Go Back   Steve's Digicams Forums > Digicam Help > General Discussion

Reply
 
Thread Tools Search this Thread
Old Mar 3, 2008, 8:22 AM   #1
Member
 
Join Date: Jan 2006
Posts: 96
Default

Hi
I have just red this article. Part about "noise as a tradein for high ISO" is a bit suspect for me. I have Nikon D40x and is easy to make noise free picture with high ISO even 1600. My preivous camera was Nikon coolpix 8400 and ISO 200 was max for noise free pic. I have tried make same test as in the article. I made several pics of flowers with ISO 100 to 1600 in "jpg fine". I can see different what is in the article but only when apply 100% crop. Even with 100% crop, loosing of detail is less then is in the article. Finding noise in 100% crop is pointless. Where is the catch? I don't believe my camera is better. Maybe somethin else. What is crop factor? Was anybody else try same test? Please comment.
Thanks
emes is offline   Reply With Quote
Sponsored Links
Old Mar 3, 2008, 11:36 AM   #2
Moderator
 
Nagasaki's Avatar
 
Join Date: Jan 2004
Posts: 897
Default

You've compared two very different cameras.

The 8400 is an 8Mp camera with a 2/3 sensor which is 8.8 x 6.6 mm = 58.08 mm squared. each pixel is 0.00000726

The D40x is 10Mp has an APS-C sensor which is 25.1 x 16.7 mm 419.17 mm squared. each pixel is 0.000041917 mm.

So although there are more pixels on the D40x the individual pixels are an order of magnitude greater in size.

The fact that the larger pixels of the D40x give better high ISO performance than the smaller pixels of the 8400 bears out what the tech corner article says. It's also the reason why we pay out big bucks to buy DSLRs with better image quality.

When viewed on screen the differences will seem less but your screen resolution is probably 1024 x 768 or similar so less than 1Mp. If that's all you want to view at then you don't need 8 or 10 Mp to start with. However if you want to be able to crop areas of the image and print at reasonably large sizes then the image quality at 100% becomes more important.

Ken
Nagasaki is offline   Reply With Quote
Old Mar 3, 2008, 1:09 PM   #3
Member
 
Join Date: Jan 2006
Posts: 96
Default

Hi
I know they are different sensors. But sensors makers are not increase number of pixels only. They have improved noise level as well. You cannot say that one pixel 10 years ago has same noise as one pixel, same size pixel, today. Nikon D1 has max ISO 1600, same as d40X, same sensor size, but pixel size is cca 3x bigger (2.7 Mpix). Does d40x have 3x more noise? When you crop 10 Mpix pic, from d40x, to 2.7 Mpix, noise will not be visible.
Anyway I don't expect ISO 1600 will produce same as ISO 100. Anyway, do you make portret with ISO 1600, crop it 100% , and print on A1??? I don't.


Nagasaki wrote:
Quote:
You've compared two very different cameras.

The 8400 is an 8Mp camera with a 2/3 sensor which is 8.8 x 6.6 mm = 58.08 mm squared. each pixel is 0.00000726

The D40x is 10Mp has an APS-C sensor which is 25.1 x 16.7 mm 419.17 mm squared. each pixel is 0.000041917 mm.

So although there are more pixels on the D40x the individual pixels are an order of magnitude greater in size.

The fact that the larger pixels of the D40x give better high ISO performance than the smaller pixels of the 8400 bears out what the tech corner article says. It's also the reason why we pay out big bucks to buy DSLRs with better image quality.

When viewed on screen the differences will seem less but your screen resolution is probably 1024 x 768 or similar so less than 1Mp. If that's all you want to view at then you don't need 8 or 10 Mp to start with. However if you want to be able to crop areas of the image and print at reasonably large sizes then the image quality at 100% becomes more important.

Ken
emes is offline   Reply With Quote
Old Mar 3, 2008, 1:19 PM   #4
Moderator
 
Nagasaki's Avatar
 
Join Date: Jan 2004
Posts: 897
Default

I thought we were discussing the tech corner.

The article doesn't suggest that sensor design and camera firmware haven't improved. It does suggest that at higher ISO there is some softening of the image caused by noise reduction firmware.

Your comparison of the two cameras seemed to me to suggest that you thought according to the article an 8Mp Camera should have better performance than a 10Mp camera without taking the sensor size into consideration.

I was suggesting that if you want to crop heavilyOR print at A1 (actually I was thinking more like A3 but heh)then the loss of image quality would be an issue but that ifwhat you want to do is look at your picture on screen as shotthen it probably isn't a major issue.
Nagasaki is offline   Reply With Quote
Old Mar 3, 2008, 1:38 PM   #5
Member
 
Join Date: Jan 2006
Posts: 96
Default

Hi
Thanks for reply. This is part of the text: "Back in 2004, manufacturers were content with just cramming more pixels into the same area, however, creating higher resolution images that consistently degraded in quality year by year. While the pixel count was going up, noise went up in proportion, bringing overall image quality down. Cramming more pixels into the same area reduced the pixel size thereby reducing sensitivity of the pixels and increasing noise."
This is not correct for me. Nikon D1 has sensor with 2.7 Mpix and max ISO 1600, D3 has sensor with 12.1 Mpix and ISO 6400. Pixel on D3 is more than 4 times smaller. Is it 4 times more noise?? What about ISO 3200 and 6400. Why D1 cannot use ISO 6400? Forget jpg and algoritams, compare just RAW and ISO 1600.
Rgds
emes is offline   Reply With Quote
Old Mar 3, 2008, 3:05 PM   #6
Administrator
 
Join Date: Jun 2003
Location: Savannah, GA (USA)
Posts: 22,378
Default

emes wrote:
Quote:
Cramming more pixels into the same area reduced the pixel size thereby reducing sensitivity of the pixels and increasing noise."
This is not correct for me. Nikon D1 has sensor with 2.7 Mpix and max ISO 1600, D3 has sensor with 12.1 Mpix and ISO 6400. Pixel on D3 is more than 4 times smaller. Is it 4 times more noise?? What about ISO 3200 and 6400. Why D1 cannot use ISO 6400?
I would not assume that what you see is what the sensor produced with raw files. I can remember Dave Coffin commenting on spikes in the histogram from Nikon RAW files in interviews, indicative of Nikon applying mulltipliers to the data from the sensor.

