Believe what you want.
35 mm movie film cameras vs HD cameras
- Thread starter suryad
- Start date
No, not necessarily. I'm saying that shooting at 2MP on a sensor that has 6MP is effectively oversampling (supersampling) the image. It will have more information than a native 2MP camera, and - provided it doesn't use some braindead profile to aggregate those full res RAW sensor data into the lower res sRGB image - that it will be effectively the same as shooting 6MP and downscaling afterward. Difference only being when you combine the samples.
Same with a newer film scanner (they are basically digicams) that has a higher than 2K max resolution. A higher sample resolution increases perceivable detail (up to a point) at a fixed output resolution.
Same with a newer film scanner (they are basically digicams) that has a higher than 2K max resolution. A higher sample resolution increases perceivable detail (up to a point) at a fixed output resolution.
Last edited by a moderator:
Well, yes and no. My point is that increasing resolution is useless if you don't pair it with same quality optics. In fact, you'll instead just get more noisy images. This is plaguing today's consumer products.
Cheap film, but what lens? I don't think that crop looks any more impressive than what I'd get on my A100 DSLR. It's quite soft and no very close to 40MP actual resolution. But that's besides the point. You may be able to find an in focus point where you might see very fine detail. However, at normal apertures the vast majority of the picture will be enough out of focus that increasing the pixel count only creates more blurry pixels. No level of quality of the optics will change that. You can use very tight apertures instead, but then you'll eventually get quality loss from diffraction, and even in focus areas will be blurry. Plus a dose of motion blur from the longer exposures. IMHO we've passed the point already where increasing resolution stopped being useful.
Not a very convincing example.
In your 0.4MP version of the image, you can still see that there is text there, in the 40MP version you can read the text, but only barely.
My cheap 10MP digital camera does much better, if I scale a 10MP picture to both 0.4MP and 40MP, I can read text in the 40MP version that is practically invisible in the 0.4MP version.
Unfortunately I can't show you examples at the moment, I do not know where neither my camera nor my charger for it is at the moment, but if you don't believe me I suppose I could start looking for them...
If I recall correctly from a conversation in the office a few years back with someone who'd looked into the issue, there are two problems caused by the pixel sensors getting smaller.
The first is that fewer and fewer photons can hit the sensor so that the recorded values are inherently becoming noisier and noisier. Related to this is that each pixel 'bucket' can only hold so much 'charge' and so you lose dynamic range. As soon as some % of pixels become fully exposed, the picture has to be finished.
The manufacturers, of course, don't want to spend silicon on a bigger sensor but appear to be locked in an almost senseless resolution arms race. <shrug>
Of course, take all of the above with a grain of salt.
The first is that fewer and fewer photons can hit the sensor so that the recorded values are inherently becoming noisier and noisier. Related to this is that each pixel 'bucket' can only hold so much 'charge' and so you lose dynamic range. As soon as some % of pixels become fully exposed, the picture has to be finished.
The manufacturers, of course, don't want to spend silicon on a bigger sensor but appear to be locked in an almost senseless resolution arms race. <shrug>
Of course, take all of the above with a grain of salt.
I think you are quite right. Let me see if I can find an explanation from dpreview.com that explained pretty much what you said.
Are you sure? It's cropped from 40MP scan and it's 1:1.
Take a shot by any consumer DSLR, upscale it to 40MP, find any text on the upscaled shot, which is 6 pixels high (as on this image) and try to read it. You will fail.
If your DSLR supports resolution up to 10MP, the font on the original shot will be only 1,5 pixel high, so there wouldn't be enough details to make it readable even after resize.
Even 20MP won't be enough and I doubt even 30MP DSLR would make a better job, because majority of DLSR are not able to retain per pixel detail. Exceptions are Hasselblad H3DII-39MS ($43k) and it's piezo-eletric shift of bayers mask and Sigmas Foveon based cameras (which are much cheaper, but their resolution is much lower).
FWIW: A 15MP consumer DSLR (EOS 50D) can defeat the test charts and resolve >2000 line pairs.
Test chart.
Detail shot.
Test chart.
Detail shot.
