AF100 / AF101 - initial report

The question (AIUI) is: would autofocus be of any value in a scenario like that? Is it "smart" enough to know to focus on the eyes?

And that, I don't know the answer to.

Thank you, yes. I mean AF would have to be pretty trustworthy.

Rather than Canon's face detection, I have come to rely on focus assist if I can't otherwise see it on a less-than-HD LCD. So, yay for the AF100 LCD.

And referring to your earlier report on focus assist, I'm wondering how difficult that would be to implement. Don't the $800 Canon HV models have it?
 
Actually, the raw data off the CMOS is going to be analog, so the question is, what is the bit depth of the A2D pipeline off the sensor. If this were 8 bit then from an acquisition standpoint, you'd be hardware limited. I have absolutely no insight to their A2D circuitry, just pointing out that it could be a hardware limit. Sure, you could pack that in 10-bit but you aren't gaining anything except wasted space :)

Now, if they do have higher resolution A2D, then it becomes a question if they do any sort of full hardware compression down into 8-bit. The reason being, hardware is ALWAYS faster than software in the same technology node, but again, who knows.

I would be surprised and disappointed if their A/D from the sensor was less than 12 bit.
 
I would be surprised and disappointed if their A/D from the sensor was less than 12 bit.

I'd actually be surprised as well :) BUT, I am not sure how much of the lower end is eaten up by converter noise. That's one of the big problems with analog is the CMOS sensor will have a noise floor and the converters will have a noise floor. Also, if they use any gain stages that are analog between them, they have noise as well.
 
I would be surprised and disappointed if their A/D from the sensor was less than 12 bit.
I think the HPX170 has a 19-bit A-D converter, so it would be quite a step backwards for the AF100 to have 12. Of course, we won't know for sure until they announce specs.
 
I am with Barry. I am DONE with the DSLR foolishness.

I don't know. I still think having a DSLR around for stills and for certain shots that the AF100 can't do, is gonna be useful.

I plan on keeping my T2i as a complimentary camera to the AF100, and even upgrading when the inevitable T3i comes out.

Also, a DSLR would be nice for location scouting. You can snap location scout pics and shoot video clips, all in a small little package.
 
Agreed that for certain purposes, such as location scouting etc., a small compact body would be good to have. I plan on keeping the GH13 for such purposes.
 
I'll keep the gh13 if I can but will probably have to sell it to pay for af100. How can I get rid of the legendary gh13 though. Will be hard.
 
19 bit!? why are we all asking for 10 bit then? no wonder they only gave us 8! you ask for 19 and they give you 10
seriously though we need 19 bit output to "futureproof" this camera
 
this is exactly the sort of thing that makes film so much better than video (if you can afford it) you make a movie now, then make a dvd of it and watch it or whatever, but in 20 years when all the tvs have 16bit colour depth you just rescan your film and it looks awesome compared to the dated video footage shot on 8, 10 and 12 bit cameras from 2010.
 
this is exactly the sort of thing that makes film so much better than video (if you can afford it) you make a movie now, then make a dvd of it and watch it or whatever, but in 20 years when all the tvs have 16bit colour depth you just rescan your film and it looks awesome compared to the dated video footage shot on 8, 10 and 12 bit cameras from 2010.

This is also why high resolution RAW motion acquisition is such a big deal. It doesn't limit you to the format and encoding standards of the present like a video camera.
 
Can I just say that I'm so freekin excited about the capabilities of this camera at this price point?

C'mon...this really is exciting!
 
7D and 5D both have more DR than the original Red One (although no Red user will ever admit it). The new MX has about 13 stops so it does have more then the 7D/5D but mostly in the highlights. Red defiantly has more Latitude (i.e. "exposure compensation", thanks to RAW format) but there is nothing special about the "Red" name-tag that makes it have magically more dynamic range than any other camera. Dynamic range is all about sensor size... Bigger sensors soak-up more light (unless they're engineered poorly).

Take a look at this video if you haven't see it already. Red, 5D, and 7D were all used in this... and you can't really tell any shots apart. Probably at 1080p you could... but the point is all the shots capture the flames in the same way. If the 7D and 5D really had noticeably less DR, it would be a huge jump in the exposure range from shot to shot.

http://vimeo.com/14681630

I'm a little late to the party, but the 7D and 5Dm2 do NOT have more dynamic range than Red One. According to dpreview, the DR for the 7D is 8.3 stops, and for the 5Dm2 is 8.4 stops. For comparison, Red One is about 11.3 stops (according to Red). Even the Sony EX1, with a DR of 10 stops (according to ProVideo Coalition) is higher than the DSLRs.

Dynamic range is not just about sensor size. Bigger pixels can soak up more light, but if they saturate quicker than smaller pixels, then they don't have better DR.
 
Yes, I will keep the 7D for stills and scouts and family videos I just will not use it any professional video situation. I have already moved all my paying gigs back to the HPX-500. It is more reliable, more stable and easier to shoot with.

I will be one of the first in line for the AF100.
 
19 bit!? why are we all asking for 10 bit then? no wonder they only gave us 8! you ask for 19 and they give you 10
seriously though we need 19 bit output to "futureproof" this camera

19 bit is a very odd number, but it probably magically happened when they turned on their sensor for the first time, but anything over 16 bit for acquisition can really be considered just a waste of hard drive space. you only need to get into higher levels of bit depth when your doing mass 3d/CGI/VFX, where you have to continually work at several times everyday quality to reduce artifacts, aliasing, and transcode degradation. For the digital video recording realm, 16 bit is gold, more than that, it's platinum.

you work in higher bit depths so you don't lose anything and gives you wiggle room to do a lot of manipulations, as far as acquisition, as of now, a lot of people would agree that acquiring beyond 16 can be a big waste of space, and processing power.

Keep in mind the higher the bit depth you go with acquisition, the more bandwidth you are going to need. 10 bit is not out of the realm of being sub 10k in your package for pricing. 12 bit might be a stretch, but they'd have to put a dual link SDI in the camera, and that's high end professional level hardware, which could easily bump the camera price up several thousand dollars more, then to take advantage of that, the cheapest thing you can get for the sdi recording is a 10 thousand dollar cinedeck extreme.

as for more than 12 bit, they'd have to come out with a even newer recording solution to take advantage of that, and probably bump the processing power, and bandwidth by a lot too. just look at your file sizes between a pro ress, pro ress hq, and pro res 4x4 file. and 4x4 is only 12 bit, it's huge differences.

Not to mention there aren't too many file formats that retain a 16 bit depth, one that I know of is TIFF. You should see the file sizes on a motion sequence at 16 bit, at 1080p...

The more you ask for, the more you have to deal with, that isn't this camera. 10 bit isn't out of reality, next dimension, mega ultra game changing spec to ask for. It's in the reality of what they can offer in this price range, anymore that expect price hikes and a redesign. No other way around it.
 
Back
Top