Home News Buyers Guide About Advertising
 
 
 
   
Page 3 of 9 FirstFirst 12345 ... LastLast
Results 21 to 30 of 87
  1. #21
    Join Date
    Jul 2005
    Posts
    6,590
    On the left two ISO 1600 100% crops where no noise filtering has taken place. This is the DPP interpretation of the RAW data.

    On the right the same crops, but with chroma noise filtered.
    Attached Images Attached Images  
    Canon EOS 350D, Tamron SP AF 90mm F/2.8 macro, Sigma 18-50mm F2.8 DC EX, Canon EF 70-200mm f/4 L USM, Tokina AT-X124 Pro 12-24mm F4, Soligor 1.7x C/D4 DG Teleconvertor, Manfrotto 724B tripod, Canon Powershot S30

  2. #22
    Join Date
    Jul 2005
    Posts
    6,590
    Same 100% crops, but now from the ISO 3200 shot.
    The dithering from the chroma noise filtering gets more apparent.
    Again, left unfiltered, right filtered.

    ** note, this file has been saved twice as JPEG, because the size exceeded the DCRP upload limit, and the "original" was already closed in photoshop.
    Attached Images Attached Images  
    Canon EOS 350D, Tamron SP AF 90mm F/2.8 macro, Sigma 18-50mm F2.8 DC EX, Canon EF 70-200mm f/4 L USM, Tokina AT-X124 Pro 12-24mm F4, Soligor 1.7x C/D4 DG Teleconvertor, Manfrotto 724B tripod, Canon Powershot S30

  3. #23
    Join Date
    Jul 2005
    Posts
    6,590
    Finally the ISO 6400 100% crops. You see a very Nikon like dithering, or "luminance noise" pattern after the chromatic noise has been filtered.
    Left no noise reduction, right the chroma noise reduction crops.

    Not bad for ISO 6400, on a 3 year old camera (2 DSLR generations ago), don't you think?

    ** the odd dithering in the red area is from the JPEG reduction, needed to fit the image in the 237k size restriction.
    Attached Images Attached Images  
    Last edited by coldrain; 01-23-2008 at 12:22 PM.
    Canon EOS 350D, Tamron SP AF 90mm F/2.8 macro, Sigma 18-50mm F2.8 DC EX, Canon EF 70-200mm f/4 L USM, Tokina AT-X124 Pro 12-24mm F4, Soligor 1.7x C/D4 DG Teleconvertor, Manfrotto 724B tripod, Canon Powershot S30

  4. #24
    Join Date
    Jul 2005
    Posts
    6,590
    Quote Originally Posted by VTEC_EATER View Post
    But there is chroma noise in my RAW image. The reds areas have a distinct pink/purple chroma noise, the blue areas have a greenish chroma noise, and the black areas have a bluish chroma noise.
    Yes, you see a very little bit left that the chroma noise filtering has not touched. And you do see the dithering you see when you filter chroma noise, as you can see in my demonstration.

    Quote Originally Posted by VTEC_EATER View Post
    The tests performed by DPReview also show the D300's chroma noise in its RAW files. Unfortunately, I can not direct link their images so I will give the link for people to check out for themselves:

    http://www.dpreview.com/reviews/sonydslra700/page18.asp

    The RAW image tests clearly show the D300 with chrominance noise present. Also, when compared to the A700's images, you can clearly see that the Nikon is not using noise reduction algorithms to clean anything up. No smudges, no dithering. So, its pretty hard to say that the D300/D3 performs noise reduction no matter what.
    You clearly do see the odd dithering, often referred to as luminance noise.
    And you do see a HUGE difference in chroma noise compared to the A700, which still uses the same sensor technology as the D300.
    Both things do point to the fact that Nikon does filter chroma noise, since the introduction of the D200 at least.
    This actually is quite a well known fact though.


    Quote Originally Posted by VTEC_EATER View Post
    Yes, this is true. The camera says it will not perform long exposure noise reduction on images with a shutter time of less than 8 seconds. However, I do want to try a few tests to see if there is some of this noise reduction at speeds shorter than 8 seconds, or if there is noise reduction performed even when you have the setting turned off. For some reason, I just can not believe that the images on the other site are accurate. Those 100% crops are terrible from the D3, and clearly look like noise reduction was turned on. I will post my results.
    chroma noise reduction and long exposure noise reduction are two very different things....
    Canon EOS 350D, Tamron SP AF 90mm F/2.8 macro, Sigma 18-50mm F2.8 DC EX, Canon EF 70-200mm f/4 L USM, Tokina AT-X124 Pro 12-24mm F4, Soligor 1.7x C/D4 DG Teleconvertor, Manfrotto 724B tripod, Canon Powershot S30

  5. #25
    Join Date
    Aug 2006
    Location
    Chicago, IL USA
    Posts
    935
    Quote Originally Posted by coldrain View Post
    Yes, you see a very little bit left that the chroma noise filtering has not touched.
    If there is chroma noise still present in the image, how can you say that there has been chroma noise removal? I'm just not following the math here.