Dave is the author of dcraw.c, now supporting virtually every popular camera made that is capable of shooting in raw. His code is open source and published for all to see.

He does the work, others reap the benefits by adding nice graphical user interfaces around the base conversion code, more features and variations in the demosaic algorithms and more.

Make sure to see the "Other Raw Decoders" section on Dave's dcraw.c page for a *partial* list of programs that use some or all of Dave's code. He probably knows more about raw conversion than anyone on the planet. ;-)

I would not assume that other manufacturer's don't modify the data from the sensor either before writing out their raw files. I've seen reviews indicative of Sony doing it (but, they do a very good job of it with the latest firmware versions in newer models like the Alpha A700, after a bit of refinement in their code).

We have seen some advances in sensors like better microlens designs, so that the light is better amplified by the microlenses over each photosite. We've also seen advances in electronics in some models to improve signal to noise ratios (so that the signal generated by the sensor for a given number of photons is higher above the noise generated by the other camera components).

So, we are seeing the ability to stuff a few more pixels into the same sensor size while improving quality.

However, in the case of many models with smaller sensors, I personally feel that quality was very significantly degraded with a number of newer cameras compared to some of the older models, as they tried to increase megapixel count faster than other sensor and camera components could keep up.

It's starting to turn back the other direction now. We'll have to wait and see how it goes from here.

Quote:
Forget jpg and algoritams, compare just RAW and ISO 1600.
Why forget? Many users want to have better "straight from the camera" images that they don't have to edit for acceptable results.

As for RAW, you can't compare images from a D1 looking at older reviews, using older raw converters, since image processing has improved signficantly since it was in production. You'd need to see how image quality compares with newer and better raw conversion software to get a better idea of it's strengths and weaknesses.

JimC is offline   Reply With Quote
Old Mar 3, 2008, 3:33 PM   #7
Member
 
Join Date: Jan 2006
Posts: 96
Default

Hi
Thanks for lots of deatil. But how we can compare to sensors? Way described in the article is bit a strange, because I have tried same test with different result. What I can understand you, old camera with less Mpix will produce better pics? So, Nikon D1 with 2.7 Mpix will produce better pics than Nikon D3? Say, on A4 format. Do you agree?
emes is offline   Reply With Quote
Old Mar 4, 2008, 3:45 AM   #8
Moderator
 
Nagasaki's Avatar
 
Join Date: Jan 2004
Posts: 897
Default

The D300 would be a better comparison as that has the same size chip as the D1, the D3 has an FX chip the same size as 35mm film.

But I'd say no way, there have been advances in sensor design and firmware since the D1 was introduced.The article you quote mentions that manufacturers wised up to the fact that they couldn't just keep cramming more pixels on a chip

"Back in 2004, manufacturers were content with just cramming more pixels into the same area"

It is possible that the D40 might produce better images than the D40x under the right conditions. Certainly viewing on screen there is no difference between 6 and 10 megapixels until you start to zoom in. And with these cameras you are talking about similar generation chips and firmware.

What the article is saying is that the image processing to get rid of noise loses some detail. You seem to agree with this but rightly say that it's only really visible at full zoom and you disagree about the extent. Basically any camera will give better results at lower ISO than higher ISO, the degree of degradation and how it is manifest differs with different cameras.


Nagasaki is offline   Reply With Quote
Old Mar 4, 2008, 8:45 AM   #9
Member
 
Join Date: Jan 2006
Posts: 96
Default

Hi
You are right. We have to compare D300 and D1.
I am talking about whole text, not about part of that.
One of the parts is: "While the pixel count was going up, noise went up in proportion, bringing overall image quality down. Cramming more pixels into the same area reduced the pixel size thereby reducing sensitivity of the pixels and increasing noise."
Pixel on the D1 is 4x bigger than on the D300 but is it correct D300 has less sensitivity or bigger noise?? I can not say, sensitivity is lower but processor is better and fixed it. Maybe somebody can.
Rgds
emes is offline   Reply With Quote
Old Mar 4, 2008, 9:47 AM   #10
Administrator
 
Join Date: Jun 2003
Location: Savannah, GA (USA)
Posts: 22,378
Default

There are more variables involved (it's not just the size of the photosites). You'd need to see a newer version of sensor with the same size and resolution as the one in the D1 (with the other advances in technology incorporated into a camera using it) to get a better idea of the difference a larger photosite makes.

See this quote from the first post I made to this thread. It's not just the photosite size.

JimC wrote:
Quote:
We have seen some advances in sensors like better microlens designs, so that the light is better amplified by the microlenses over each photosite. We've also seen advances in electronics in some models to improve signal to noise ratios (so that the signal generated by the sensor for a given number of photons is higher above the noise generated by the other camera components).
JimC is offline   Reply With Quote
 
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



All times are GMT -5. The time now is 1:32 PM.