No salt needed; that's absolutely true. And it's also a major reason why the picture quality of an SLR beats compact cameras. They simply got much larger sensors, and even if you might get a slightly higher resolution as well you still get an order of magnitude more light per pixel. That's why dpreview.com has added pixel density info to their camera database so you can easily compare. For instance some Nikon models:
D700: 1.4 MP/cm² (Fullframe SLR)
D90: 3.3 MP/cm² (APS-C SLR)
P6000: 33 MP/cm² (Highend compact)
The bottom row here is common compact camera sensor sizes:
http://en.wikipedia.org/wiki/File:SensorSizes.svg
From what I've been able to google up average cellphone camera sensors are another factor 2-4x smaller than compacts.
If you don't mind non-english forum, here is a bit more complete list of camera sensors, even with some examples:
http://4um.overclocking.cz/showthread.php?p=772299#post772299
the last one (1/4") is used in many web-cams
http://4um.overclocking.cz/showthread.php?p=772299#post772299
the last one (1/4") is used in many web-cams
Holy crap that cam is awesome! If I only had 1500$ laying around.
Sensor size is absolutely an important issue. I hate those consumer cameras pushing for the so-called high pixel counts. With those small pixels, these are just noises. They are just a waste of time to store and process these useless pixels. Not to mention that those noises have serious impact on JPEG compression.
Color resolution is another issue comes from the Bayer pattern. In most cameras (consumer or professional) the color resolution is essentially only half of the "true" resolution because of the Bayer pattern. Only the Foveon X-3 sensor has true color resolution, but it's only used by Sigma DSLR. Though some obscure industrial camera also uses it.
Digital film cameras such as Red One uses 35mm size CMOS sensor, that means it's pixel size is comparable to current high end DSLR (i.e. much better than those cheap high pixel count consumer cameras). However, they have their own problems. Many CMOS sensors use rolling shutter and have this "shearing" problem, that's because the data is read line-by-line, so when the camera is panning very fast, the recorded image will be "shearing." The most comical example is the Nikon D90's video recording mode.
Some very high end digital flim cameras use CCD and global shutter so they don't have this "shearing" problem. They also uses 3 CCD to record RGB on different sensors, so they have full color resolution. But these cameras are very expensive. However, if you are making a feature film, the savings on the film will easily outpace the price of the cameras.
Color resolution is another issue comes from the Bayer pattern. In most cameras (consumer or professional) the color resolution is essentially only half of the "true" resolution because of the Bayer pattern. Only the Foveon X-3 sensor has true color resolution, but it's only used by Sigma DSLR. Though some obscure industrial camera also uses it.
Digital film cameras such as Red One uses 35mm size CMOS sensor, that means it's pixel size is comparable to current high end DSLR (i.e. much better than those cheap high pixel count consumer cameras). However, they have their own problems. Many CMOS sensors use rolling shutter and have this "shearing" problem, that's because the data is read line-by-line, so when the camera is panning very fast, the recorded image will be "shearing." The most comical example is the Nikon D90's video recording mode.
Some very high end digital flim cameras use CCD and global shutter so they don't have this "shearing" problem. They also uses 3 CCD to record RGB on different sensors, so they have full color resolution. But these cameras are very expensive. However, if you are making a feature film, the savings on the film will easily outpace the price of the cameras.
It's also used in their DP1 and DP2 compacts. I'd buy one if they had a zoom lens.
pcchen: bayer-mask sensors have half color resolution only in theory. Low-pass optical filter and in-camera processing has huge impact and causes additional loss of color detail.
The 4.7MP Foveon has more details in red chanel even if you upscale the picture to 21MP and compare to 1Ds Mk3 (I'll not argue, that 1Ds has much lower noise, much better in-camera processing, much better speed atc.), but I feel, that more and more products are developed to excell in conventional theoretical tests. There is not any single website, which measures resolution of color channels. So manufacturers don't have any single reason to solve this issue...
The 4.7MP Foveon has more details in red chanel even if you upscale the picture to 21MP and compare to 1Ds Mk3 (I'll not argue, that 1Ds has much lower noise, much better in-camera processing, much better speed atc.), but I feel, that more and more products are developed to excell in conventional theoretical tests. There is not any single website, which measures resolution of color channels. So manufacturers don't have any single reason to solve this issue...
Wow that Sigma DP1 craps all over everything for color resolution. Read a review of it and they attributed it to superior camera software IIRCThis is true only for luminance resolution. Color resolution is much lower.
just compare detail in color (esp. red) on these crops:
Canon 40D
Canon 50D
Nikon D300
Sigma DP1 (upscaled to Nikon size)
