'm trying to understand the semantics of Vegas calling it 23.976 fps progressive. And I think the answer is just semantics. HDV versus Mpeg. In other words, HDV 23.976 fps progressive is another name for mpeg2 29.97 fps Psf with pulldown added.

Below, click on the link for a screen capture from Ulead MF clip properties. This is the clip captured from tape by Vegas and reported by Vegas as "HDV 23.976 fps progressive."

http://vsdrives.com/graphics/XH-A1/24F_3.png

It's important to note that Ulead MF authored that clip to HD-DVD format without re-encoding it, not possible if it were true 24p. Mpeg 24p is not a supported format by HD-DVD, just VC1 24p. But it plays.

Salah Baker noted in this thread that HV20 records 24p (not PsF), and sent a clip (which I am downloading now) to David Newman of Cineform. David said that clip had pulldown added inside of 60i, i.e. it *is* PsF.

Brian Brown April 1st, 2007 12:04 AM

Hmm... well I also have screen grabs from 24F footage from the XH-A1. However, on mine both ProCoder Express and MainConcept's encoder show it as progressive, 23.976fps footage.

http://www.brownland.org/video/24F.jpg

I did capture this clip in Premiere Pro... so does this NLE ignore/ remove pulldown when putting streams on my hard drive?

I'll have to try to capture with HDVslpit or something and see what happens.

This is strange indeed!
Brian Brown

Geoff Dills April 1st, 2007 12:21 AM

1 Attachment(s)
A1 24f clip captured using HDVSplit, opened in MPEG Streamclip. Attached is the readout showing indeed it is 23.976 fps. Everything I've read on the forum indicates the A1 records 24 frames to tape with NO pulldown.

Barlow Elton April 1st, 2007 12:41 AM

24F is a 24 fps stream with no pulldown HOWEVER...I believe there are repeat flags in the stream intended to make it easier for the decoder to add pulldown if necessary. Before I had native HDV 24F support in FCP, I captured raw m2t's with MPEG Streamclip and it indicated that the files were 29.97 and VLC player would play 24F clips with repeat frames that had a noticeably odd cadence if you knew what to look for.

24F is 23.976 fps.

Tom Roper April 1st, 2007 02:44 AM

Quote:

Originally Posted by Barlow Elton (Post 652159)
24F is a 24 fps stream with no pulldown HOWEVER...I believe there are repeat flags in the stream intended to make it easier for the decoder to add pulldown if necessary. Before I had native HDV 24F support in FCP, I captured raw m2t's with MPEG Streamclip and it indicated that the files were 29.97 and VLC player would play 24F clips with repeat frames that had a noticeably odd cadence if you knew what to look for.

24F is 23.976 fps.


Some apps are reporting 23.976 and some are reporting 29.97. I think you are right about the repeat flags inside the stream, but a repeated frame is never actually encoded because it wastes compression efficiency. The flags themselves are the repeat mechanism for pulldown, so I think you make a case for it being there.

Tom Roper April 1st, 2007 02:48 AM

Quote:

Originally Posted by Geoff Dills (Post 652155)
A1 24f clip captured using HDVSplit, opened in MPEG Streamclip. Attached is the readout showing indeed it is 23.976 fps. Everything I've read on the forum indicates the A1 records 24 frames to tape with NO pulldown.

Your readout also says "upper field first." <---

Mikko Lopponen April 1st, 2007 03:46 AM

Quote:

Originally Posted by Tom Roper (Post 652175)
Your readout also says "upper field first." <---

Why don't you guys just take a player and go through a file frame by frame? It should be blatantly obvious if its 29.97 or 23.97.

Piotr Wozniacki April 1st, 2007 04:17 AM

Quote:

Originally Posted by Justin Avery (Post 652032)
Which part is wrong Piotr? Can you set this thread straight?

Sorry for having been so sparse in words, I didn't want to be rude - just was going to bed:)

What was wrong was using the term PsF in the Canon A1 context; it's a term reserved for true progressive, written within interlaced stream in the form of two half-frames (fields or segments) of half-resolution, but no temporal difference, per each frame (Progressive Segmented Frames).

Canon A1 uses "F" which also is progressive, but named differently because it comes from interlaced chips, and is achieved with some very clever signal processing.

Also, I had the impression "PsF' was confused with "FpS", which would of course stand for frame rate.

Geoff Dills April 1st, 2007 10:51 AM

Quote:

Originally Posted by Tom Roper (Post 652175)
Your readout also says "upper field first." <---

and is that wrong?

Brian Brown April 1st, 2007 02:14 PM

Quote:

Originally Posted by Geoff Dills (Post 652282)
and is that wrong?

Geoff, "upper field first" would indicate interlaced footage. "True" progressive would presumably make no mention of fields.

Boy, this is an interesting thread. It seems like we're all getting different indications from various programs about 24f streams.

Brian Brown

Geoff Dills April 1st, 2007 03:10 PM

(voice of Homer Simpson) DOH!!!


that is interesting that it shows up 23.976 interlaced as that's not a supported HDV format from what I can tell. Maybe this is a reason Canon can't call it "P" .


All times are GMT -6. The time now is 06:34 AM.

DV Info Net -- Real Names, Real People, Real Info!
1998-2025 The Digital Video Information Network