AG-DVX200 4K Large-Sensor Handheld Camcorder

greytail said:
Another nice feature of the DVX 200 is that you can assign the CAPTURE function to a user button to take still pictures. The size of still picture captured depends on the record mode setting. The great thing about this feature is that you can take stills images even while the camera is recording video. The maximum still picture size you can capture is 8.8MP (4096×2160) in a 17:9 aspect ratio. The only limit to the number of stills images you can capture is the size of your recording media.

I am very excited about this...How big are the prints that you can make from this?.
Click to expand...
Well... that's a photography question, based on how dense of a DPI you think looks good printed. At 300 DPI, you could make about a 13.5" x 7.2" print; at 100 DPI you could make a 41" x 21.6" print. So it depends on your opinion of how tight the DPI needs to be, and at what distance the print will be viewed. If it's for something hanging on a wall above people, you might be fine with 100 DPI or even less. If it's something that you're going to be handing to pixel-peepers, you might want 300 DPI or even more. Most digital cameras set the default at 72 DPI; at 72 DPI you'd get a 57" x 30" print. But 72 DPI is generally used for computer monitors, not for printing.
 
Ron Evans said:
Wide angle has been mentioned before a few times. Not sure there is a low light hit compared to any other shooting rate though.
Click to expand...
I think the thought that 4k 60p gives a low light hit comes from the thought that a smaller sensor (or cropped area of sensor) gives lower basic sensitivity than a bigger sensor. In general that's certainly true.

At first I was sceptical, on the basis that PHOTOSITE SIZE doesn't change with crop - so why should sensitivity? Didn't seem to make sense.

But - think of it this way. If a given resolution was achieved by downsampling, such that (say) four photosites got averaged to a single output pixel for full sensor readout, then if only half the area got used, each output pixel would (on balance) then be formed by an average of only two phoosites. (The true maths and process is far more complex, but the principle holds good.)

This "averaging" leads to lower s/n ratios than if single photosites were used. Effectively four small photosites averaged give the same result as a single large one of the equivalent area. Reduce the no by half, and the equivalence is the same as halving the area of the single large one. So yes, it seems likely that fundamental sensitivity of the cropped chip for 4K will be lower than fundamental sensitivity if the whole chip area is used. (All else equal.)

But really, I have to agree with Barrys point - "it's pretty much the only camera out there offering UHD/60p at all. Most other cameras max out at UHD/30p." (At least in this price range.)

It's 4K 60p mode may be compromised compared to 4K at lower frame rates - but at least it does do 4K at 60p at all! Maybe it depends if you see glasses half full or half empty.......? :)
 
harddrive said:
.... It's 4K 60p mode may be compromised compared to 4K at lower frame rates - but at least it does do 4K at 60p at all! Maybe it depends if you see glasses half full or half empty.......? :)
Click to expand...
If DVX 200 crops to 8.8 MPX or so from the oversampled 12.9 MPX, the difference will largely be in the resolution, but marginally at that. Sony claimed that its digital zoom with FS5 is "visually indistinguishable" at a 1.5 ratio in 4K and 2.0 in 1080P. One could call it internal upscaling a la ARRI Alexa's Faux 4K and, depending on the crop, the visual difference might indeed be very minor - noticeable on a 200% resolution chart but not with own eyes.
 
DLD said:
If DVX 200 crops to 8.8 MPX or so from the oversampled 12.9 MPX, the difference will largely be in the resolution, but marginally at that.
Click to expand...
If the DVX200 crop is to 8.8Mp in the 4K 60p mode, that's only what many cameras are starting with in the first place!! And that includes higher end cameras like the F55 and Varicam for just two examples. So I really don't see any compromising for resolution.

The biggest negative - as others have already said - will be a far more restricted wide angle end to the zoom range. About equivalent to 37mm on a ff camera. But as said before, it will do 4K 60p, even with the restrictions - some cameras won't do that at all.
DLD said:
Sony claimed that its digital zoom with FS5 is "visually indistinguishable" at a 1.5 ratio in 4K and 2.0 in 1080P. One could call it internal upscaling a la ARRI Alexa's Faux 4K ........
Click to expand...
I think that feature is very impressive in HD, and can well believe the 2x claim there. But for 4K, a 4K Bayer chip is on the limit of what may be called 4K after the deBayer in the first place. Start doing a crop, and whilst it may still look good, it ain't 4K - and same with the Alexa, may look good.... but it ain't 4K.

It may also become relevant not just on a chart, but if the 4K mode is used to give a reframing ability for HD. You'll still be able to crop and reframe - just not as much as if the digital zoom wasn't used.
 
harddrive said:
I think the thought that 4k 60p gives a low light hit comes from the thought that a smaller sensor (or cropped area of sensor) gives lower basic sensitivity than a bigger sensor. In general that's certainly true.

At first I was sceptical, on the basis that PHOTOSITE SIZE doesn't change with crop - so why should sensitivity? Didn't seem to make sense.

But - think of it this way. If a given resolution was achieved by downsampling, such that (say) four photosites got averaged to a single output pixel for full sensor readout, then if only half the area got used, each output pixel would (on balance) then be formed by an average of only two phoosites. (The true maths and process is far more complex, but the principle holds good.)

This "averaging" leads to lower s/n ratios than if single photosites were used. Effectively four small photosites averaged give the same result as a single large one of the equivalent area. Reduce the no by half, and the equivalence is the same as halving the area of the single large one. So yes, it seems likely that fundamental sensitivity of the cropped chip for 4K will be lower than fundamental sensitivity if the whole chip area is used. (All else equal.)

