Reading mode: Light Dark

SD800IS - Why oh why the smaller sensor???

Go to last post
Matthew Miller said:
ljfinger said:
Do you realize the 1/1.8" sensor is only 1.25x (linear dimension)
bigger than the 1/2.5" sensor (same as the difference between the
5D and the 1DII)?
Click to expand...
However, in two dimensions, it's 1.55x (going by the size values in
the dpr glossary). Since the sensor is a 2d surface, not a line :)
that's the number that matters, isn't it? The bigger sensor
receives 55% more light.
Click to expand...
It's still 2/3 of a stop (log base 2 (1.25^2)).

--
Lee Jay
(see profile for equipment)
 
Lee Jay said:
Do you realize the 1/1.8" sensor is only 1.25x (linear dimension)
bigger than the 1/2.5" sensor (same as the difference between the
5D and the 1DII)?
Click to expand...
i realize it and even better would be a 2/3" sensor or bigger. as a 5D/20D shooter you should know how much sensor size has to do with dynamic range and color accuracy.

compare the ixus750 1/1.8" 7MP to the new ixus800/850 1/2.5" 7MP and in my opinion there is a huge differnce in image quality - color, sharpness, noice etc...

regards
thomas
 
Why do you think higher mp sensors are worse? On the same process more pixels will result in less chip coverage of the light sensitive area but i think we are probably 2 generations further now wich means the circuitry is 1/4th the size so even if you double the mp you still have higher efficiency.
 
The only real advantage of more pixels is to crop the picture - this would make the 900's lens seem longer (closer to the 700IS) but as I noted there is no IS so you would have to be concened about camera shake. I've never been that impressed with downsampling saving a noisy image. Downsampling is much more valuable in the context of sharpening - and these cameras usually provide fully sharpened pictures at the max resolutions.
 
There is very little real benefit of the larger sensor in the 5D over the one in the 20D, compared to the 20D it means that instead of iso 100-1600 it has iso 40-640 equivalent, so all you are missing with the 20D is the lower iso range and since iso 100 on the 20D is already so good i dont think thats worth a lot?

The main practical advantage to get the 5D is that nearly all lenses are made for 36-24mm and except for the telephoto lenses they are a complete waste on crop sensors while the few crop specific lenses are overpriced.

In the future with new technology even the low iso advantage of larger sensors will disapear so i wonder if there will be any reason for them.

Maybe it is possible to make higher quality lenses for larger sensors but i dont have that knowledge.

There is still a real advantage of larger sensors in compact cameras though, i think up to the 1inch type there will be a good improvement, at iso100 it would have the same quality as iso800 on the 5D wich might be good anough? Ofcource with the same size lens it would mean you get f/5.6-11 or worse so the use of it will be limited.
 
Isn't it weird what they have done to these cameras? It's as if they want you to have to buy two or more cameras to get the features you want...

:-o
--
Sue * Pro One

 
Canon is always into market segmentation. If you want all the goodies, you pay through the nose.

I could live with the small sensor, but with umpteen useless scene modes, you think they could quietly slip in manual control to make it actually usefull?
 
Suepete said:
Isn't it weird what they have done to these cameras? It's as if
they want you to have to buy two or more cameras to get the
features you want...
Click to expand...
Canon does this more because they are on top and get away with it. Buy two cameras, more money for Canon. You don't even get manual Shutter/Aperture control, Heck you don't even get readouts of the values. You are shooting blind with respect to what your camera is doing.

I own an S400 and a G6. The G6 is often too big and the S400 frustratingly lacks any controls.
 
malcolm82 said:
There is very little real benefit of the larger sensor in the 5D
over the one in the 20D,
Click to expand...
Wow...that is just totally wrong. The 5D produces about a 1/2 stop improvement in noise at the pixel level and substantially more detail. If you use software to trade that extra detail for better noise performance, you can get about 1 1/3 stops better performance. That's the difference between f4.5 and f2.8.
malcolm82 said:
In the future with new technology even the low iso advantage of
larger sensors will disapear so i wonder if there will be any
reason for them.
Click to expand...
The above-mentioned advantage will remain.
malcolm82 said:
Maybe it is possible to make higher quality lenses for larger
sensors but i dont have that knowledge.
Click to expand...
A lens for a 1.6 crop sensor has to be 1.6x sharper than a lens for a full-frame sensor for the same final image sharpness.

--
Lee Jay
(see profile for equipment)
 
Thomas K. said:
Thomas K. said:
Do you realize the 1/1.8" sensor is only 1.25x (linear dimension)
bigger than the 1/2.5" sensor (same as the difference between the
5D and the 1DII)?
Click to expand...
i realize it and even better would be a 2/3" sensor or bigger. as a
5D/20D shooter you should know how much sensor size has to do with
dynamic range and color accuracy.
Click to expand...
The problem is, you need a huge difference in sensor-size to get a significant difference in dynamic range. The 5D's sensor is 39 times bigger than a 1/2.5" and it has about 2 stops more dynamic range. The difference between a 1/2.5" and a 1/1.8" isn't very much in this area. You do still end up with either 25% more detail or 2/3 of a stop better noise performance.