    And you do see the dithering you see when you filter chroma noise, as you can see in my demonstration.
    Okay, maybe I'm not as much of a pixel peeper as you are here, but I do not see any more "dithering" in my D300 image than your XT images show. In fact, Im not quite sure exactly where you see any dithering in the images.

    You clearly do see the odd dithering, often referred to as luminance noise.
    Could this in fact be just luminance noise, and not "dithering"? If the answer is no, then I would counter you with saying that Canon uses Luminance noise removal in all cases because you can clearly see that the luminance noise is far too low for a sensor of that generation.

    We could go back and forth for days on this, but the fact is, you can not "prove" your statement with any definitive evidence.


    Edit: Unfortunately it just finished snowing here and it is overcast tonight, so no astro-photography tests tonight. Sorry.
    Nikon D300 | MB-D10 | Nikkor 12-24/4 | Nikkor 50/1.8 | Nikkor 70-200/2.8 VRI | Sigma 18-50/2.8 | SB-800 | SB-80DX (x4) | Radiopopper JrX Studio |

  6. #26
    Join Date
    Dec 2006
    Location
    God's Country - Australia
    Posts
    10,421
    ok...what the hell is dithering ?
    D800e l V3 l AW1 l 16-35 l 35 l 50 l 85 l 105 l EM1 l 7.5 l 12-40 l 14 l 17 l 25 l 45 l 60 l 75
    flickr

  7. #27
    Join Date
    Aug 2006
    Location
    Chicago, IL USA
    Posts
    935
    Quote Originally Posted by Rooz View Post
    ok...what the hell is dithering ?
    Exactly!!!

    It took me half the day to find some logical explaination of it:

    http://www.cambridgeincolour.com/tut...mera-pixel.htm
    Last edited by VTEC_EATER; 01-23-2008 at 05:54 PM.
    Nikon D300 | MB-D10 | Nikkor 12-24/4 | Nikkor 50/1.8 | Nikkor 70-200/2.8 VRI | Sigma 18-50/2.8 | SB-800 | SB-80DX (x4) | Radiopopper JrX Studio |

  8. #28
    Join Date
    Dec 2007
    Posts
    30

    Exclamation

    Quote Originally Posted by coldrain View Post

    Not bad for ISO 6400, on a 3 year old camera (2 DSLR generations ago), don't you think?
    Hmm it seems you are the only one who can shot at iso6400 with eos 350 ... Nevrmind, yes, we all know that your camera is the best of the best ! But please don`t turn this topic into praise of it !Maybe it`s time to begin praise the new EOS 450 - don`t you think Thank you in advance .

    2All Please stick to the theme and don`t get into total offtopics ! Or just open your own new topic ...
    Last edited by dolphin; 01-24-2008 at 12:27 AM.

  9. #29
    Join Date
    Jul 2005
    Posts
    6,590
    Anyone can shoot ISO 6400 with an EOS 350D, you can shoot ISO 6400 shots with your D40x too.

    I even explained how, in my post above. When you shoot with 1 f-stop under exposure, and you then push the exposure up 1 f-stop again in your RAW converter, you basically do the same as your camera does when you go a step higher in the ISO settings.

    So, if I shoot with 2 f-stops under exposed at the camera's ISO 1600 setting, and then push the exposure in the RAW converter up 2 f-stops, you get an ISO 6400 photo. And you see the exposure time go down 4 times.

    Rooz, I use the term dithering because the filtered chroma noise pixels make the image LOOK dithered. It is not an official term to describe what the sensor does. Dithering is when you use different coloured dots to together form another colour, it is the way your inkjet printer prints its photos.

    When you look at those prints very up-close, you lose the solid colours it seemed to have, and you see the dithering.

    And that is what this noise removal makes a photo look like too. To counter that, noise removal programs like Noise Ninja smooth over the pixels, which is why noise removal often makes photos softer.

    VTEC, why do you want to hear the D300 does not filter chroma noise?
    Explain why the D300 has a lot less chroma noise than the A700, with the same sensor generation and sensor technology?
    Then explain how a bayer array sensor can actually cause so much luminance noise (it can't)?

    You can filter chroma noise very harshly, or you can do it less severe. You can use all kinds of algorithms to go about it too. If you do it too harshly the results are not going to be great.
    Nikon has chosen to filter chroma noise, and in their own chosen strength and with their own chosen way.
    So, you still can see choma noise and colour blotches. But way less than you would without the filtering. And in its place you get lighter dots, that you call luminance noise.

