Film-Tech Forum ARCHIVE: 2K vs. 4K Dcinema

This is topic 2K vs. 4K Dcinema - Quality with DCI Specs in forum Film-Yak at Film-Tech Forum ARCHIVE.

To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=8;t=004943

Posted by Lyle Romer (Member # 1266) on 01-24-2007, 06:22 AM:

I was just thinking about this this morning. DCI Specs limit both 2K and 4K to the same data rate. What will actually look better on screen, 2K material or 4K material compressed at 4x the compression ratio?

Does anybody have the software to take a still picture with the equivalent resolution of a 4K D-cinema image, down rez it to 2K and then compress it down to a 1.3 MB (approximate size of a frame at 250 Mbps) JPEG2000 image? Basically I'd like to see what the quality difference is of the resultant image in both 2K and 4K form to see if 4K actually looks better at the much higher compression ratio.

Posted by Bobby Henderson (Member # 840) on 01-25-2007, 03:19 PM:

If the image compression is done properly the 4K material should still look better. A 4K image will have 4 times the pixel count and much better potential for image sharpness.

I don't get the point of down-sampling a 4K image to something much lower in resolution as a means of doing a comparison to 2K. Any fair comparison should use full resolution images for either format at the compression ratios described. Print both images out on good quality paper, each one at the same size. It would be like the difference between printing a photo at 72ppi or 144ppi. The higher resolution image is going to look more sharp, provided it is a good quality image and there isn't a bunch of compression artifacts trashing up the image.

It should be said 4K is pretty useless if the entire work flow (CGI, digital intermediate, etc.) is not all natively 4K based. If 2K is involved at any point in the process then the entire 4K endeavor turns into an absolute waste of time. There's no point in blowing up a downsized 2K image back up to 4K. Once you throw away 75% of the pixels in a down-rez step you won't get them back. And interpolation (up-rezzing) creates no new image detail, just blurry bull crap.

Posted by Lyle Romer (Member # 1266) on 01-26-2007, 07:34 AM:

My experiment request is to simulate the resultant quality of a 4k produced source displayed in both 2k and 4k JPEG2000 but with the image file size for each being the maximum that it could be under DCI specs.

I guess in reality, if the files are sourced and distributed at 4k, the 4k will look better on screen because the server will be downrezzing after decompression.

I'm just trying to ascertain the quality (with DCI specs of 2k vs. 4k).

Posted by Bobby Henderson (Member # 840) on 01-26-2007, 01:42 PM:

You can do that experiment, but not by down-sampling the 4K image out of its native size. If you size it down to a similar pixel count of a 2K image it won't look any better than a 2K image.

As far as actual projection goes, it makes no sense at all to feed a 2K projector a 4K file. It would just be a huge waste of hard disc real estate. Why take up 4K worth of file space when the project can only show 25% of the pixels?

There's no graceful, organic process involved on down-sampling a digital image. It's nothing like making 35mm reduction prints from 65mm filmed material. The pixel grid is absolute.