D50 better than D300 in ISO

Discussion in 'Nikon' started by rob_piontek, Jun 5, 2009.

  1. I had seen some reports that this was true, but after buying a D300 and making the comparison now I know for myself. There is so much hype around the D300 and it's ISO capabilities, it's hard to know what is what.
    What I found, comparing RAW images in Lightroom with no NR, was that the D50 is better at both ISO 200 and 1600 than the D300. The D50 is close to noise free at 200, but the D300 is definitely not. At 1600 I expected the D300 to be better, but it's just not the case.
    I tried resizing the ISO 200 D300 image to a 6MP image, the same as the D50, and then the noise is basically the same between the two. Slight edge to the D300, as the resolution of the resized D300 image seemed better than the original D50 image if I looked at fine text which was just below the resolution limit of the D50.
    Anyway, I just wanted to post this. If I had known this beforehand I might not have bought the D300.
     
  2. I'm not sure if this is the case but often times lower MP sensor show less noise. With the noise reduction on and at higher ISO's the advantage would probably go to the D300. Usually when adding MP to the same size sensor will increase noise as a side effect.
     
  3. why don't you post some samples. I find it hard to believe- I'm curious now
     
  4. Rob, I guess you know that, If you don't post some images, many people is going to find difficult to take you seriously.
    <p>
    Also, It really impresses me that the only reason to buy a body like the D300 is the High ISO capability. I only have the D40x myself and I am allways thinking about buying the D200. They have the same ISO quality but D200 offers much more to me than the D40x.
    <p>
    The same happens with the D300, even if what you say It's true, I guess that D300 It is just much more than ISO noise compared to the D50.
     
  5. Yes, of course you are right. But years have passed since the D50 was introduced. I had hoped that better ISO performance would offset the increase in pixel count.
     
  6. I have the D300 and the D50. I shoot almost exclusively raw with both cameras and process with Capture NX2. I dearly love the D50 because it is compact and light. But you are wrong. When processed properly the D50 images aren't as good as those from the D300. Maybe it's because you are processing them with Lightroom rather than NX2. NX2 is the ONLY program that reads all the information in a Nikon raw file.
     
  7. There was an old published report and graph that compared dynamic range of Nikon and perhaps few other brand cameras. Surprisingly to some and challenged by others, the D50 (6 MP CCD sensor) camera had higher dynamic range than some 10 MP CCD and 12 MP CMOS cameras. I bet someone remembers that report and graph ? - perhaps from 2007.
     
  8. wow, I didn't know my little d50 (bought used for $299) had such a fine little engine inside!
    But what about the hype about the new cmos sensor high iso performance of the d90, 300 or 700?
    I am still planning to upgrade, for other reasons as well as the (may be not?) better high iso performance.
     
  9. Even the noisiest of images can have the noise removed. But reduced noise generally results in loss of detail. A lot has to do with camera settings, format being shot (RAW vs JPG) and of course post processing. While the D300's sensor is not a huge improvement over older cameras (like the D3/D700 and now D90 do), it does offer noticeable/visable improvement over older cameras. If you refer to the DXOMark site, there is little difference in the Low Light ISO rating between the two cameras - the important differences can be found elsewhere - the D300 has higher dynamic range, color range and and overall higher sensor score. These differences combine together to give the D300 a solid advantage over older Nikon DSLR cameras including the D50.
     
  10. I'm still waiting to see samples before I believe any of this, because high-ISO performance is something the D300 is touted for.
     
  11. I've used both cameras. This is false.
    Both cameras have been tested thoroughly by many publications and pro's so this is not just me saying this. The d90 ISO performance is no better then the D300 either. If anything they are nearly same w/ a slight edge going to the D300.
    One issue I see with the original poster is the fact he's using lightroom and shooting in RAW. All camera settings are lost so I'm sure the images look pretty dull compared to a JPEG in Lightroom or using RAW images in Capture NX. As far as I know lightroom cannot read the camera settings on a D300 shooting in RAW. Can anyone confirm this?
     