    But... sensor diodes do not see all colours. In fact, they only see light intensity. This makes them colour blind. Not so nice, one could only make black and white photos if it was left that way.
    So, what have they done? Then put filters in front of the pixels. Some pixels only get to see the green component of the light. Some only get to see the red component, and some only get to see the blue component.
    Now... if a pixel gives an erroneous value (and that happens a lot with higher ISO settings) the value will be TOO red, or TOO green, or TOO blue.
    Not TOO red and TOO green and TOO blue.
    So... a pixel gets a wrong colour, not just the right colour but with a wrong intensity. And that is why chroma noise always increases with a bayer array sensor... Also when they are put in a Nikon DSLR.

    So yes, when you see "luminance noise" in the way the high ISO results of the D200 and D300 (and D3) show it, you know chroma noise has been filtered. Bad? no, why would it be bad. But do take it into account when blindly comparing such results as dolphin has compiled from some website.
    As my examples show, chroma noise can be filtered with the other cameras too.

    The same sensor technology in the Pentax K10D and Sony A100/200 shows this chroma noise. Yet, the Nikon D200, D40x and D80 show it in different ways much less. Why? Because Nikon filters it. Nothing strange about that.

    Same with the A700 and D300. A700, much more chroma noise, much blurrier also because Sony uses a "wrong" kind of noise filtering.

    This is NOT a controversial point.
    Canon EOS 350D, Tamron SP AF 90mm F/2.8 macro, Sigma 18-50mm F2.8 DC EX, Canon EF 70-200mm f/4 L USM, Tokina AT-X124 Pro 12-24mm F4, Soligor 1.7x C/D4 DG Teleconvertor, Manfrotto 724B tripod, Canon Powershot S30

  10. #30
    Join Date
    Aug 2006
    Location
    Chicago, IL USA
    Posts
    935
    Quote Originally Posted by coldrain View Post
    VTEC, why do you want to hear the D300 does not filter chroma noise?
    Explain why the D300 has a lot less chroma noise than the A700, with the same sensor generation and sensor technology?
    As silly as it is to argue that my D300 has more noise than your 350D (which is subjective at best), I'm merely trying to point out that the broad generalization of "Nikon uses noise filtering in its RAW files" is in fact not true, or there is not enough evidence to prove this theory. As I mentioned above, the same could be said of Canon filtering luminance noise in its RAW files. It may be true, but is not something you can easily prove unless you were a programmer for Canon.

    As for comparing the sensor noise to the A700, well its not something you can directly compare with any conclusive evidence. Yes, both use the same sensor technology and were probably both developed side by side by the same engineers. The problem with trying to compare them has to so with Sony's noise reduction. A lot of the "detail" was lost with its noise reduction and smudged the characteristics of the noise pattern. The D300 has a much "sharper" noise pattern to Sony's "blotchy" pattern, and unfortunately that blotching spread the chroma noise over more pixels, I assume, in an attempt to remove it. So, it is difficult to say exactly how much chrominance noise the A700 sensor produces.

    If both cameras did in fact produce "digital negatives" in its RAW files it would be much easier to compare their images.

    So yes, when you see "luminance noise" in the way the high ISO results of the D200 and D300 (and D3) show it, you know chroma noise has been filtered.
    Im still not sold on this theory. I can believe that the D200, and Nikon's other CCD sensor cameras, filter some chrominance noise, but it [chrominance noise] is clearly present in the D300's RAW files. I know you are saying that it is filtered, even if just a small amount, in the D300, but no one can really prove how much, if any, really was. The detail is still sharp, there is no artifacting, or blotching, or even dithering, that I can see in the 100% crops. To me, this looks like a digital negative. If any chrominance noise reduction is in the D300's RAW files it is very small, and very good, IMO.


    Same with the A700 and D300. A700, much more chroma noise, much blurrier also because Sony uses a "wrong" kind of noise filtering.
    Yeah, something just went a little hooey with Sony's filtering. Maybe it inst bad when you look at the entire image scaled down, but 100% crops show some serious smudging, which is unfortunate. I try to keep as little of the processing to the camera, and prefer to do my own processing in Photoshop, or Capture NX, because there is greater control of the settings. This is why I shoot without high iso noise reduction on. I have never been satisfied with the images when it is turned on, even at its lowest setting. I feel Neat Image produces a much nicer, and sharper, image than Nikon's noise filtering.

    I will probably use long exposure noise reduction, but I have not had the need to take 8 second exposures yet. Maybe when I try out those astro-photography tests, Ill try a few different settings.
    Last edited by VTEC_EATER; 01-24-2008 at 06:46 AM.
    Nikon D300 | MB-D10 | Nikkor 12-24/4 | Nikkor 50/1.8 | Nikkor 70-200/2.8 VRI | Sigma 18-50/2.8 | SB-800 | SB-80DX (x4) | Radiopopper JrX Studio |

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •