1080/24p Is More Detailed Than 720 Mode

toke lahti said:
So which one is it: pick-ups only or also compositing shots?
Is this another marketing gimmick?
The movie was shot last summer!
Are you saying that Spielberg & co. went back to shoot some more in last November with prototype hvx?
And then rushed back to make a new version of the movie?
They had to begin to make release prints in December, because it premiered in Europe in January in almost every country. Thousands of copies was needed.
Click to expand...

They even shot with a Sony Hi8 camcorder for a Spielberg film, -with extreme work of 2K specialists...- it appears in the movie, ( it was a scene with a homemovie shown in the background on a tv set while the rest was shot 35mm film) :)
 
Hey, we can all scoff if we want, but we just wish we could shoot on Hi8 and get it projected in every theater in the entire nation, not to mention getting the nod for Best Picture academy award. Seriously, Spielberg would be a genius even if he was shooting through a cardboard box with a pinhole in it. Some people are just great filmmakers.
 
Alex Leith said:
Thanks Kevin for doing that. It's nice to see some uncompressed footage.
Click to expand...
The footage was presented in DVCPROHD. That is hardly uncompressed. Better than the .wmv or h.264 stuff for sure, but not uncompressed. Unfortunately, none of us PC users can even see it unless it's in one of the other formats.

holditnowfilms said:
Has anyone seen HD TV? It is so compressed and bastardized prior to distribution nobody but the editor is ever going to see the pristine picture.
Click to expand...
Broadcast HDTV, yes. But the upcoming high definition optical formats are a completely different story. Blu-Ray is a 1080p format with far superior bitrate that gets eons closer to that "pristine picture" you're shooting than broadcast HDTV ever will. I plan to distribute most of my work, not broadcast it. Things like resolution and compression artifacts during the acquisition process are very important to me.

penfever said:
Seriously, Spielberg would be a genius even if he was shooting through a cardboard box with a pinhole in it.
Click to expand...
Completely subjective.
 
ullanta said:
Whew! That took about 8 hours to download at something like 2K per second with lots of stalls... kind of old-school and fun.

Anyway, yes, this test clearly shows much more resolution in the 1080p image. What is that resolution? Who knows, it's not a chart (thank g-d)! But 1080 shows far more fabric detail, most noticeable in the furthest-zoomed-in shot, near the top of the image.

Unfortunately, the exposure is a little different - the 720 shot is darker, so it's tough to get a perfectg feeling for the differences... but the fabric detail differences are extreamly clear; there are large sections on the 720 image where you see mush, but in 1080 you see clear weave.
Click to expand...

I agree. The reason for the exposure difference is the difference in sensitivity between 1080 30p and 720 24p. The iris is set exactly the same for both shots.

Sorry about the long download times. You are probably one of the only people that waited it out. The clips are on my development server at home which runs off of a DSL line. There have been a ton of people downloading the clips. Or at least attempting to.

If someone has bandwidth to spare they can feel free to re-post them or let me know where to FTP the files.

K
 
I think people could get just as much out of one or two still images as from the video; you may want to use those instead...

Anyway, thanks for posting this. Again it says to me that there's something going on in the HVX that makes chart-type, monochrome resolution a weak point. I haven't seen good comparisons (to other cameras) of other material... but from looking at different bits of footage, perceived resolution and overall image quality of the HVX is much better than charts indicate. It's a bit odd, but i can imagine several scenarios that would account for this. In any case, Kevin's test is certainly good evidence that there's a difference in real-world material between 1080 and 720.
 
ullanta said:
I think people could get just as much out of one or two still images as from the video; you may want to use those instead...

Anyway, thanks for posting this. Again it says to me that there's something going on in the HVX that makes chart-type, monochrome resolution a weak point. I haven't seen good comparisons (to other cameras) of other material... but from looking at different bits of footage, perceived resolution and overall image quality of the HVX is much better than charts indicate. It's a bit odd, but i can imagine several scenarios that would account for this. In any case, Kevin's test is certainly good evidence that there's a difference in real-world material between 1080 and 720.
Click to expand...

I will post frame grabs tonight when I get home. My reason for posting the clips is to keep from recompressing the images. They were in DVCPROHD from beginning to end with no adjustment or recompression as to give a true indication of what the camera puts out.

I will save the stills as .tifs (instead of jpgs) with no compression and that should give a pretty true representation.
 
Lossless PNG would be much better. And JPEG max quality will not deteriorate the images or anything:) Do a test.
 
Yeah, lossless PNG's would be great. Might want to export 'em in two different sizes, as native (1280x1080 & 960x720) and as aspect-ratio-corrected 1920x1080 and 1280x720.
 
Barry_Green said:
Yeah, lossless PNG's would be great. Might want to export 'em in two different sizes, as native (1280x1080 & 960x720) and as aspect-ratio-corrected 1920x1080 and 1280x720.
Click to expand...

Ok - here are the links to the still frames exported from the original HVX clips.

http://homepage.mac.com/kphoto/hvx/960x720a.png
http://homepage.mac.com/kphoto/hvx/960x720b.png
http://homepage.mac.com/kphoto/hvx/1280x720a.png
http://homepage.mac.com/kphoto/hvx/1280x720b.png
http://homepage.mac.com/kphoto/hvx/1280x1080a.png
http://homepage.mac.com/kphoto/hvx/1920x1080b.png

Each size has an a and b version, a is wider angle and b is zoomed in all the way.
They are all uncompressed png files as requested by Barry.