  12. I do find this hard to believe.
    Please provide NEF RAW files from both cameras of the exact same subject shot under the exact same circumstances so we can compare ourselves by processing them. Maybe you're confusing it with all the consumer blows & whistles (I think that's the expression) that come with a consumer body just to ensure that jpgs look good out of camera. Oh & I should add that I now own a D70 (not the D50 for sure) but it's no where near the camera my D300 is.
    I don't want to see any small 700 pixel shots for this one.
    Thank you & eagerly awaiting those RAW files. :)
    Lil :)
     
  13. Indeed, this can not be taken seriously without examples.
    The D50 is from before my time in the digital world - however, I started off with a D40, 6mp like the D50, perhaps the same sensor? When I switched to my current D90, similar to the D300, I noticed a major improvement in high ISO performance.
    Please post sample images to corroborate your observations.
     
  14. Look here:
    http://dustylens.com/D300-ISO-200.jpg
    These were converted with ACR, which should be the same as Lightroom. Look closely at the color chart and you can see the noise at ISO 200. If the same scene was shot with the D50 you wouldn't see this. At ISO 400 from this same site, you don't have to look close any more to see the noise in the D300 image. I'm telling you the D50 would keep a slight advantage to 1600. I have heard other people on the web talking about the noise present on the D300 at low ISO. I'm not the only one. Since I had seen these comments before I bought the camera, I figured at least the D300 would shine at 1600, but I don't see it.
    If you look around, you will find plenty of people, from a few years ago, saying that the D50's ISO was/is better than the D200/D80. Based on this as well it's not off the wall to think that it can keep up with the D300.
    Of course it's also 6MP vs 12MP. A better test would be in large prints - who wins resolution or noise? Probably the D300 enlarges a bit better.
    Anway, I don't really feel the need to convince anyone of anything. I Just wanted to share my experience. To summarize I would simply tell those of you looking at a new camera not to expect anything spectacular in terms of noise performance from the D300. If you're coming from the D40X/D60/D80/D200 maybe it's a bit better. But not from the D40/D50/D70. Wait for FX if better ISO is what you really want. If you don't believe me, fine, but I'd challenge you to make the same side by side test, using Lightroom. Maybe I wll try NX, too.
     
  15. Not to highjack this thread, but regarding the reference to LR and processing RAW files: LR uses ACR to convert the RAW images. While ACR may not translate every bit of NEF metadata, ACR is certainly capable of translating the sensor bits. With modest effort one can easily come up with default settings in ACR that will do more than an adequate job of rendering sharp, colorful images from the RAW files. As to the OP's original statement - as someone else suggested, please post images.
     
  16. I have and use both cameras...I bought the D50 upon it's release the images from the little guy are pretty incredible..I've done 8 x 40 inch pano's and had some 36 x 48 images printed from that camera....and they are tack sharp......
    but the D300 is superior in all aspects...including noise....I have at 1600 iso with both and D300 is much better.... images from the D50 at 1600 are usable..but require some tweaking....
     
  17. Here's another guy who came to the same conclusion as me, but with camera jpegs:
    http://www.flickr.com/groups/d40/discuss/72157606973945716/
     
  18. Rob, you are making statements which are contrary to results most others experience or expect so substantiating your findings can be important.
     
  19. ShunCheung

    ShunCheung Administrator

    Rob, since you have both cameras, could you provide some side-by-side A/B comparisons between the D50 and D300 at the same ISO, same lens, same aperture ... essentially same everything else to demonstrate your claim?
    Since you mentioned both ISO 200 and 1600, could you provide two sets of comparisons at those ISOs, respectively?
    P.S. The dustylens.com link Rob provided above is Steve Bingham's site. That is not Rob's own image.
     
  20. Rob -
    now I see you mention camera's jpgs.
    Stop right there - we want RAW NEFs side by side comparison please.
    So - please since you have both cameras, would you mind taking some side by side, same lens, shots at the same time with the same RAW settings so that we truly can compare.
    Thank you
    Lil :)
     
  21. I recall being very impressed with the D50's high ISO performance when it first came out. Heck, I may still grab a used one to replace my ailing Olympus C-3040Z P&S. Most objective tests done at that time showed it was better at ISO 1600 than my D2H.
    But the reviews I've seen that use standard testing methodology, such as dpreview, do show that the D300 has an observable edge over the D50.
     
