+++ DVXuser Exclusive! +++ Sony V1 Footage and Stills! DV, HDV, 50i and 25p!
- Thread starter mikkowilson
- Start date
Fugitive said:
Yes, this is obviously correct, this is laws of physics/optics
There are other reasons why smaller isn't better ... the smaller the pixels, the less light each photosite receives, so more noise (although a low quality 1/3in might be noisier than a high quality 1/4in). Also, a smaller sensor needs sharper glass. If you shoot a comparison of 1/3 and 1/4in sensors with the exact same glass, then the 1/4in will be using a smaller amount of the glass (i.e. the middle)-- and any problems with the lens will effectively be magnified very slightly, and you will be trying to resolve the same number of lines through a very slightly reduced area of glass.
But, that's not to say that Sony's CMOS chips won't be super low noise and the Zeiss glass pin sharp! Time will tell. But certainly the odds would be better if we were talking about 1/3in CMOS chips ........... shame.
The DOF issue is unavoidable, and a blow for those striving for shallow DOF for film.
Personally I'm expecting to pick up the XH-A1-- I see it like this:-
Canon- H1 pedigree, famous for great glass, high res 1440 sensor, should be a smooth film-like image.
Sony- 1/4in DOF (bad), low-light MAY also suffer, heavy reliance on pixel shift from relatively low-res chips ... Deal-killer for me will be the almost-certain Sony super video look ... shite for anything other than documentary/corporate/weddings. BUT, true progressive COULD it a real edge for P work.
IF Canon mess something up (unlikely), and Sony delivers staggering high-res footage with awesome low-light performance, good image/cinegamma control, and (critically) a P mode that doesn't look Video-as-hell, then maybe (just maybe) it will offer something special.
I think it's unlikely.... and that the A1 will be better and cheaper.
Last edited:
I am not too worried about the video look, because I strongly believe its a function of lighting and post. I have lit several sets and shot on Sony cams before and they always turn out really film-like when lit properly.
I am just worried about 1/4 inch, but then again, as Heath pointed out, the image could just be so good that any other factors would become a non-issue. You could always use an adapter, for example...
I am just worried about 1/4 inch, but then again, as Heath pointed out, the image could just be so good that any other factors would become a non-issue. You could always use an adapter, for example...
Heath McKnight
Active member
I wonder, again, if 1/4 in. CMOS chips are more similar to 1/4" CCD.
hwm
hwm
Well it gets worse...
And Barry, correct me if I'm wrong but with the sony's 'diamond' shaped CMOS pixels, you're going from a diamond sensor to a square or rectangular television pixel, so the data collected on that individual pixel has to be ?shrunk? to fit within a rectangle?
I really don't even understand it, but I know they sense light with diamonds, but we all know our televisions have rectangular pixels, so each pixel has to get 'cropped' to the right shape.
Wouldn't this make the effective pixel size even smaller? I mean, it collects light like a 1/4" chip but then if the pixels get shaved to shape it would be like shaving the whole chip? Right?
I just don't understand why they'd to that.
And Barry, correct me if I'm wrong but with the sony's 'diamond' shaped CMOS pixels, you're going from a diamond sensor to a square or rectangular television pixel, so the data collected on that individual pixel has to be ?shrunk? to fit within a rectangle?
I really don't even understand it, but I know they sense light with diamonds, but we all know our televisions have rectangular pixels, so each pixel has to get 'cropped' to the right shape.
Wouldn't this make the effective pixel size even smaller? I mean, it collects light like a 1/4" chip but then if the pixels get shaved to shape it would be like shaving the whole chip? Right?
I just don't understand why they'd to that.
Zim
Veteran
epicedium said:Yes, this is obviously correct, this is laws of physics/optics
There are other reasons why smaller isn't better ... the smaller the pixels, the less light each photosite receives, so more noise (although a low quality 1/3in might be noisier than a high quality 1/4in). Also, a smaller sensor needs sharper glass. If you shoot a comparison of 1/3 and 1/4in sensors with the exact same glass, then the 1/4in will be using a smaller amount of the glass (i.e. the middle)-- and any problems with the lens will effectively be magnified very slightly, and you will be trying to resolve the same number of lines through a very slightly reduced area of glass.
But, that's not to say that Sony's CMOS chips won't be super low noise and the Zeiss glass pin sharp! Time will tell. But certainly the odds would be better if we were talking about 1/3in CMOS chips ........... shame.
The DOF issue is unavoidable, and a blow for those striving for shallow DOF for film.
Personally I'm expecting to pick up the XH-A1-- I see it like this:-
Canon- H1 pedigree, famous for great glass, high res 1440 sensor, should be a smooth film-like image.
Sony- 1/4in DOF (bad), low-light MAY also suffer, heavy reliance on pixel shift from relatively low-res chips ... Deal-killer for me will be the almost-certain Sony super video look ... shi*e for anything other than documentary/corporate/weddings. BUT, true progressive COULD it a real edge for P work.
IF Canon mess something up (unlikely), and Sony delivers staggering high-res footage with awesome low-light performance, good image/cinegamma control, and (critically) a P mode that doesn't look Video-as-hell, then maybe (just maybe) it will offer something special.
I think it's unlikely.... and that the A1 will be better and cheaper.
Maybe with the 24p the V1 won't have the "Sony super video look". Maybe it wil have the 24p film look. Maybe it will even be better that the 24F look!
Why would they put 24p on a camera only for it to end up looking like the sony video look?
I'm hope both cameras come out and are great.
Barry_Green
Moderator
But it doesn't work that way. People really, really want to believe that the sensor pixels are directly related to the output pixels, and it just isn't so.meta4 said:
With CMOS it's closer, you can at least individually address the sensor pixels which is something that's impossible in CCD, but understand that even though you can get at the individual pixels, you're getting 960x1080 pixels, which get interpolated up to 1920x1080 by the EIP, then scaled down to 1440x1080 for recording, and then interpolated by the TV back up to 1920x1080 (or, far more likely, scaled down to 1280x720 because the overwhelmingly vast majority of pixel-based televisions out there are 720p-native).
So I wouldn't worry about diamond-shaped pixels not fitting into TV-shaped pixels. It just isn't relevant. All that matters is a) does the footage look good, and b) does it introduce any weirdness we need to be aware of as shooters.
mikkowilson
Steadi-Guru
Regardeing the "shape" of pixels.
Mathematically a pixel is just a single point of light. It has no shape.
The only time a pixel actually has a "shape" is at a display where those points are aranged in a pattern to fill up the correct aspect ratio. But when stored in any form, they are just single points.
An image doesn't have a bunch of values drawing a little rectangle for each pixel, it's just one value for the one point.
- Mikko
Mathematically a pixel is just a single point of light. It has no shape.
The only time a pixel actually has a "shape" is at a display where those points are aranged in a pattern to fill up the correct aspect ratio. But when stored in any form, they are just single points.
An image doesn't have a bunch of values drawing a little rectangle for each pixel, it's just one value for the one point.
- Mikko
Why is deep dof even considered an issue? On a 1/3" sensor it'll be marginally shallower - and you'll only notice this when a number things like zoom and object distance are maximised. Even then you'll lose your field of view. Framing becomes more difficult.
Fit an adapter, and as long as your image is sharp to start with, you're laughing.
Fit an adapter, and as long as your image is sharp to start with, you're laughing.
StMad said:
Well, the difference in size (diagonal) compared to 1/3in is 33% ((6-4)/6), difference in area is 55.5% (((4.8*3.6)-(3.2*2.4))/(4.8*3.6)) .. is it not? (numbers in mm) Far from insignificant ... and actually makes it clear just how much less light 1/4in would receive, especially when you're already dealing with the tiny photosites needed for HD.
I don't know precisely how that change in area will relate to DOF, can anyone save me an hour of reading up on the maths involved?
For those of us who always strive for as shallow DOF as possible, this may be significant.
I completely agree that if the cam is sharp enough then it will take a 35mm adapter, and this is something that I was seriously thinking about (35mm on a V1, fed via HDMI into the new blackmagic card), but the 35mm adapter will be really impractical for any kind of doc / run-and-gun type of work.... I wish that all of my filming was music promo and cini style, but it isn't
EDIT: looking at the numbers (but not actually researching it properly yet), my best guess is that the difference in DOF would be proportional to the diameter ... so a 33% difference, where aperture and focal length are equal?
Last edited:
I saw this camcorder at NAB in NYC, Looks like the old VX1000 in size and shape, a bit heavier. It will compete with the HVX and the new small Canon. Sony has a new way of extracting HDV from 1/4" CMOS chips. Looks like the Canon A1 and will probably be its closest competitor, for now. I would like to see a side-by-side between these two camcorders.
Last edited:
I recently saw some v-1 footage at the miami digital expo. It was an hdv tape playing back on a hdv sony deck. The footage looks like the real deal, 24p, nice color, and no motion artifacts. The blacks looked a little milky. I think sony is going to hit a home run on this one.