Let me know if there is anything more I can do for you. Wash your car? Walk your dog?

Just kidding. Happy to do it.

I'm interested to hear thoughts on this camera based on the clips and stills.

I owned an FX1 prior to getting the HVX and I am never going back. Don't get me wrong, the FX1 was a fine camera, etc, etc.

K
 
kevinM said:
I'm interested to hear thoughts on this camera based on the clips and stills.
Click to expand...

Thanks Kevin, this was very helpful!

I thought it would be interesting to do a side by side compare of the the 720 and 1080 "a" images, so I did a color balance on the 720 clip to match the 1080, as well as enlarge and adjust it to overlay and match to size. I used the anamorphic 720 by 960 and the 1280 by 1080, not the scaled to correct aspect ratio versions, just to keep the quality of FCP scaling out of the equation;

Here is the result:

http://ftp.datausa.com/imageshoppe/images/DETAIL_1080vrs720.png

I also need to mention again, that all the image processing adjustments were done to the 720 image, leaving the 1080 virgin, doing my best to favor it in the test as much as possible.

Of course, this test would be even better if the camera was locked off between grabs so the image sizes were the same, and also you could use the shutter to compensate for exposure, but this is a good start to testing.

Regards,

Jim Arthurs
 
That's interesting.

I'm assuming the enlarged 720 image is the one on the left. It held up better than I expected. However, the difference is quite visable in the small type on the back of the iPod.

Did you do any sharpening on the 720 image?

Are you thinking that the HVX may be doing an internal uprez as some are saying?

After using the camera in different types of situations and seeing the difference on moving footage, I would be surprized if this was the case.
 
Emanuel said:
I'd like too but as......it was said (you're welcome as always -- after some silence), only if you will do it@our_disposal 'cause after 3h of downloading it was so much frustrating to see nothing. Specially regarding an important post like yours:

Thanks in advance!
Click to expand...

I'd be glad to post the full clips in another format as long as I can do it out of Final Cut Pro. As far as I know, I can't do Windows Media files from FCP but if anwone knows how to do it, I'll give it a shot.

Just tell me what you want. However, any other format will be recompressing the clips so the results may be affected.
 
Thanks Kevin! It will be very useful.

Maybe following Haakon's tip (BTW, you're welcome -- after some silence, Haakon and your thoughts about the recent posted footage such as Steev's or this one, it would be so much appreciated):
Haakon said:
The footage was presented in DVCPROHD. That is hardly uncompressed. Better than the .wmv or h.264 stuff for sure, but not uncompressed. Unfortunately, none of us PC users can even see it unless it's in one of the other formats.
Click to expand...

EDIT -- By keyboard handling error, the prior post was deleted; that's why I re-post again and BTW I'm grateful for your care, Kevin. Unfortunately, I can't help but I'm sure it will not be difficult.
 
No, I did no sharpening on the 720 portion of the image. In fact, it's a bit softer, as it's enlarged/scaled with Photoshop bicubic. As I said, all the damage I did, was only to the 720 side to bring it up to match the 1080 in size and color balance.

Nothing conclusive, but it was as fair a test as could be done with the given frames.

Again, thanks for making these available. I hope to return the favor to the community when I can...

Jim Arthurs
 
Jim Arthurs said:
Thanks Kevin, this was very helpful!

I thought it would be interesting to do a side by side compare of the the 720 and 1080 "a" images, so I did a color balance on the 720 clip to match the 1080, as well as enlarge and adjust it to overlay and match to size. I used the anamorphic 720 by 960 and the 1280 by 1080, not the scaled to correct aspect ratio versions, just to keep the quality of FCP scaling out of the equation;

Here is the result:

http://ftp.datausa.com/imageshoppe/images/DETAIL_1080vrs720.png

I also need to mention again, that all the image processing adjustments were done to the 720 image, leaving the 1080 virgin, doing my best to favor it in the test as much as possible.

Of course, this test would be even better if the camera was locked off between grabs so the image sizes were the same, and also you could use the shutter to compensate for exposure, but this is a good start to testing.

Regards,

Jim Arthurs
Click to expand...
Which is which?

Left's 720p and Right's 1080p? Am I correct?
 
So, there's a difference and it's not true that there isn't a real resolution increase from 720p to 1080p. Maybe not so unnoticeable if to theatrical screen...

Where's the beef PappasArts?
 
filmmaker1977 said:
Where's the beef PappasArts?
Click to expand...




To me they look the same. Atleast to close to call it an advantage.

As for the the beef, I am the messenger on this one. That's what PlasterCity told the audience about 1080P vs 720P at Moviola.

I could care less anymore, nothing is going to come from this. It's a 720P camera IMHO. That's fine by me.

The HVX is still one bitchin camera no matter what. It has an outstanding look that no other camera on the market in this class has an out of the box film look like the HVX200..... Colors are gorgeous and it's DVCPROHD etc etc etc!
 
Pappas, I'd prefer hearing that this camera has a look that no other camera in this class has, not just in an "out of the box" undertone.

My reasoning for saying this is because as much as I like the H1's sharpness, it doesn't compete in any other dept. as far as I can see.

It's pretty clear that the HVX is the all around champ of this class.
 
You must log in or register to reply here.
Back
Top