6mp CCD v 10mp CCD high ISO

[bfitzgerald]

Shooting the Km5d and A200 at ISO 1600
Metered (honeycomb/multi segment) reading for the KM5d was 1.3 sec at f5.6 ISO 1600
A200 metered and exposure of a stop less.

Both cameras set to Manual mode with an exposure of 1.3 seconds at f5/6 ISO 1600 raw

Imported into LR3 linear tone curve (my normal setting) NR sliders both at 0 no sharpening.
Kelvin set to 2345 for both to match the output of the energy saving light output

Capture 5d A200.jpg

(217.11 KiB) Downloaded 2900 times

Note the stronger colour noise on the right hand (A200 image)
This is what the metered setting from the A200 gives you in raw with no NR applied at all (LR3)

Capture2.JPG

(139.7 KiB) Downloaded 2900 times

This is fairly typical of the A200 in low light with a metered reading which is inappropriate for the exposure (not always but often) the image is pretty much useless so manual intervention is required at higher ISO speeds.

Downsizing the A200 image trying to remove colour casts from both modest sharpening of USM radius 1 and 55

Capture6.JPG

(200.82 KiB) Downloaded 2900 times

A200 left 5d right hand side

Conclusions are as follows:
The A200 high ISO performance is not bad "if" you ignore the in camera metering
The 10mp CCD falls apart with colour noise if underexposed in low light more than the 6mp CCD
The 10mp CCD has more colour noise at high ISO luminance is also a bit more but not the main problem
I'd rate the 6mp CCD as better for high ISO though if you give the 10mp CCD enough exposure the results are decent and not too far off
Most of the noise complaints could have been addressed by making sure low light exposures are not underexposed (A200 does this too often and where you get away with it on the 5d you don't on the A200)
Limiting Auto ISO to 400 for the A200 didn't solve the high ISO issues

[bossel]

Time machine - who cares about KM5D vs A200 noise these days? The KM5D users have most probably moved on and the A200 users as well or they don't care Don't even know why I am writing this ...

[bfitzgerald]

I always find it helps to have a point when you make a post, might want to bear that in mind before you press the button next time

[bakubo]

Interesting. Personally, I don't mind stuff about older products. Barry is using these two cameras so they are relevant to him.

[redsim74]

My 5D still gets the occasional run, despite having the A700. I still marvel at the colours and its a nice camera to use. It feels a bit ponderous now in terms of focus speed and shutter lag but I don't shoot much in the way of action.

I had an A100 for a while but it didn't really feel like much of an upgrade over the older KM, so the A100 went in favour of the A700.

[bossel]

I always find it helps to have a point when you make a post

Bravo! For years your only point is that Sony sucks. Then you get a Pentax and start making NR tests with the A200. Your only point is that Sony is evil and that's of no value at all. Post something useful once in a while!

[bfitzgerald]

Bossel I'm not exactly sure what the problem is would you care to explain a bit?
I thought this was at least mildly interesting (for some) of course those who find it boring can simply ignore it and post in threads they find meets their excitement criteria. Or is there another agenda here?

I didn't actually make any reference to Pentax in this thread, nor am I making a complaint (I think the metering aspect is relevant) Seems you are complaining about me not moaning that's a new one for me I shall take that on board.

This is useful to post it's my white fluffy cat and a topic of interest to cat lovers..feel free to discuss if desired

DSCF2623.jpg

(102.32 KiB) Downloaded 2846 times

Not shot on a Minolta, Sony or a Pentax. I was trying out the EXR mode on a Fuji compact.

[pakodominguez]

Conclusions are as follows:
The A200 high ISO performance is not bad "if" you ignore the in camera metering
The 10mp CCD falls apart with colour noise if underexposed in low light more than the 6mp CCD
The 10mp CCD has more colour noise at high ISO luminance is also a bit more but not the main problem
I'd rate the 6mp CCD as better for high ISO though if you give the 10mp CCD enough exposure the results are decent and not too far off
Most of the noise complaints could have been addressed by making sure low light exposures are not underexposed (A200 does this too often and where you get away with it on the 5d you don't on the A200)
Limiting Auto ISO to 400 for the A200 didn't solve the high ISO issues

The A200 supposedly have an A100's (my reference cuz I had to "endure" this camera for 1 year) improved sensor. I recall the A100 been quite sharp, mostly because too contrasty -that means you'll also see better the noise... But DK loved this camera, the A100, and I'm sure he will give you the best LR/PS settings in order to avoid all that noise.

In other hand, I think that the standard settings you are applying favor the 5D over the A200. Because you are in the mood, why don't you try to get the best result of each file and then compare them? that includes exposing both files to the right...

Regards

[bfitzgerald]

Pako I did the A200 a favour by using the 5d metered reading trust me as above the one stop less exposure is much much worse for noise. I agree mostly with the 5d metered reading sure I could add 1/3 of a stop but it's def exposing more appropriately than the A200 is.

Had I shot both as per metered reading in each camera I'd get howls of complaints about not matching the exposure or not being fair. I think it's a decent stab at some kind of level playing field. The reason I use the 5d more for low light is it's mostly consistent (not always but fairly good) you can't really wander around with the A200 fixed at +1 exp comp as it can be too much for some scenes. I've seen as much as 1.5 stops adrift on exposure or under on the A200 though that is less common. In daylight it's great even in tough scenes the metering is better than the 5d good balance between shadows and not blowing them out. When the clouds arrive and overcast it goes under almost always, 5d can as well but never as much.

Had Sony tweaked the metering they'd have got a more positive response regarding it's low light performance.

[chriznatch]

My a100 consistently underexposed in low light -- sometimes by as much as a full stop. And by low light, I mean not only indoors, but outdoors on a cloudy day. I always wondered why Sony hadn't done a better job with the metering. My wife often got better exposures using a small Canon point and shoot. I would have thought the a100 could do at least as well as a point and shoot in auto mode.

On the flip side, I did get a lot of beautiful pictures on sunny days, and once I figured out that I should be using exposure compensation or manual exposure, on cloudy days as well. I now have an a700, and although my pictures are generally better with the new camera, there are still a small handful of shots I got with the a100 that are better than anything I've been able to produce with the a700 so far.

[rush2112]

When I do test image comparisons now, I pay more attention to the histogram reading than a metered reading or relying on ISO sentitivies. There seems to be too many variations in the actual/stated ISO values anymore. Best to try and read the histogram and get them as close as possible when making an exposure.

I like Barry's comparison here, shows pretty clearly the difference extra pixels make on a sensor, given even the advances that came with the A200's improved sensor over the A100.

C