HVX200a 1080/24pA or 720/24pN upconverted to 1080 looks the same

M

mrbrycel

Well-known member
I know the HVX's 1080 vs. 720 controversy has been gone over too many times, which is why I didn't put this in the HVX forum. I didn't want to anger anyone.

But owning only one 16GB P2 card, I've been giving a lot of thought to how important it is to shoot in 1080 over 720 in situations where I don't need variable frame rates. Of course it always seems as though 1080 would be the best way to go, since its a higher resolution. But since in 720 mode, the image is still being taken from a 1080 scan, I wondered how different a 1080 scanned/720 recorded/1080 output image was from a 1080 scanned/1080 recorded/1080 output image.

I took a shot at 1080/24pA, and then the same shot at 720/24pN. I dragged them both into a 1080/24p sequence in FCP, then exported it and viewed it through quicktime fullscreen on my 20in. Mac display. I could not see the difference. The 720 recorded image did not really look any softer than the 1080 recorded image. Maybe quicktime/mac display isn't the best way to compare, or maybe my shot wasn't the best one to test, but it looked the same.

Now looking at it as 16 minutes vs. 42 minutes, I'm wondering if recording in 1080 is worth the hassle of offloading, or the hard drive space. Can anyone tell me their personal experiences with these two, or any advice on which to shoot on?
 
Most people would agree with you that the image quality improvement of 1080 isn't worth the 2.5x space increase.

From my observations the 1080 mode adds about 15% more resolution in the horizontal direction, and not much in the vertical. But there is significantly more chroma resolution (640 samples instead of 480, for a 33% increase) and much less compression in the 1080 mode. Again, you'll have to determine whether it makes enough difference to justify the space increase. For maximum sharpness and quality I always use the 1080 mode, but many still feel that 720/24pN is the "sweet spot" for the HVX/HPX170.
 
The chroma resolution point Barry made is an interesting one. Would this lead to noticeably better results color correcting in Color? Could someone more technically minded than myself explain a little more about how the chroma resolution increases by 33% while the image resolution increases by only 15%. This seems counter-intuitive to my non-engineer's mind.

Thanks,
Kevin
 
It's because chroma is recorded at a different resolution than luma.

The chips are capable of delivering luma resolution in the realm of about 1440 x 810. The codec then cuts the chroma resolution in half, to 4:2:2. So there's a whole lot more potential chroma resolution in the source image than what gets finally recorded; as you add room to record more data more of the original chroma information can be recorded.

In addition, DVCPRO-HD prefilters the horizontal resolution, so you end up with 960x720 for 720p. And then 4:2:2 chroma sampling happens, so that means a net effective 480 x 720 for chroma resolution.

But in 1080 mode, it prefilters to 1280x1080, and therefore chroma resolution is recorded at 640x1080. The vertical detail is about the same (there's not all that much difference between the 800 lines of detail the chips are generating, and the 720 lines that 720p records) but horizontally you go from 480 chroma samples to 640 chroma samples. You're starting with about 1400 chroma samples to begin with, which get prefiltered and sampled down to either 480 or 640. Obviously more will survive more accurately in the 640 version.
 
I would also like to see a comparison of both clips zoomed in post; say 10-15%, and then see if you can tell a difference.
 
You must log in or register to reply here.
Back
Top