You'll need to increase your Google skills, then, I think. ;
Hey, you've got no real argument from me on that. That's why i said the 2/3 video cams aren't going away any time soon.
This is a little tricky, but doable. I think that for quality purposes, you'd never want to go all the way to 6x though. And I think you'd really want to do this in steps rather than continuously.
I refer to the minimum number of pixels to create a 1080x1920 image ~ 2.07MP
Thom, I think your punctuation skills have saved you from being extremely rude.
I completely agree John. We tested the D3s before purchasing another 5DII for broadcast promo work. Suffice it to say we chose the 5DII. Besides the reasons mentioned above, for the majority of work Canon's 30fps offers more "apparent" resolution and less noise due to the higher frame frequency (although 24p is also coming to the 5DII soon). And being progressive, there is little loss of the cinematic quality. The increased bandwidth of the 5DII not only looks better, it holds up amazingly well to heavy post processes like compositing.
I don't see this. But I spent some time trying to find optimized Picture Control settings.
I personally wouldn't shoot with audio enabled in the DSLR. Even with a mic in the hotshoe there's too much risk of camera noise. But, yes, this is something Nikon needs to address.
Indeed, I've commented on Nikon's stairstep issue. It's interesting in that it appears very angle dependent. It appears to be angle dependent and triggered by horizontal resizing issues.
D3 has already been there ;~). But for stop frame. I'm not sure any of these DSLRs are going to make it to Hollywood. Even 1080i isn't going to cut it. When they can do 2k or 4k video they might have a shot, but that requires they do video entirely differently than they're doing it today. The real question is whether the videographer doing serious DVD work, commercial work, broadcast work, is going to use them. To get the DOF isolation, yes. To shoot in really low light, yes. But those aren't nearly 100% of the things they need to do in their shoots.
Serious video shooters are using the 5DII with third party firmware. I haven't been brave enough to try it in mine, but there have been no reports of any issues using the firmware. It's held in RAM and disappears once the camera is turned off.
Which just goes to prove my Inverted Razor Theory (see other threads).
Is 2k really that much of a stretch? That's 2048x1080, i.e. take the 16:9 1080p and expand it a tad to 17:9. Only 6% more pixels.
Yes, and the first 3 points have come out of the blue and within a year of introduction, just imagine the next 3 years. The 4th point is critical also, I have seen so many low budget video productions that turn out absolutely stultifying and dreadful because they are so slow to the shot and so cumbersome and so "safe". Yet more than once I have seen beautiful work from the hands of one person alone with 5DM2, that's revolutionary and not to be taken lightly.
A true 2k camera is going to generate JPEG2000 with 12 bits per pixel, a 2.6 gamma, and 24 bit audio at 48 or 96Khz. All of those things are a bit of a stretch for Nikon at the moment (as is 1080 for that matter ;~).
I'll give you the audio, but it still seems to me like the video isn't much more of a problem than 1080p (which nikon may very well see as a problem
). JPEG2000 only adds a little bit of extra work since that's only the last step in the already 12+ bit pipeline. I'm not up on exactly how much more work it is than plain JPEG, but I imagine the fast fourier transform circuits can be expanded to 12 bits without much trouble. Maybe JPEG2000 can be handy for stills, too 
Well, first of all, Nikon doesn't do 1080i, let alone 1080P ;~). The real issue is bandwidth. No doubt it can be solved easily enough and will be, but generally you don't want to get too far ahead of the bandwidth curve because then you up costs.
True. Is it really a bandwidth limit? It would be for the 720p-> 1080p jump, but 1080p-> 2k is only a 6% boost. I'm assuming the 1080p jump will happen anyway, or that Canon could do it. JPEG2k et. al. don't add bandwidth requirements, just switch steps in the existing pipeline.
Yes, I think so. Nikon is generating JPEGs on the fly (their file format is Motion-JPEG). That means that the EXPEED ASIC is involved. So you've got all these bits coming in from the sensor, going into memory, being pulled out of memory by the ASIC, being put back into memory by the ASIC, and, oh, we have to coordinate the audio. Not that it can't be done. But the solutions may need to be faster ASIC or dedicated AVCHD processor, either of which implies more bandwidth.
I think you're forgetting the 12-bit versus 8-bit part. I believe 2k records 12-bit, which means its files are larger than the 8-bit being written in Motion-JPEG. All these little bandwidth grabs add up. Again, I'm not saying it can't be done, because obviously it is by Arriflex, RED, and others. But increasing internal bandwidth often means increasing parts costs. You prefer for that to scale naturally with Moore's law, not get ahead of the curve.
