My resolution test (amateur) for Firmware 2.0.4

[Pat Reddy]

I have been shooting in 24p with the new firmware, and the image quality and resolution seem improved over the old firmware in 30p. This makes sense given the higher bandwidth for 2.0.4 and the greater availability of that bandwidth when shooting 24p.

I downloaded the ISO 12233 chart made by Stephen Westin of Cornell, printed a relatively high resolution print, and started testing the 5D II with the new firmware. I tested the camera with a 70-200 f4/L at 100 mm and a 135 F2/L. I am seeing between 700 and 800 lines of resolution in both the horizontal and vertical. Moire and/or aliasing kick in at 600 to 700 in the vertical. The true lines persist almost to 850 in the vertical, but false lines are just as "numerous" at that point.

Interestingly, the moire and or aliasing was much more severe with the 135 F2. It's a very sharp lens, and I had it at a more optimal aperture when I tested it. (I manually focused both lenses, and may also have been off a bit with the 70-200.) I've been shooting wide with the 17-40 stopped down to f16 to f22 lately, and the images look plenty sharp and quite clean. I guess this is just another confirmation that less than perfect focus can often benefit the image quality of video for this camera.

I also tested the XH-A1 today, and it had a bit higher resolution and far less moire/aliasing (in my tests at least).

Pat

[Régine Weinberg]

U got it, the XH-A1 has three CCD's the hype on the 5d... it has a full frame CCD by the way it uses only a fraction for 1920 p the rest is for stills and it is no real camcorder.
All once started with 1 CCD and now in the highend and prosumer market we have 3 CCD.
All three colors on 1 CCd takes time to decode and makes some artifacts, less with stills as the max is 7 frames second and in the picture taken is no movement....

the 5D is not bad, but think, ok not a must, all that Zacuto gear and the price tag will be not far from a XH-A1

[Pat Reddy]

I love my XH-A1, but I'm starting to use the 5DII more for video. I have used it a lot for landscape still photos this winter. I might go ahead and spent the extra $ for a Hoodloupe or Zacuto finder and a shoulder brace. It would be nice to be able to just take one device for multiple purposes.


I still can't get over the amount of moire on the 5D test shot compared with the XH-A1. Aside from the fact that it's not a camcorder that shot explains the whole problem with the 5D. But we also know why so many will choose to use it instead of an Xh-A1 or EX1.

Pat

[Ben Denham]

The 5D is CMOS not CCD. My understanding is that these are quite different technologies that have different possibilities and limitations. My guess is that the CMOS chip in the 5D isn't the limiting factor when it comes to resolution. The explanations that I've read seem to suggest that the 5D is more limited by the algorithms that take us from 21 to 2 megapixels. Please correct me if you think I'm misunderstanding something here.

[Pat Reddy]

Your right, and the 5DII sensor really excels at low light compared to most camcorders we talk about around here. The 5DII microprocessor samples rows to get to around 2 MP for video and can't do it fast enough to avoid some aliasing. It is also my understanding that the implementation of the codec is not all that efficient. So while 38 Mbps (the old firmware) and 44 Mbps sound pretty good compared with the 25 Mbps of HDV, the codec may not be able to handle as much information in detailed scenes or scenes with high motion as you might expect. The combined result is aliasing and moire and some smudging of detail.

The new firmware seems to be an improvement in this regard. I don't think that my little test proves this (I didn't test the resolution with the old firmware, and it's difficult to compare these results with other tests), but it at least satisfied some curiosity I have had lately. I would be interested to hear what others are seeing.

Pat

[Evan Donn]

I've had an XHA1 for a few years now, and the 5DmkII since it started shipping. I think I'm finally ready to sell the XHA1 - I only used it 3 times last year, all times when I wanted a second fixed angle that I could let run without monitoring.

Resolution isn't everything...

Aliasing is a problem to be aware of with the 5D, but once you're aware of it's not too hard to avoid. Noise is a big difference though, and that's where the XHA1 falls really short in comparison. It simply needs more light, and exhibits more noise even in situations that I wouldn't consider particularly dark. There's absolutely no comparison when shooting with fast lenses on the 5D.

And lenses are a big part of the equation too - there's simply no option for an XHA1 similar to Canon's 14mm f/2.8, or any tilt shift lens. Even just a good true wide angle requires a heavy WA adapter. Sure you can use any of these lenses with a 35mm adapter, but now your camera is really cumbersome, there's a lot more things to go wrong, and you've made the low light performance even worse. Oh, and added quite a bit to the cost. And if you want shallow DOF that's the route you have to go.

Speaking of size - the 5D with a normal prime is about 1/3 the size of the XHA1, even with a loupe/viewfinder. It's lighter and easier to carry around, and you can shoot with it without drawing nearly as much attention as with a bigger video camera.

My experience with the codec has been largely positive. I think it actually holds up better than HDV to high motion, although neither is perfect. I also find I have a lot of room to push things in Color (after conversion to prores), more so than with the XHA1's footage - not sure if that's due to the codec or just the camera's greater latitude.

So this isn't to say that the 5D is better than the XHA1 - only that both cameras have their shortcomings and strengths, and depending on your needs you may find that differences in something like resolution aren't that important when other features are considered. For me, the strengths of the 5D outweigh it's weaknesses and make it an easy choice over the XHA1.