But really, I have to agree with Barrys point - "it's pretty much the only camera out there offering UHD/60p at all. Most other cameras max out at UHD/30p." (At least in this price range.)

It's 4K 60p mode may be compromised compared to 4K at lower frame rates - but at least it does do 4K at 60p at all! Maybe it depends if you see glasses half full or half empty.......? :)
Click to expand...

You might have a point if it was cropping into 2x or larger which basically a DSLR did to get HD from such a high resolution sensor but in this case we are not talking enough extra photosites to really impact low light all that much. If they did use the full sensor not only would you have aliasing and moire up the ying yang for such a small subtle conversion but the sensitivity and noise smoothing impact would be little to none. We are talking a 12% change in size here and not a 100% - 150% change in size like we have seen in DSLR cameras in the past. The over sampling advantage here is pretty much nonexistent. The tradeoff is you get a true 1:1 pixel readout which means a cleaner moire free image compared to what most DSLR cameras can achieve by binning or line skipping. The only way I would ever consider it acceptable to shoot 4k with the full sensor is if the sensor only had 8.8 MP on the entire sensor. This would mean each photosite would be a tiny bit larger and able to capture a tiny bit more light. It wouldn't be a very big advantage but could maybe help 1/3 of a stop.
 
I think we will have to wait and see how the DVX200 at UHD 60P really performs against a HC-X1000 or FDR-AX1/PXW-Z100 at wide angle. It cannot go as wide and starts with a slower lens but a bigger sensor. As a HD camera it will likely outperform my NX5U most of the time and be backup in 4K for my FDR-AX1 so is of real interest but I have a feeling that at UHD 60P ( the only thing I am interested in with the camera) its performance may not be that startling better in low light compared to my FDR-AX1 which is not that good !!! Would love to be wrong. In most of the stuff I shoot as a full stage fixed camera 37 mm may be fine and then it will definitely have an advantage of F2.8 against a slightly zoomed position of the others which would move from their F1.6 to maybe F2 with lens ramp. So either it will not be wide enough or it will catch up a little on the others because they have to zoom in . Will wait and see.
 
Thomas Smet said:
If they did use the full sensor not only would you have aliasing and moire up the ying yang for such a small subtle conversion but the sensitivity and noise smoothing impact would be little to none.
Click to expand...
Point is they DO use the full sensor at other framerates! That's straight from the Panasonic specs:
Effective Pixcels: 15.49 megapixcel (FHD), 8.71 megapixcel (UHD/59.94p),
12.89 megapixcel (UHD/29.97p), 13.35 megapixcel (4K/24p)
Click to expand...
I'm assuming by UHD they really mean QFHD (3840x2160), so even the QFHD/29.97p mode is a crop (12.89Mp) from the full sensor as used in HD (15.49Mp).

As for aliasing/moire look back at the chart Barry posted a while back. (There is very little.) Conclusion is that the processing on this camera is far more advanced than on such as the AF100 - the DVX200 is able to fully read, deBayer and downscale the whole chip at the lower video framerates. So forget all the lessons from performance of earlier generation DSLRs.

Thomas Smet said:
We are talking a 12% change in size here and not a 100% - 150% change in size like we have seen in DSLR cameras in the past.
Click to expand...
No - in QFHD we are talking about the difference between 8.71 and 12.89 Mp - 4.18Mp, or about 48%.

Compared to the full frame as used in FHD mode the difference is 8.71 and 15.49 - nearly 80%!
Thomas Smet said:
The over sampling advantage here is pretty much nonexistent.
Click to expand...
I assume you're thinking in terms of resolution etc - but the big difference is really in limits of angle of view. Fully zoomed out, the DVX200 lens is equivalent to 28mm in FHD mode - but only a little over 37mm in QFHD/60p. That's a big difference, and if wide angle is important, then on a fixed lens camera there's nothing that can be done about it. Of course, for other types of work - when the telephoto end of the lens is most important - this crop could be seen to be a benefit......
 
According to the DVX User Group on Facebook, the delivery of the camera for sale has been delayed until November. Anyone know if that is accurate?
 
Auto Focus Face Detection?

Auto Focus Face Detection?

I'm close to pulling the trigger on pre-ordering the AG-DVX200, but I haven't seen anything definitive on whether the camcorder will auto-focus based on face detection. Has anyone seen evidence that the AG-DVX200 can focus based on facial recognition?
 
MikeWilmette said:
I'm close to pulling the trigger on pre-ordering the AG-DVX200, but I haven't seen anything definitive on whether the camcorder will auto-focus based on face detection. Has anyone seen evidence that the AG-DVX200 can focus based on facial recognition?
Click to expand...
There is no face recognition feature, like there has been on some of the other cameras. You can point at the screen to tell it where you want it to auto-focus, but there is no automatic face recognition feature.
 
dvbrother said:
I just downloaded the DVX200 version of the V-Log LUT. I applied it to some sample V-Log clips. To my eye and on the scopes, it is identical to the Varicam 35 version of the LUT.
Click to expand...
Looks like it to me as well. Looks like they renamed it so DVX200 users would know to use it maybe.
 
lol just got my invite by way of email to the EVS event Barry will be at...

but a ? for you Barry.. have you heard if someone will make a wide angle adapter like on my old DVX... but for 4k... probably cost and arm and a leg some maybe they won't ...
 
The images do look nice. Am I reading correctly that it's 8 bit 4:2:0 and 150mb/s in UHD? If so surely that's a bit poor isn't it?....though, again, images do look nice - on the web at least.
And it does 120fps, but at what resolution / compression?
Thanks for the links.
Steve
 
You must log in or register to reply here.
Back
Top