1080p 25 vs 720 p25 - PAL

Did some tests that actually seem to go contrary to what you guys are saying. Shot a strait brick wall at close range with my NTSC HVX200 (bricks are grainy and provide A LOT of detail to examin). At first I wasn't seeing much of a different untill I realized that, for this test to be as accurate as possible, I should turn detail coring all the way off. I saw a BIG difference at that point (it's amazing what the difference between +2 and -2 is). 1080p definatly seemed to retain more detail. The various flaws and wholes in the bricks managed to maintain a lot more distinctive shape in the 1080p shots compared to the uprezed 720p one. A lot of fine scratches seemed to fade away as well in the 720p one and th 1080 just seemed sharper and more defined. I'd give the 1080p one about 25% more detail overall. Not a huge amount but definatly something.

The next thing I looked at was the claim of compression and noise. I shot a somewhat dimly lit scene in Cinelike D since that's the noisiest setting of the bunch. In the case of noise though, I found that the 1080p was quite a bit worse. The 720p uprezz, no doubt because of the uprez, had softer looking noise that was much less pronounced.

I wish I could include some grabs but I don't have anywhere that'll let me upload full 1080p shots.

As for the claim of more compression, I'm not sure how accurate that is. I recall that in Barry's book, he said that the compression for both was the same. 720 is a progressive stream so even though it has half the resolution it has 60 progressive frames in full bandwith. 1080i (which progressive signals are encoded to) has 60 interlanced fields which translates to 30 progresive frames. Hence, the difference is offset. The PAL distinctions of 50i and 25p should be about the same. You'll get 3.333 mgs per frame (or 4 for PAL) in 1080i and 1.6666 mgs per frame (or 2 in PAL) for 720p. 1080 is taking twice the space but it's twice the resolution. The compression amount is the same.
 
No the datarate in 720p25 is the same per frame as 720p30, they used filler bits to match the stream unfortunately.

I used -1 in sharpness in my test pretty close to -2. I will make a new test when I get the time. Just to be sure I didn't **** it up :)
 
Back
Top