  22. I can't show you RAW files right now because I'm home and I don't have fast internet. Looking again, right now, the ISO 200 files are pretty close, there is not much noise in either, but still a very slight edge to the D40. By the way, these crops are with the D40, not the D50. So, here, ISO 1600. First the D300.
    00TZcX-141299584.jpg
     
  23. And here is the D50 file...
    00TZcl-141299884.jpg
     
  24. jbm

    jbm

    Yeah, there is some noise even at 200 in my D300. It doesn't really matter, though, in practical use...still it's worth noting that the D50 sensor is pretty solid. Isn't that the sensor that was in the Epson RD-1, also?
    Jay
     
  25. To me the D40 looks better. You might say they are the same, but there is no way the D300 is better. If you say they are the same, then you are better off with 12MP than 6MP. This all only applies to this test, of course! YMMV!
    I guess I will add that these were shot inside, of course, window light, tripod. 17-55 2.8 at 5.6 and 1/25s at 1600. I adjusted the WB by hand in LR.
     
  26. I love my D50,but I wish it had a larger viewfinder.
     
  27. The thing that bothers me most about the D50 is how slow it is. Slow to clear the buffer, and slow to preview images. And the small screen. The D40 fixes these things. And supposedly the viewfinder is better as well, but to tell you the truth I've never noticed a difference between them.
     
  28. For the matter... Dpreview tests:

    D50
    http://www.dpreview.com/reviews/nikond50/page17.asp

    D300
    http://www.dpreview.com/reviews/nikond300/page18.asp
     
  29. A question Rob,
    what is you NR setting in camera? The general consensus is to have it at no NR in camera if I'm not wrong....
     
  30. To make a better comparison I downsampled Rob's D300 example so that it was the same size as the D40 (or is it D50?) one. After doing that they look much the same to me.
    I am not surprised since I remember looking at the JPEG noise measurements on Dpreview for D40 and D300 and finding them much the same at a given ISO.
    Don't people say that the D40 is about 1 stop better than D200/D80? Don't people also say that the D300 is about 1 stop better than D80/D200? If that's so then D40 and D300 should be similar.
    Mind you, noise and sharpness are always a tradeoff so be sure to compare sharpness as well as noise.
     
  31. I am looking at the same files posted above in my new Demo copy of NX. To me there isn't much difference. The D300 preview maybe looks a bit better compared to the D50, at 1600, but it also looks like there is more NR going on. The blacks especially show much less noise, but if you look at the shadows, the D50 retains more detail. It's interesting.
    So I would challenge those of you who say the D300 is clearly better than the D40/50 to show us the goods! To tell you the truth I would be happy for you to show me that I actually got something in terms of image quality for the 1400 euro I just spent!
     
  32. The D50 was my first DSLR, and I still have it, and use it mostly as a backup to a D90. I don't know if it has better noise reduction than the D300, but I do know that it is pretty good at 1600 ISO, particularly with some mild Noise Ninja applied with the proper profile. In the attached image, I apologize for cutting off the bird's tail, but I wanted to keep within the 700 pixel limit for online presentation. I don't care which is better, only that the D50 is very usable.
    00TZeM-141309584.jpg
     
  33. While it is not at all impossible that the statement is true and the dpreview test can be interpreted to point in this direction, the actual side by side test is fairly easy to execute, so why not do it and show? It is, of course, a bit hard to compare the considerable difference in resolution, but that can be approached in several ways.
     
  34. Evaluating noise in images is a tricky business. So it is no surprise that without a controlled side by side comparison anybody can claim anything and nobody will believe it :p
     
  35. I have a D40 and a D300. The truth is that improvements in CCD/CMOS technology has been relatively undramatic. Both cameras can shoot really good pictures - even enlarging up to 20x30. The D300 can do more, work with more lenses, do CLS, etc.
    But the trusty little old D40 sure is nice and light weight when you are out hiking.
     
