CineForm

Ernesto,

The trial peroid is 15-days from when you first install it, you don't get additional 15-days by downloading it again as you have found (just stating it so others don't get messed up.) For users who make this mistake there are methods that allow the support staff to issue an additional 15-days. You would normally have to go through support, but I will private measage you a download with 15 more days.
 
David. I'm assuming the new Mac Pro towers will run CineForm under bootcamp. Which graphics card would you recommend?

This is what I've been waiting for. Thanks.
 
Yes the Woodcrest Intel Xeons are awesome so Prospect HD will work great under bootcamp. For graphics cards, any modern dual-head ATI or NVidia card will work for desktop playback. I prefer ATI as it YUV video levels are correctly calibrated out of the box (not so with many NVidia cards.)
 
Well, I just downloaded a new trial and I'm ready to produce a little piece in order to get things together about a month from now, when we'll begin a real production, so can you tell me what settings for camera? (HVX, 1080 24p)
I need to transfer the final work to 35 mm

what editing workflow? (Premiere Pro 2, 2.2 GHZ AMD 64, 2 GB Ram, 2 sata raid 0)

Thanks a lot

PS: I know that the answer to this must be anywhere in this forum, but, you know...

:Drogar-BigGrin(DBG)
 
With HDLink you will want "Remove 3-2 Pulldown" checked, as the 1080p24 modes actually use 60i. Set the frame format to progressive, and the quality to High. No other filters need to be active. You should consider resizing the video as the capture format of 1280x1080 is not commonly used. Either scale to 1440x1080, due to HDV this pixel format is now common, or scale down to 1280x720. Scaling down has some advantages with the HXV, as it resolves less than 720 lines vertical anyway, why waste the disk space (no real quality loss.) The down scaling helps reduce noise and gives you more editing performance. Plus 1280x720 is square pixel, makes effects work easier. A high quality 1280x720 workflow can tranfer to 35mm just as well as the 1280x1080 source.
 
Hi David

well I just finish my first real test on Aspect HD with Premiere Pro 2.0

It work well, slow but well, I notice some odd behavior in clips converted, but it can be a leak of experience, I going to "play" a little more with it.

The clips show like relented and motion blurred, my test was recorded at 1080/24p, converted at 1080 and also 720, pulldown removed.

Also Rec at 720 24pA and 720 24pN, showing the same "effect".

I think that I have to set properly the shutter, going to see it.

Mean while I have some questions about it:

You said:

David Newman said:
... or scale down to 1280x720. Scaling down has some advantages with the HXV, as it resolves less than 720 lines vertical anyway, why waste the disk space (no real quality loss.) The down scaling helps reduce noise and gives you more editing performance. Plus 1280x720 is square pixel, makes effects work easier. A high quality 1280x720 workflow can tranfer to 35mm just as well as the 1280x1080 source.
Click to expand...

but in DvFilms, wich is the place where we usually transfer to 35mm, the say:

"Recommended Settings for the Panasonic HVX200 and Transfer to Film


Shoot 1080/24PA format (1080/25P for European model) with 1/48th sec (or 1/50th sec) shutter speed.

Or

Shoot 720/24pn (720/25pn for European model), 1/48th sec (or 1/50 sec) shutter speed.

The 720-line format will have about 20-30% reduced resolution on a transfer to film vs. the 1080 line format.

Other settings are up to the user and can be accurately judged using an HD monitor with greater than 700 lines resolution...


... 1080/24PA (Advanced, or 2:3:3:2 Pulldown)

The Panasonic HVX200 can record 1920 x 1080 at 24P (actually 23.976), with 2:3:3:2 (Advanced) pulldown, and this is the recommended recording mode for transfer to film.

The movie is stored on the P2 card as a 1920 x 1080i MXF file, at 29.97 frames/sec for compatibility with standard HD editing systems such as FCP-HD and Avid Express Pro HD. "

it is in contradiction with you, can you tell me something about?

PS: especially..."A high quality 1280x720 workflow can transfer to 35mm just as well as the 1280x1080 source" in contrast with: ..."The 720-line format will have about 20-30% reduced resolution on a transfer to film vs. the 1080 line format"


:shocked:

Thanks, David
 
erdiaz said:
Hi David

well I just finish my first real test on Aspect HD with Premiere Pro 2.0

It work well, slow but well, I notice some odd behavior in clips converted, but it can be a lack of experience, I going to "play" a little more with it.
Click to expand...

We don't add any motion blur, so either that is on your source or your display device is doing it. If you place the resulting AVI into a free tool like VirtualDub you can see each frame pixel for pixel. The CineForm AVI are visually identical to the source. Make sure you haven't selected "de-interlace" as that can blend fields, as you are shooting 1080p24 you don't want that.

The 720-line format will have about 20-30% reduced resolution on a transfer to film vs. the 1080 line format.
...
... 1080/24PA (Advanced, or 2:3:3:2 Pulldown)

The Panasonic HVX200 can record 1920 x 1080 at 24P (actually 23.976), with 2:3:3:2 (Advanced) pulldown, and this is the recommended recording mode for transfer to film.

The movie is stored on the P2 card as a 1920 x 1080i MXF file, at 29.97 frames/sec for compatibility with standard HD editing systems such as FCP-HD and Avid Express Pro HD. "

it is in contradiction with you, can you tell me something about?

PS: especially..."A high quality 1280x720 workflow can transfer to 35mm just as well as the 1280x1080 source" in contrast with: ..."The 720-line format will have about 20-30% reduced resolution on a transfer to film vs. the 1080 line format"
Click to expand...

Yes, you have quoted from another source is not completely correct. The camera does NOT record 1920x1080, the DVCPRO-HD data within the MXF only contains 1280x1080 (which is only 2/3rds that of full 1080p.) If the camera could resolve near the 1280x1080, then editing in that mode will make a lot of sense, but it doesn't. As for "20-30%" addition res than the 1080p offers, that is primarily in the horizontal direction, which is why I make my recommendation. The camera only uses 960x540 sensors. Due to pixel shifting the camera will resolve a little more than the sensor would suggest, about 30% more in each direction (testing shows slightly less than this.) But using the 30% estimate the 960x540 sensor will generate a 1248 x 702 image (pretty closes to standard 1280x720.) When you use the 720pN mode the camera only stores 960x720, so you do lose horizontal information when in HVX200's 720p. However, using CineForm tools you are not limited to 1280x1080 or 960x720, you can resize to more standard formats. Resizing the 1280x1080 image (that only has under 700 lines of information) to 1280x720 doesn't lose vertical detail, while it maintains the horizontal advantage for the DVCPRO-HD 1080 mode. Does that make sense?

Do the test for yourself. Import a 1080p clip convert it once (pulldown on) but no rescaling (so you get 1280x1080) then convert it again with resize to 1280x720. Then load the new scaled 720p clip onto VirtualDub and rescale back up to 1280x1080p (using the filter "resize" with the "Lanczos3" Filter mode.) Then compare the two, you will find it is difficult to tell the difference.

I did the test with a 1080p24 sequence that is all high frequency, this is the worst case senario, so you can see a subtle difference if you know what to look for -- I would still chose resizing to 1280x720 when shooting the HVX200.

Now this is not a glamous scene, shoot though a window into a bunch of trees. Reader, can you tell which is the 1280x1080 source and which has been resized to 1280x720p then scale back to 1280x1080?

Image A : http://www.cineform.com/examples/hvx_scaling_a.png
Image B : http://www.cineform.com/examples/hvx_scaling_b.png
 
Last edited:
Yes, I did what you said and yes, you are right, so it leads me to conclude that it is better to shoot 720 pN, convert to 1280 x 720, edit it and send to the film house in this format.

Do you have any advise about camera settings? or may be this is a question for Barry

I'm going to do that test and tell you about.

Thank a lot again Dave.
 
Last edited:
Ernesto, my argument does still suggest shooting 1280x1080 and scaling down for the slight improvement in horizontal resolution over the 960x720 modes. Although I understand why the 720pN modes are compelling. Here is an additional argument for shooting 1080p and resizing to 720p. The 720p mode is actually heavily compressed at only 40Mbits/s for 24p, I can even argue that it is more compression than the 720p HDV equivalent. Yet shooting 1080p for the same detail, gives you 80Mbits/s of frame information (20Mb/s is lost to redundant fields.) So you are get half the compression artifacts using the 1080p24 mode -- that has value for a film project.

P.S. Where you able to tell the difference between image A and image B?

Image A : http://www.cineform.com/examples/hvx_scaling_a.png
Image B : http://www.cineform.com/examples/hvx_scaling_b.png
Click to expand...
 
Gosh, those pictures look familiar!!!!

