Base ISO

Discussion in 'Nikon' started by tonkee, Sep 18, 2010.

  1. The new D7000 has the base ISO at 100, while the previous models D90 has it at 200 and D80 at 100.
    What is the significance of changing the base ISO back and forth?
     
  2. I think only who designed the sensor knows the answer. But I can guess from D80 to D90 they improved the base ISO to 200 but when they try to squeeze 16mp to the D7000 they could n't keep the base iso 200 anymore, have to go back to 100
     
  3. There are times when a lower ISO is useful to keep the shutter speed down. Some examples would include shooting moving water to make it blurry to suggest motion (needs longer shutter times), flash photography where the ISO + aperture selected might cause the shutter to exceed the flash sync speed and possibly spoil the shot, and needing to use a wide aperture for shallow depth of field in bright conditions, where a fast lens with a very wide aperture might exceed the maximum shutter speed available on the camera if the ISO is too high. There are work-arounds for these conditions (filters, gels on the lights, etc, but it's simpler and cheaper to just lower the ISO instead (if you can).
    Even with the tech advancements, one thing that hasn't changed is that image quality (color, tonality, saturation, contrast, low digital noise, etc. is at it's best at a lowest native ISO the camera has. Increased resolution may have something to do with wanting to try to keep the D7000's ISO low to avoid digital noise.
     
  4. What is the significance of changing the base ISO back and forth?​
    None; other than sensor design parameters and software interpretation down stream from the sensor.
    It is my understanding the D-7000 has a Nikon designed sensor, where in the past Sony sensors were used.
    There is a arbitrary relationship between exposure and sensor sensitivity as set by the signal gain of the sensor. This is determined by design.
    To say the D-7000 @ Native ISO 100 is "cleaner" than the D-300 @ Native ISO 200 is not a mathematical certainty.
    Because sensor data is interpreted by camera software, conclusions concerning overall sensor sensitivity are difficult to draw if we try to compare one sensor with another.
     
  5. Ray House

    Ray House Ray House

    I don't consider moving the base ISO from 100 to 200 an improvement.
     
  6. I don't consider moving the base ISO from 100 to 200 an improvement.​
    Neither do I. I would think moving up the base ISO diminishes the capabilities of a camera. I would rather go to ISO 25 than using ND filters.
     
  7. I personally like base ISO 100, but it depends on the photographer. When base ISO is 200, and you use Lo 1 or such, the camera is actually overexposing at ISO 200 and then adjusting levels down.
    ISO 100 can be useful for the following:
    When you shoot in bright sunlight, and use a fill flash, your flash sync speed is usually limited to 1/200s, which is overexposure conditions at wide apertures. This is a frequent problem when you use fill flash and you want to have shallow depth of field with a wide aperture prime. You often have to use an ND filter in this situation anyways, but being pushed to ISO 200 just makes twice as difficult.
    Secondly, if you wanted to do a long exposure, you can accomplish by using the slowest ISO and/or a smaller aperture. When you want to do a long exposure in bright sunlight, you often have to do *both*, and probably use an ND filter too. ISO 200 base makes that harder.
    In either case, you could solve the problem by using a darker ND filter, but they are more expensive, more difficult to manufacture with quality, and somewhat more difficult to use (since they darken what you see in the viewfinder).
     
  8. It is my understanding the D-7000 has a Nikon designed sensor, where in the past Sony sensors were used.​
    Maybe, but keep in mind that Sony announced a 16.2 MP APS sensor camera about a month ago. (SLT-A55) That sounds suspiciously familiar to the D7000's sensor.
     
  9. The D90 default settings is ISO 200 - 3200 but can be increased to ISO 100 - 6400. I stop below and 1 stop above the default settings (Lo 1 and Hi 1). I notice that the D7000 has standard setting of ISO 100 - 6400 and can then be extended with Hi 1 to ISO 12,800 and Hi 2 to 25,600. Both can be extended by two amounts. the D90 one above and one below the so called standard range and the D7000 has both adjustments above the so called standard range.
    Therfore it could be argued that the base ISO for the D90 is also ISO 100. For what it is worth, I own 2 D90's and always have it set on ISO 100 for studio shots.
    Cheers
    John
     
  10. "Maybe, but keep in mind that Sony announced a 16.2 MP APS sensor camera about a month ago. (SLT-A55) That sounds suspiciously familiar to the D7000's sensor."
    curiously enough, base ISO on the sony NEX-5 -- which uses sony's 16mp sensor -- is 200. perhaps they are not as "suspiciously familiar" as supposed...
     
  11. "To say the D-7000 ..."cleaner" than the D-300 @ Native ISO 200 is not a mathematical certainty." True, but it is likely.
    A body with native ISO of 100 will give better results than a similar camera with a base ISO of 200 shot at ISO 100 (L1).
     
  12. ShunCheung

    ShunCheung Administrator

    Therfore it could be argued that the base ISO for the D90 is also ISO 100.​
    No. The base ISO for the sensor used on the D90 is 200, where you are supposed to get the best results from that sensor. The D5000 and the D300/D300S use a similar sensor, all from Sony. Lo 1 is merely an extended range and so is Hi 1. When you venture into the extended range, you lose quality. However, the quality degradation at Lo 1 is subtle; you typically lose a bit of dynamic range in the highlights, which is not always easy to observe.
    Since the D3100 announced last month is using a Nikon sensor, when photo.net received advanced information from Nikon about the D7000, I verified with them that the D7000's sensor is a Nikon design.
    • The Sony A55's sensor has 4912 x 3264 pixels and has an ISO range from 100 to 12800.
    • The Nikon D7000's sensor has 4928 x 3264 pixels and has an ISO range from 100 to 6400. Note that the Nikon sensor is wider by 16 pixels.
    The D3, D700, and D3S have been using Nikon sensors since 2007, so it looks like Nikon is phasing out Sony sensors now.
     
  13. The D3, D700, and D3S have been using Nikon sensors since 2007, so it looks like Nikon is phasing out Sony sensors now.​
    They really had to, since Sony sells DSLRs too now.
     
  14. the camera is actually overexposing at ISO 200 and then adjusting levels down.​
    Not really as this assumes a custom internal curve is applied; it is not.
    Signal gain is reduced which narrows the (S)+s/n ratio, thereby bringing the amplified signal closer to the noise floor; hence a slight reduction in dynamic range, more so in the higher luminosity range in a non-linear fashion.
    The result is "some" clipping but with the benefit of reduced noise.
    As far as "quality" in the native ISO; this is a subjective call depending on your definition.
    I find in (low .3), the overall (Noise) is greatly reduced when shooting evenly lit subjects or as another example, large expanses of blue sky. I shoot in (Low .3) more often than not where in my opinion the clipping of high luminosity values is a small price to pay for the visually obvious reduction in noise.
     
  15. ShunCheung

    ShunCheung Administrator

    BTW, the Nikon D3100's 14MP sensor is different from the Sony 14MP sensor in the following manners:
    • Sony NEX 3 and 5: 4592 x 3056 pixels, ISO 200 to 12800 (not entirely sure about the ISO range)
    • Sony A33: 4592 x 3056 pixels, ISO 100 to 12800
    • Nikon D3100: 4608 x 3072, ISO 100 to 3200
    So most likely Sony's mirrorless NEX cameras are using the same sensor design as the A33 SLR, but the Nikon D3100 sensor is clearly different.
    Back in the 1990's, I used to work for AT&T. At that time AT&T's main business was long-distance telephone and other communication services, but AT&T also had a communiation equipment division that was selling hardware to AT&T's competitors such as Sprint, MCI, and other so called "Baby Bells" (now Verizon and also SBC, which in turn bought AT&T later on and renamed the combined company back to AT&T). However, those competitors were wondering why they should buy equipment from AT&T to help their competition. That eventually led to the AT&T-Lucent spin off in 1996 so that the equipment division became a separate company to avoid that conflict.
    When Sony bought Konica-Minolta a few years ago and started selling DSLRs, they put themselves in the same awkward position where major customers for their sensors (mainly Nikon and Pentax at the time) are also competitors. It should surprise no one that Nikon is finally pulling away from Sony as a sensor supplier.
     
  16. I found Philip Tam's comment quite intriquing. I have always wondered why ISO "L" on my 1Ds is a custom function. Does this mean that the "L" ISO setting, activated using the custom function, it isn't ISO 50 in the same way that ISO 100 is ISO 100? If that is the case, is there an advantage, in terms of image quality, to using ISO 100 and an ND filter rather than using "L"?
    Sorry if this is off topic. Maybe I should start a new thread....
     
    • Sony NEX 3 and 5: 4592 x 3056 pixels, ISO 200 to 12800 (not entirely sure about the ISO range)
    • Sony A33: 4592 x 3056 pixels, ISO 100 to 12800
    • Nikon D3100: 4608 x 3072, ISO 100 to 3200
    So most likely Sony's mirrorless NEX cameras are using the same sensor design as the A33 SLR, but the Nikon D3100 sensor is clearly different.​
    With the same RAW file from the same sensor, different conversion algorithms may give images of quite different resolutions. The number of pixels specified in the camera user's manual is not the number of pixels on the sensor (these pixels are only R, G, or B not a full color pixel specified in the manual). So your observation only shows that Sony's algorithm is used for many of their cameras (why rewrite it?) and not exactly the same as Nikon's algorithm. And the difference may just be some personal difference between programmers
     
  17. I am very pleased that Nikon is producing digital cameras with astronomical ISO settings. However, since I prefer images with high detail, low noise, and a wide dynamic range, I tend to shoot at lower ISO settings. I am much more likely to shoot at ISO 25 (film) than ISO 2500.
     
    When I shoot a solar eclipse or a transit of Venus across the Sun, I use the lowest ISO setting available. I would much rather have an ISO lower than the 200 minimum on some cameras.
     
    When I shoot weddings, still lifes, and studio portraits with 35mm and 120 film, I routinely use ISO 100. Since I had standardized my lighting and exposures on ISO 100, when I went shopping for a digital camera, I did not even consider a camera unless it had an ISO of 100 or lower.

    .
     
  18. ShunCheung

    ShunCheung Administrator

    John, those are dimensions for the RAW files straight from those respective cameras and sensors. By definition, there is no conversion and algorithms applied to RAW files. That is why they are called raw.
    OTOH, the Sony A900 and Nikon D3X have identical raw dimension sizes (6048 x 4032). That is why people think Nikon is using a variation of Sony's 24MP FX sensor on the D3X. In the D3 family of cameras, the D3X is the only body that is not using a Nikon sensor.
    Nikon has explicitly pointed out that the D3 sensor is a Nikon design but they don't disclose who manufactures it for them. The D700 uses the same sensor and the D3S uses an improved version. Nikon also points out that the D3100 and D7000 use Nikon sensors.
     
  19. Shun, I meant the "colour filter array (Bayer) conversion" which created the image with such and such resolution. Sony and Nikon may (very likely) program this Bayer conversion a little difference. That's all. Take the Fuji Super CCD as a good example of what is the difference between the pixels on the sensor and the pixels on the resulting image
    You cannot see the image if it is not thru the Bayer conversion yet. The numbers you gave are definitely not the number of pixels (R,G,B) on the sensor
     
  20. Nikon has explicitly pointed out that the D3 sensor is a Nikon design but they don't disclose who manufactures it for them​
    Can we safely assume that Nikon is not "who manufactures" it? Well, then we may start guessing who can that be: Canon? Kodak? Foveon? Samsung? Matshusita? any others?
     
  21. ShunCheung

    ShunCheung Administrator

    John, again, Nikon says they design the sensor themselves but they do not disclose who the manufacturer is. As far as I know Nikon has no IC fabrication facility themselves, although Nikon also happens to be one of the largest semiconductor stepper producer, so they are extremely familiar with the IC manufacturing process. Potentially, even Sony could produce the sensors based on Nikon's design and specifications, so could a few other fabs. But to me, that is irrelevant. You can start guessing who the manufacturer is, but since Nikon does not disclose the answer, you cannot confirm who is right and who is wrong, so I think that is a pointless exercise.
     
  22. I think that is a pointless exercise.​
    I believe Shun is the first to point our attention to this pointless point in this thread. But maybe it is related to the fact that the two cameras have the same specified resolution in the manuals.
    My point is only that even with different resolution specified in the manuals, the cameras may have the same sensors. And as the camera companies didn't disclose it, we can not confirm that they are the same or different
     
  23. As long as the print on my wall looks great I really don't care about the "specs" of the sensor or who makes it.
     
  24. As long as the print on my wall looks great I really don't care about the "specs" of the sensor or who makes it.​
    I learned to think that way when I was in my mid twenties and everybody was bragging about the processor speed they had on their laptops while the operating system became the big equalizer that slowed everything down. Today's processors are commodities and few people care/know what's inside their portables and/or smartphones. Most of the time it doesn't make any difference as long as your facebook status is up to date.
    I believe DSLR sensors are still a few years away from that point but the focus shift has begun as evidenced by the proliferation of HD video capabilities. Image sensors will become ubiquitous and who knows what new killer features we'll be discussing in the next years.
     
  25. Not really as this assumes a custom internal curve is applied; it is not.
    Signal gain is reduced which narrows the (S)+s/n ratio, thereby bringing the amplified signal closer to the noise floor; hence a slight reduction in dynamic range, more so in the higher luminosity range in a non-linear fashion.​
    Well, I may be wrong, but this is what I've read (and experienced) about the D5000/D90. The recent Nikon cameras achieve higher ISO by applying signal gain before the digital to analog conversion. There is however, no signal gain setting to reach the lo or hi settings. Instead, the camera uses the closest capable ISO (for example, ISO 200 for lo 1, or ISO 3200 for hi 1), does the analog to digital conversion, then digitally boosts the file (or cuts it), resulting in the final picture.
    This is why on testing websites, like DXO mark, when they test ISO 200 and lo 1 on the D5000 and D90, etc., the actual measured ISO is exactly the same.
    This is the explanation I've read. Whether or not it's correct, I cannot confirm. I'm a Mechanical Engineer, not an Electrical one.
    I found Philip Tam's comment quite intriquing. I have always wondered why ISO "L" on my 1Ds is a custom function. Does this mean that the "L" ISO setting, activated using the custom function, it isn't ISO 50 in the same way that ISO 100 is ISO 100? If that is the case, is there an advantage, in terms of image quality, to using ISO 100 and an ND filter rather than using "L"?
    Sorry if this is off topic. Maybe I should start a new thread....​
    The short answer is yes, you would do better using an ND filter... but only slightly better. I'm not sure if Canon is different, but at least in Nikon's case, you lose dynamic range (most particularly do to highlight clipping). Of course though, the difference is small enough that when I'm not in a high contrast situation or when I don't care that much, I use lo settings rather than take the time to screw on setup up ring + ND filter.
    See below:
    http://www.dpreview.com/reviews/NikonD5000/page18.asp
     

Share This Page