  36. It just proves a couple of things. To the average Joe, the need to upgrade continuously is a marketing con because unless you are printing big images, you don't need a big sensor. Another thing is that NX2 certainly gets more out of a .nef file than Lightroom...as you'd expect. Nikon aren't going to give all their IP to Adobe, are they? I've just spent a week discovering that. Had the images been uploaded to NX2, saved as a .tiff, and then imported into Lightroom, you would see the difference.
    I bought a D300 because of its quality and body integrity, depth of picture controls and compatibility with my old, but fast, manual lenses. It's purchase will protect my long accumulated Nikon investment. I doubt that my limited expertise will ever exercise its capability, so why would I want anything more? One day I might go to a D700, but not for now.
    Any 6mp+ digital camera with a good lens will produce nice pictures. My Canon G7 with its 2.8 lens is one. I bought it for that lens, not the body. Its only when you want instant startup, fast focus, big viewfinder, big LCD, interchangeable lenses, would you need to go further. Horses for courses.
     
  37. "Anyway, I just wanted to post this. If I had known this beforehand I might not have bought the D300." Some people don't get it and never will.
     
  38. If you just bought camera, simple return it to store for refund. D300 never was hi ISO camera and never meant to be, pixels too small. That's why D700 better in hi ISO then D3X.
     
  39. So first off, the science behind matters such as these is beyond the comprehension of most of us.....but let’s get beyond the crazy enlargements, and ridiculous comparison of minor amounts of noise difference. Until just a few days ago I owned the d70, d50, d200, and d300, and their noise performance was in my opinion, ranked in the same order. I know from actual use, that the d300 is better in all situations.... I don't need to post controlled tests, at a standard setting, and blow it up to 300x to know that. Why? because I see it in real life! If I sound irritated it’s because people continue to post hostile threads like this that seem like they are intended solely to stir up controversy. Seriously, do these infinitesimally small details really matter to the photographer? Or just the gear jockey? I implore all of you to get out and shoot, use your experiences to be the judge of a cameras performance, and stop staring at numbers and graphs and giant enlargements of another guys OFFICE! Ps....I love all of those cameras, and it’s a pity I couldn’t keep them all!
     
  40. Rob, you still haven't posted RAW photos as requested multiple times. Please take a number of identical pairs of exposure using the D50 and D300 on a tripod and post the RAW files.
     
  41. This should not surprise you if you take the following in account:
    1. CMOS is a inherently noisier technology than CCD. The main reason for the move to CMOS is that CCD is much more power hungry.
    2. With the same sensor size the higher the resolution the higher the noise
    3. It was common knowledge that the D50 of the 6MP nikons D50, D70, D100 all using basically the same sensor, had the cleanest files.
    In the end the D300 wins out with better processing, and higher Dynamic range.
    I do not understand that the OP is not willing to put up two shots in nef; they do not even have to be processed as this is the nikon forum most of us posses the tools to download the file and procesit for them selves.
    An other point is that when using the latest version of nx you are basically processing the D50 file with the same engine as in the d300, as that is a software redering of the expeed chip and the same algorithms are used in both.
    In the end the D50 is profiting from the improved noise reduction on the cmos camera's.
     
  42. So I would challenge those of you who say the D300 is clearly better than the D40/50 to show us the goods!​
    So a claim is posted, backed up with at best anecdotal evidence and then the naysayers are challenged to prove their point? This isn't politics, this is science and technology and seems to me that you don't know how scientific research and debate is done.
    FWIW, I upgraded to a D300 from D70. The ergonomic improvements are considerable (better manual focus, better AF, better viewfinder, more rigid body, meters on manual lenses, PC sync and cable release sockets), but in a 8x12" print there is a clear difference in favor of the D300 due to the resolution advantage.
     
  43. I will post the corresponding RAW files on Monday. You can see for yourself and decide for yourself. I have only dial up this weekend.
    Looking at them in NX2, though, I would argue that Nikon is not passing NR algorithms from new cameras to the old. It makes sense. Why would a camera company want to make images from old cameras look better. In LR the RAW images look pretty similar, slightly different color, different resolution, slightly different noise. NX2 must be designed to give you the same JPEGs you get out of the camera, not necessarily more recent improvements which come from software. This is just my interpretation.
    Also, I had other reasons for buying the D300 besides noise. If I only cared about noise and thought I could get it from the 12mp DX chip I would have bought the D90 or D5000. Nor did I say that the D50/40 is a better camera. It was not my intention to start controversy or imply that the D300 is not a great camera. I was simply surprised, and a bit disappointed when I compared, and felt it was worth sharing. Nothing more. Take it or leave it it's up to you, if you don't care then I totally understand, but there's no need to insult me. I'm just a guy with these cameras who took a closer look.
     
  44. I used to have a D50, and now own a D300. So this thread made me curious (curious enough to sign up). I checked RAW files from both yesterday in ViewNX. Unedited files, shot at ISO800 and 200.
    It is somewhat true that the D300 is not noise-free at ISO200, though it's more a fine grain than anything else. At 800, I find the D300 really good. 1600 is very usable.
    The D50 at 800 is good, I found 1600 quite bad; hardly ever used it.
    Another thing: I've used ACR in the past, but the handling of ACR on D300 ISO1600 files made me buy Capture NX2. I find it seriously below par, far too much need to use noise reduction with the inevitable loss of detail. Nikon software handles them far far better (IMHO). For the D50 and D80 RAWs, I never had any issue with ACR, so using Adobe products to compare does not do the D300 any justice in my view.
    However, to me the biggest difference is not so much the amount of noise but the type of noise. The D50 has colour noise (at ISO800), the D300 luminance noise. Colour noise is plain ugly, luminance noise looks like a grain which is fine by me.
    And because of that, I would not say that the D50 is better at higher ISO than the D300. Maybe the amount of noise may be equal, but the way it looks sure doesn't.
     
  45. All 3 points Erwin listed are incorrect:
    • Modern CMOS image sensor technology is much superior to CCD in noise, speed and power. The pixel architecture termed '4T' or 4-transistor per pixel is actually a little CCD in each pixel, complete with a pinned photodiode, a transfer gate and a sense node.
    • A smaller pixel will typically hold less charge so the signal-to-noise ratio is degraded, but not the noise, which is due to the readout circuitry, not the pixel size. Higher resolution lowers the SNR but has no impact on noise.
    • From this review: "The CCD sensor the D50 uses is unique. No, it's not the one used in the D100 or D70 models, though it is certainly derived from the same technology." In the same review, it is noted that the D50 tends to overexpose leading to blown out highlights. Since noise is typically less evident in overexposed images, this may help explain Rob's observations.
    Erwin Baeyens , Jun 06, 2009; 01:49 a.m.
    This should not surprise you if you take the following in account:
    1. CMOS is a inherently noisier technology than CCD. The main reason for the move to CMOS is that CCD is much more power hungry.
    2. With the same sensor size the higher the resolution the higher the noise
    3. It was common knowledge that the D50 of the 6MP nikons D50, D70, D100 all using basically the same sensor, had the cleanest files.
     
  46. These were shot in M, there was no metering involved, and the exposures are pretty close, about 1/3 stop I'd day.
    I think I basically just figured it out, why everyone thinks the D300 is so much better. Here I will post the JPEGs, from NX2, then imported to LR to do the crop. I'm sorry the crops are slightly different than the above. I think I left the exif in this time so take a look if you're inclined. Here is the D300 from NX2...
    00TZy0-141513584.jpg
     
  47. And here the D50 NX2 image...
    00TZyD-141513684.jpg
     
  48. And one more time, converted from LR, with the same crop... D300....
    00TZyJ-141515584.jpg
     
  49. D50 convert in LR, same crop as above...
    00TZyK-141515684.jpg
     
  50. What I really notice flipping back between these, is that NX2 really crushes the blacks in the D300 RAW, but this is not the case with the D50. Look at the black between the two D50 images. Looks the same, but less noisy in NX2 compared to LR. Now compare the D300 LR and NX2.
    It's also interesting to compare the file sizes. From NX2 to LR the file size change is more dramatic for the D300 file than the D50. This could be interpreted, I think, as NX2 being more aggressive with NR in the D300 image.
    I agree with the comment about color noise. But LR can remove color noise no problem. It's the luminance that is more difficult, at least this is my impression, and I agree that I have never been happy with what LR is doing here.
    By the way in NX2 I left everything to whatever default settings are. I only adjusted the WB.
     