It was raining that day and I didn't want to get my camera wet!

Sorry they weren't more exciting.
 
Sorry Mike, I would have credited them to you but I forget how I got those images and they were all I had on my home PC. :) It doesn't matter that the aren't exciting as they work great for proving the HVX's resolution limitations.
 
David Newman said:
P.S. Where you able to tell the difference between image A and image B?
Click to expand...
David,

Great thread, CineForm Connect HD will be my entry to editing HVX footage in Vegas. Thereafter, perhaps P Pro, Prospect HD and a Xena card.

ok, I will have a go at answering and distill my observations:

1. There is a slight difference between the 'a' and 'b' images, the difference is here. Sorry, its a jpg, no png allowed, so not a true difference key, as it has been compressed. The difference between image 'a' and 'b' is not perceptible to the human eye, imo.

2. Loading image 'a' and 'b' on top of each other in a compositor or PS, zoomed in to say 1600%, you see that the images are basically 1 pixel out of alignment in the vertical axis, part of the truncation/scaling, totally normal imo because of the down and up scaling.

3. If I was to guess, i would say 'b' is the 1280x1080 source file, and 'a' was the resized to 1280x720p then scaled back to 1280x1080. My reason, very hard to tell, especially trees which are hard for human eye to perceive, though a great compression tester imo, however, zoomed in at 1600% and a/b referencing on/off in alignment, looking at the pattern in the pixels, 'a' image gave the impression of slightest scaling softness. But who looks at still frames at 1600% and a/b tests them? CineForm has my interest and attention.

So David, was I wrong? Or right answer, wrong theory?

May I ask David, does CineForm use one of the scaling methods available in compositing applications or is it proprietary? Name?

Thanks for the education David, clears up some things I was not sure about and learned some others.

imo, CineForm = :thumbsup:
 
mmm, after thinking and considering CineForm may use pixel morphing type scaling, not typical scaling, the effects may also result in a type of image sharpening, especially when I refer to observation at 1600% zoom. So now I am not sure, perhaps 'a' is the source file....

The answer is 'visually lossless'. Put me out of my misery and please disclose David. :crybaby: Thanks.
 
David Newman said:
Ernesto, my argument does still suggest shooting 1280x1080 and scaling down for the slight improvement in horizontal resolution over the 960x720 modes. Although I understand why the 720pN modes are compelling. Here is an additional argument for shooting 1080p and resizing to 720p. The 720p mode is actually heavily compressed at only 40Mbits/s for 24p, I can even argue that it is more compression than the 720p HDV equivalent. Yet shooting 1080p for the same detail, gives you 80Mbits/s of frame information (20Mb/s is lost to redundant fields.) So you are get half the compression artifacts using the 1080p24 mode -- that has value for a film project..
Click to expand...

Now I understand !!! thanks David, it was not clear to me untill you reveal your argument.

So I'm going to shoot 1080p, and 720pN only for slow motion (if I need)

Thanks again and again!!!

:dankk2: :dankk2:

PS: Barry, what do you say about it? I want to know you opinion
 
Last edited:
I love the "its A" no "its B" posts. Yes is it hard to tell, proving the point that editing at 1280x720 is the way to go (if you can handle the short recording times of 1080p24.) The answer is 'B' is the source and 'A' is the twice scaled image (down then up again.) Yes the half pixel shift is the main give away that scaling up is involved -- but all the source detail is preserved. I won't comment on our scaling algorithms as they do change from time to time as put even better tech, they are pretty good today.
 
David Newman said:
I love the "its A" no "its B" posts.
Click to expand...
After I had pressed the send button on the 1st reply, I said to myself: "Why didn't I take the blue pill?"

In the end, I should have not sent the 2nd message and kept my mouth shut. :D
 
I LOVE this camera and LOVE Cineform. Wow what a great duo. We are shooting a long project for United Way and using cranking at 60pfs to capture the faces of people helped by United Way. We had a tough one at a preschool, lots of children, overhead fluorescencent, white walls. Our lighting guy killed the overhead lights and used only two Lowel Totas, no gels, bounced off the ceiling. One Lowel Omni with Follies pink gel, close to the white wall just out of frame on the left as a wash. The HVX just LOVES (boy I use that world a lot) soft light. Here is a still (unretouched) right off the Cineform Premiere Pro time line.
Hands.jpg
 
You must log in or register to reply here.
Back
Top