DV Info Net

DV Info Net (https://www.dvinfo.net/forum/)
-   Canon XL H Series HDV Camcorders (https://www.dvinfo.net/forum/canon-xl-h-series-hdv-camcorders/)
-   -   Kaku Ito's XL H1 video clips -- here's the second batch (https://www.dvinfo.net/forum/canon-xl-h-series-hdv-camcorders/52110-kaku-itos-xl-h1-video-clips-heres-second-batch.html)

Tom Roper October 5th, 2005 06:57 AM

The "carspassing" clip from the XL H1 has edge to edge sharpness in spades. But HD DESTINATIONS "Bora Bora" it's not.

On the Vasst site, they probably did their best to highlight the capabilities of the FX1. Filming a duck floating on water is let's admit, not too demanding but can be pretty.

Shooting passing cars and tourists is not as artistic, but tells a lot more about what to expect from a run and gun.

But I want my videos to look like Discovery HD Network, not my GR-HD1. And I want to go to Bora Bora and shoot them, and never come back...sigh.

Robert Niemann October 6th, 2005 01:14 PM

Does somebody realize block artifacts at the grass in the video with the dog? I wonder, how it will look on a blow up/big screen.? I guess, there you will see it still more obviously.

Michael Pappas October 6th, 2005 03:22 PM

This camera is very sharp. Clean image, great low light and very bright nice colors. The lens performs very very good. The best HDV camera out there by far. It's a winner of a camera.

Pappas

Nick Hiltgen October 6th, 2005 04:35 PM

Tom I work for the company that did "Bikini Destinations Bora Bora" (sort of comparable) we used to sell to HD Net and they would only accept HDCAM. So it wouldn't make a difference what you shot in as long as it was a sony HDCAM. They have the 1080i standard, and I've heard of some OTHER companies that would actually shoot with smaller cams (think Z1U) and insert the footage onto HDCAM and no one ever new the difference. OF course MY company wouldn't do anything like that, but let's not anyone overlook the power of good post correction on a prosumer camera...

Michael Pappas October 6th, 2005 05:39 PM

The XL-H1 is a professional camera head. HD-SDI 4:2:2 1080i is anything but "consumer". Those specs can only be found on professional line level of products. The lens is "prosumer". The LCD VF & Mic is "prosumer". But the H1 core head is a professional broadcast system.

Pappas

Nick Hiltgen October 6th, 2005 06:29 PM

HC1? DO you mean the XLH1 are they the same? I thought the HC1 was sony?

Michael Pappas October 6th, 2005 06:45 PM

Quote:

Originally Posted by Nick Hiltgen
HC1? DO you mean the XLH1 are they the same? I thought the HC1 was sony?


Thanks Nick! Fixed it.....

Barlow Elton October 8th, 2005 07:56 PM

Kaku,

Like so many others, I want to sincerely thank you for making your clips available on the web. To me, the material is perfectly adequate in showing what the camera is capable of, because if it looks sharp, detailed, colorful, with filmic motion (24f) and (relatively) artifact free in simple "run-and-gun" scenarios, then I can deduce that the technology will absolutely excel in production scenarios.

I've been watching the clips on both hi-res CRT and on a 12 ft. HD projection screen (720p projector) and all I can say is I think Canon's hit a home run. The HDV stuff compares quite favorably to HDCAM and DVCPRO HD material I have on my system. It's really sharp and I can only imagine what a good SDI acquisition might look like. Just the option of it is amazing.

I've been talking with a good friend and we've both come to the conclusion that Canon has taken dead-aim at the F900, and come up with 99% of what that camera really provides. I think the H1 is the true "poor man's" CineAlta--and most definitely the highest bang-for-the-buck in this "affordable" HD realm...at least until Panasonic can prove otherwise.

Tom Roper October 8th, 2005 09:32 PM

Looking at the 24F clips, they are all 29.97 frame rate actual. Stepping through frame by frame with an mpeg editor, every 5th frame is repeated. That's where the stutter gets in. It totally corrupts the motion, fluidity of the water in the video scene by the bay.

On the other hand, Bikeseq24 is actually not 24F, but true 29.97 fps. It has wonderful fluidity and continuity.

The crisp detail and lack of motion breakup is the best I've seen from HDV. The shadow detail is the best also. I think it could use some adjustments to the gamma and color matrix.

I think you'd have to avoid the XL H1 if your priority was 24 fps. But for "live quality" 1080i, it's impressive already and could be a knockout with some tweaks to the color.

Tony Tibbetts October 8th, 2005 09:59 PM

Quote:

Originally Posted by Tom Roper
I think you'd have to avoid the XL H1 if your priority was 24 fps.