  51. ShunCheung

    ShunCheung Administrator

    I assume both D300 images are from the same original. Somehow the NX2 coverted version has noticably less noise than the LR conversaion. Pay attention to the white wall on the right side. The LR version is significantly soiser.
     
  52. Both NX2 images are from the same files as the LR images. Noise reduction was off in LR. This is how I prefer to compare the two cameras. The NX2 images have whatever default NR is set to. This could depend on how my camera is set, which didn't occur to me until now, and actually I don't know, because I don't normally shoot JPEGs. Anyway, you are right, in both the D50 and D300 NX2 is doing some NR.
    If you want to come to some conclusion about which has better noise performance, then I think the LR images are the better comparison, because there is no NR applied. But, if you convert with NX2, then I could see why you'd choose these to come to whatever conclusion.
     
  53. Dan Z, thats the canon 50D not the Nikon D50.

    Umm is it just me or is there something wrong here. The exif says it's a D40, not a D50. Also there is a -0.33 EV in the D40 images and 0 in the D300

    Also the in-camera picture settings in the D300 are set to 'contrast - "hard"' and 'saturation - "high saturation"' as opposed to "normal" in the D40.
    Maybe minor things, but still not a true side by side comparison.
     
  54. Rob, thanks for posting those samples. Certainly, the D50 and D300 look much closer in noise than I would have guessed. I don't really care so much about why, and I still tend to think that when you parse the details the D300 (or D90) is the better high ISO camera. However, it is interesting to see how close the older generation is in this regard, at least in some instances. My D40 never looked half as good as my D90 at high ISO.
     
  55. Rob, you need to have all settings flat and default (no optimization, no d-lighting, etc.) and also use the exact same exposure in Manual mode on both cameras for an accurate comparison.
     
  56. This thread is very amusing. Not informative, but amusing. Sorry to be so flippant, but if you are going to make statements that go against the common wisdom, at least compare apples to apples. The OP may well be correct. However, he has offered no proof at this point. As others have requested, take identical images with both cameras and post them along with teh EXIF data. How hard can this be?
    Also, if you are doing this with NX(2), you should be aware that it will pickup the camera settings and appky them to the raw file, by default. ACR does not.
     
  57. Ok, you are definitely right, the NX2 comparison is flawed, but this was not the original test I set out to do. I did the NX2 conversion after the fact just because I was curious. I don't shoot JPEGs, I shoot RAW and use LR.
    The LR comparison is fine.
     
  58. I found the D90 (similar sensor to D300) noise levels were not better than the D50 at ISO 200 in JPEG, and only slightly better at ISO 400. However, at ISO 800 and higher the difference was obvious. I'm surprised you would find the cameras even close at ISO 1600 -- I think something needs to be reexamined there. Still, I'm glad you posted your observation because many people can't believe that the D50 is that good a camera, but it really is.
     
  59. This is why I still use a D1H on the rare occasion i shoot a frame of digital.
    The D300 is a much better camera than the D1h, but the 2.7MP sensor does have HUGE pixels which reduces many of the noise problems.
    Some simple math will show that the size of a D300 pixel is just 2.9313x10^-5 mm^2 while the size of a D1H pixel is 1.4222x10^-4 mm^2 . The D1H pixel is 6.14872 times as large as that of the D300. Thus, it can absorb 6.15 times as much light in a given time and use that to produce an image that has more accurate information at a smaller number of points.
    One could argue that doing a well conducted survey of 2.7 million people is better than doing a poor survey of 13.1 million people.
     
  60. Personally I think Lightroom, even though I use it for it's workflow, does a particularly bad job with noise reduction on D300 files. When compared to NX it's pretty shabby and there seems to be a lot of low ISO noise. Sadly in some ways what it says is the D300 is in fact surprisingly high noise and Adobe doesn't know how to hide it as well.
    Whatever - I still am happy with my D300 and it performs way better than my old D80. I always have my S5 for the areas it doesn't cover (it does not have low ISO noise at all that I can find).
    Ultimately isn't the D50 a repackaged D70 sensor? In that case, are we saying the D70 is better than the D300 here?
     
  61. Well, DxO does seem to support the contention:
    http://www.dxomark.com/index.php/eng/Image-Quality-Database/Compare-cameras/(appareil1)/270|0/(appareil2)/295|0/(onglet)/0/(brand)/Nikon/(brand2)/Nikon
    Given the smaller pixels ("light buckets" as Thom Hogan calls them I think) it makes some sense...
     
  62. As Hamish already pointed out...
    The exif info in the D50 shot reads that the camera is a D40.
    This is a silly discussion.
     
  63. Well having a D300 myself I checked the rather extensive testing on dpreview.com, and it appears that between 200-800 iso niether the D50, or D300 out shines the other by any significant amount. Each being slightly better than the other depending on iso, which is about what one would expect. outside of this range the D300 is clearly in charge.
     
  64. I happened to be in a Best Buy today, so here are some D90 vs D40 shots (quite similar to D300 vs D50, given the sensor similarities). Optimizations, d-lighting, noise reduction and any other extra in-camera processing was turned off on both cameras. I used JPEG fine/large on both cameras (sorry, I have no software that recognizes D90 NEFs).

    I couldn't use the same lens on both cameras because the D90 was alarm-anchored to an 18-105, so I did the best I could by using exactly the same exposures, white balances and focal lengths. I also had to use 18mm for the ISO 200 shots so I could use the lens' maximum f/3.5 aperture to get the shots as shake-free as possible. I used the 18-55 II for the D40 shots.
    D40, ISO 200, 18mm, f/3.5, 1/15:
    http://img.phyrefile.com/epp_b/2009/06/06/d40_iso_200_full.jpg
    D40, ISO 1600, 35mm, f/5.6, 1/60:
    http://img.phyrefile.com/epp_b/2009/06/06/d40_iso_1600_full.jpg
    D90, ISO 200, 18mm, f/3.5, 1/15:
    http://img.phyrefile.com/epp_b/2009/06/06/d90_iso_200_full.jpg
    D90, ISO 1600, 35mm, f/5.6, 1/60:
    http://img.phyrefile.com/epp_b/2009/06/06/d90_iso_1600_full.jpg
    Just for the heck of it... D90, ISO 6400 , f/5.6, 1/160, 105mm:
    http://img.phyrefile.com/epp_b/2009/06/06/d90_iso_6400.jpg
    I think the results speak for themselves. The D90 is the clear winner, especially as the ISOs climb higher and higher. The D90 manages to keep details that the D40 mushes over. ISO 6400 is pretty amazing, I'd even consider that usable! By contrast, ISO 1600 is only just barely usable on the D40 and you can simply forget ISO 3200 altogether.
    As a side note, after playing around with a D90 more extensively today than before, I discovered that there are a huge number of useful features I was unaware of and that I now need one ;)
     
  65. The D50 controls noise very well, and is still an excellent camera-even by today's standards. For web work and A4-sized prints it does a commendable job.
    With the D90 I purchased two months ago however I can shoot at ISO 800 with no appreciable gain in noise. And ISO 1600 is also excellent. With my D50 I rarely went higher than ISO 400.
    Also, the quality of the lens used has a lot to do with image noise in my opinion.
     
  66. <p>Here some pics from my D700+300AF-S F/4, most of images are with ISO 800-1600, in a theater!!<br>
    Before the D700 i used a D50 and I never go over the ISO 400<br>


    MODERATOR NOTE: No signing posts with websites please as per photo.net policy
     
  67. To the OP.
    Your D50 / D300 experience is vastly different to mine. I wouldn't use my D50 above ISO800 unless there was absolutely no other choice and even then ISO400 was my practical day to day limit. I can happily use my D300 up to ISO1600 with no qualms at all. The higher ISO performance of my example of the D300 absolutley slaughters the higher ISO of my example of the D50.
     
  68. To tell you the truth, I hated to shoot above 400 with my D50. But I also knew I could underexpose a stop and still be fine.
    But what can I say, after sitting down and making the side by side comparison, I have to think that the reason the D300 looks so much better to most people is that the NR is more aggressive or better in camera and NX2. And I would not automatically assume, using NX2, that the same NR is applied regardless of which camera the RAW file comes from. Probably you can afford to be more aggressive with the 12MP file than you can the 6MP file, and still end up with a good print.
    If you try the same thing I did, those of you with both cameras (D40/50/70 and D300/D90), and look at the RAW files in LR, I would be very surprised if you don't come to the same conclusion as me. The comparison in NX2 is not very fair because you don't know what Nikon is doing behind the scenes. Of course if you use NX2 then maybe you prefer to compare in NX2.
    I will post my RAW files tomorrow morning in case anyone cares to have a look.
     
  69. Newer technology wins out. They've had a few years to improve the sensors between the D50 sensor and the D300/D90 sensor, and they're better at denoising as well. Noise reduction doesn't just make a mushy image anymore, now you can shoot at 3200 and retain detail and texture.
     
  70. Noise reduction doesn't just make a mushy image anymore, now you can shoot at 3200 and retain detail and texture.​
    That's partly what I was getting at. Notice that the D90 still retains details better at ISO 6400 than the D40 does at ISO 1600. I did have all the NR features switched off, but it could be like a Mercedes where turning the traction control off doesn't actually turn it completely off ;)
     
  71. Here is a link to a zip file with the two RAW files, D40 and D300. These are the same files that the above JPEGs were converted from.
    http://www.robertpiontek.com/d40-d300-ISO-1600.zip
     
  72. I have to think that the reason the D300 looks so much better to most people is that the NR is more aggressive or better in camera and NX2.​
    It's not so much the camera and NX2 that do a good job, in my view. It's Adobe that does a pretty bad job with high ISO D300 files. Before I got NX2, I tried several RAW converters, including Capture One 4.x, Bibble Lite 4.x and some freeware options. Adobe in my view made the worst of my high ISO files, it was the only one to show colour noise at ISO1600. I don't think it's a matter of the noise reduction, but rather how the files are interpreted, which seems better in the non-Adobe products.
     
  73. I have both and have to agree - the D50 does a hell of a nice job at at high ISO, and, I believe, the D50 AF may be a little more reliable than the D300 at focusing on the closest subject.
    00TbC3-142219584.jpg
     
  74. I have both cameras. View photos from a 6 mpx camera--especially on a screen, can make the images look sharper than those from a 12 mpx sensor. But if you enlarge both raw images to even A4 size and print them on a good printer, the 12 mpx image is superior. No one here has offered files from both cameras with absolutely identical settings. In addition, some comparisons were used using Lightroom, which does not use all the D300 image info when processing.
    Let me pick my settings and my software and let me use them however I want, and I can make a good case that an image from my Canon A80 4 mpx P&S is as good as the image from a D50.
    This thread is a waste of good bandwidth.
     
  75. that's why i still use my D70s professionally. i don't look good to some pros and clients but i make money just the same. and yes, even with the 18-70mm kit lens. and i just use picasa 3. sometimes even just the microsoft picture viewer. funny, huh?
     
  76. Ramon,
    In the arms of the master, even the most innocent instrument becomes a weapon of the choice. :)
     
  77. I just downloaded the demo of DxO and had a look at the raw files posted above (by me). You could do the same. I left everything on default, except turned off the "DxO Lighting" feature, which I think is similar to Nikon's ADR. I installed the modules for the D40/D300 and 17-55.
    From simply looking at noise, I'd say it's a wash at 1600. Trying to compare sharpness and resolution is difficult. At the default settings, you can see that the resolution of the D300 is better, but it also looks softer. I think in large prints there would not be much difference between the two. Probably the D300 would win, but not by much, and only if you look close.
    So, after comparing 3 different raw converters, I just don't see it. In terms of image quality, the D300 has a slight resolution advantage over the D40, but that's it.
     
  78. Just my two cents. I just bought a used D300 with 17K actuations on the shutter. The camera is in mint condition and feels great in the hand. As for the noise issue, my "little" D50 is better in noise control at all iso. This is just my experience using my two eyes to view the images on screen. No scientific BS, just plain ol' carefully reviewing the images on screen. I don't know why or care, just stating the facts with my two cameras...D50 and D300. Keep shooting...Peace. BTW, my D50 has 8K actuations on the shutter.
     

Share This Page