--
Lee Jay
(see profile for equipment)
 
Thats exactly what i meant with the advantage of the 5D having lower iso equivalent levels of the 20D.

An f/2.8 lens on a 20D is the same as an f/4.5 lens on the 5D, meaning they send the same amount of light to the sensor, thats what i meant with iso 100-1600 on the 5D being equivalent to iso 40-640 on the 20D. So if they give you these lower iso modes on the 20D the advantage of the 5D is gone, is it that important to have better than iso 100 on the 20D?

I am not sure about what you said about lens quality, i think the quality is mainly dependent on things like chromatic abberations, apperture, astigmatism and others wich i think also scale along with the sensor size meaning percentages in the width of the image. Can you give an example of one that gets worse with smaller sensor size? For example i know the wavelength limitation would be one if visible light had 10micron size but thats not the case.
 
Isnt dynamic range determined by noise? Before you said a compact camera has no advantage for depth of field implying that a 1/2.5 sensor at iso 100 has the same level of noise as a full frame sensor at iso 3200.

I thought dynamic range was determined by the level of certainty of the values of the pixels, if you have 32 times higher noise doesnt that mean the pixel values have 32 times larger random variations? It seems to me that would mean that brightness levels at that size(per pixel or whatever) can only be accurately determined at a 32 times higher brightness meaning a 5 stop smaller dynamic range. It depends if the software handless it correctly though. I dont think we should directly see what the sensor captures but an interpretation of what the data means, at the per pixel level at a high iso it means you would only see very high contrast details while any low contrast details couldnt be shown since they would be indistinguishable from noise but over a larger number of pixels the brightness levels would be determined more accurately allowing high dynamic range even with extreme noise levels. Less noise would only allow smaller or lower contrast details to be shown.
 
ljfinger said:
The problem is, you need a huge difference in sensor-size to get a
significant difference in dynamic range. The 5D's sensor is 39
times bigger than a 1/2.5" and it has about 2 stops more dynamic
range.
Click to expand...
Could you explain how you got to that 2 stops number?
 
malcolm82 said:
Isnt dynamic range determined by noise?
Click to expand...
Partially, yes. Two things can happen to kill dynamic range. You can blow highlights or you can cover shadow detail with noise. If you always expose perfectly so that the highlights are just barely preserved and if you apply a tone curve the keeps scene midtones in the right midtone range, then DR is determined by noise. However, these compacts don't do that. They choose to blow a few highlights and keep the midtones in the right places in order to preserve more shadow detail. That's why this image looks like it does. The midtones are in about the same places (other than a color difference), the shadows are similar but the highlights are different. This is caused by smaller well capacity rather than increased noise:

http://photos.imageevent.com/sipphoto/samplepictures/20D%20versus%20S3%20highlight%20preservation.jpg

--
Lee Jay
(see profile for equipment)
 
malcolm82 said:
ljfinger said:
The problem is, you need a huge difference in sensor-size to get a
significant difference in dynamic range. The 5D's sensor is 39
times bigger than a 1/2.5" and it has about 2 stops more dynamic
range.
Click to expand...
Could you explain how you got to that 2 stops number?
Click to expand...
I roughly measured it between my 5D and S3.

--
Lee Jay
(see profile for equipment)
 
Well i dont think it matters how much you expose the shot as long as there is a large anough dynamic range coming trough the lens. It will always be the ratio between the maximum level and the minimum level that can be determined.

A simple way to see that a full frame sensor could have 1/32th the amount of noise as a 1/2.5 is the fact that you could simply have the same pixels as the S3 spread over the larger area (200+mp). If you simply combine the pixels in 32pixel groups you would have 1/32th the randomness per pixel right?

The level of brightness the sensor would show even when not receiving any light would still be the same though so simply using this level as a minimum to determine the dynamic range would not increase it at all. I think your 2stop number might come from this? This floor might be at brightness 64 while the randomness in it might only vary between -1 and +1. I think the dynamic range should depend on how accurately you can determine the amount of light the sensor received wich means predicting this floor. I think the 5D allows much higher dynamic range per pixel than is thought.
 
malcolm82 said:
Well i dont think it matters how much you expose the shot as long
as there is a large anough dynamic range coming trough the lens. It
will always be the ratio between the maximum level and the minimum
level that can be determined.
Click to expand...
The "that can be determined" is the tough part. How much detail can you see below the noise floor? Depends on the type of detail - the color, the contrast etc. - and the type of noise. Since this is a human perceptual thing, you will always be stuck with imprecise definitions of dynamic range.
malcolm82 said:
A simple way to see that a full frame sensor could have 1/32th the
amount of noise as a 1/2.5 is the fact that you could simply have
the same pixels as the S3 spread over the larger area (200+mp). If
you simply combine the pixels in 32pixel groups you would have
1/32th the randomness per pixel right?
Click to expand...
No, 1/sqrt(32).

--
Lee Jay
(see profile for equipment)
 
Thanks for that comparison link. I had no idea they were that bad! I think you can do it right manually though?
 
malcolm82 said:
Thanks for that comparison link. I had no idea they were that bad!
I think you can do it right manually though?
Click to expand...
Do what right?

--
Lee Jay
(see profile for equipment)
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top