Why? I'm sure most of the major NLE's will have a pulldown method for extracting only the 24 frames needed.

Nate Weaver October 9th, 2005 09:32 AM

Quote:

Originally Posted by Tom Roper
Looking at the 24F clips, they are all 29.97 frame rate actual. Stepping through frame by frame with an mpeg editor, every 5th frame is repeated. That's where the stutter gets in. It totally corrupts the motion, fluidity of the water in the video scene by the bay.

Tom, that's not really true. The Canon 24F stream only contains 24 frames per second, and uses repeat flags to tell the decoder when to insert redundant fields to add pulldown.

It's up to the decoder to correctly display this, and not all decoders will do it right. Also, some decoders will report the stream as 29.97 (which it is after repeat flags have been accounted for), and some will report as 23.98.

The 24F files DO exhibit perfectly smooth motion if you're using a program that reads the stream right.

Tom Roper October 9th, 2005 10:34 AM

Quote:

Originally Posted by Nate Weaver
Tom, that's not really true. The Canon 24F stream only contains 24 frames per second, and uses repeat flags to tell the decoder when to insert redundant fields to add pulldown.

It's up to the decoder to correctly display this, and not all decoders will do it right. Also, some decoders will report the stream as 29.97 (which it is after repeat flags have been accounted for), and some will report as 23.98.

The 24F files DO exhibit perfectly smooth motion if you're using a program that reads the stream right.

24 fps is perfectly smooth on the motion picture cinema screen because that's the frame rate of the film projector (or 48 since each frame is flashed twice). 24 fps is inherently *not* perfectly smooth on ATSC-HDTV (or NTSC) because the sync rate is 60hz. If 24 was evenly divisible into 60, there would be no need for repeat flags or 3:2 pulldown. But since it's not, motion becomes discontinuous at the repeated field/frame, no matter what.

Nate Weaver October 9th, 2005 11:18 AM

Quote:

Originally Posted by Tom Roper
24 fps is perfectly smooth on the motion picture cinema screen because that's the frame rate of the film projector (or 48 since each frame is flashed twice). 24 fps is inherently *not* perfectly smooth on ATSC-HDTV (or NTSC) because the sync rate is 60hz. If 24 was evenly divisible into 60, there would be no need for repeat flags or 3:2 pulldown. But since it's not, motion becomes discontinuous at the repeated field/frame, no matter what.

I understand all that. It has nothing to do with what I was explaining above.

I was just trying to clarify that if you see a repeated frame while stepping through a Canon 24F stream, then that's a product of your MPEG2 decoder adding it. It's a product of how your MPEG2 decoder was adding pulldown to the 24fps MPEG stream to get to 29.97.

If I demux the Canon 24F .m2t stream using MPEGSTREAMCLIP on my Mac, my end result is a 23.98 .m2v. In this .m2v there are 24 discrete frames per second, and no more.

Tom Roper October 9th, 2005 12:45 PM

Quote:

Originally Posted by Nate Weaver
I understand all that. It has nothing to do with what I was explaining above.

I was just trying to clarify that if you see a repeated frame while stepping through a Canon 24F stream, then that's a product of your MPEG2 decoder adding it. It's a product of how your MPEG2 decoder was adding pulldown to the 24fps MPEG stream to get to 29.97.

If I demux the Canon 24F .m2t stream using MPEGSTREAMCLIP on my Mac, my end result is a 23.98 .m2v. In this .m2v there are 24 discrete frames per second, and no more.

You are not stepping through a native 23.98 m2v stream unless *you* demux it into that!

The native *Canon 24F* m2t stream that comes out of the XL H1 is @ 29.97 f/sec, same as the 60i. The 24F option adds the pulldown, the 60i doesn't need it. Canon 24F mode wouldn't play on an ATSC compliant HDTV monitor unless it did, because 23.98 sync rate (48hz) is not ATSC compliant!

Tell you what. I too will demux this clip to 23.98 (as you did), and then step through that frame by frame to evaluate whether the 24 frames are periodic. That's really the salient point about whether the 24 frame mode introduces stutter or not.

Tom Roper October 9th, 2005 01:40 PM

They are not periodic. The Canon 24F mode appears to be capturing at 60i (30 fps). You can thus drop every 5th frame and replace it with a flag to repeat the 4th, (so that the it doesn't play back speeded up), to get a filmic look. And though you get 24 discrete frames every second that way, the cadence is irregular.

To do this right, you have to have a variable frame rate like the HVX200, which this cam does not.


All times are GMT -6. The time now is 04:10 PM.

DV Info Net -- Real Names, Real People, Real Info!
1998-2024 The Digital Video Information Network