Open question to DPR- Scene Comparison Tool and high ISO samples

Started Feb 23, 2013 | Questions thread
ForumParentFirstPreviousNextNext unread
Flat view
il_alexk
Senior MemberPosts: 2,007Gear list
Like?
Open question to DPR- Scene Comparison Tool and high ISO samples
Feb 23, 2013

Hi,

It is no secret that there is many heated discussions related to possibility of cheating by some manufacturers with high ISO performance of their cameras. While DPR has not discovered any signs of the cheating, I still have a few concerns regarding the methodology used by DPR to obtain the samples for the DPR Scene Comparison Tool (SCT). I believe this discussions were caused by the confusion SCT caused while presenting High ISO images.

Let's consider for example how Scene Comparison Tool presents samples for Pentax K30, K5 and K5IIs cameras with ISO6400:

Scene Comparison Tool - ISO6400

While observing these images, one can easily notice a much cleaner images produced by K30 and K5II/s cameras, which could lead to some interesting conclusions about the improved low light handling in the newer cameras.

Unfortunately these conclusions are not not backed up by other reviews or published samples, including DxO and Imaging Resource. This is not surprising, since all four cameras use the same sensor (and probably almost the same electronics).

So I coulnd't help but notice that these shots are made under different lighting conditions, For example, the K5 shots are done with F9, 1/1603 while the newer cameras use F8 and 1/1250 allowing more light to reach the sensor.

It is easy to see that in practice the difference in the shutter times and apperture results in almost a full stop difference in amount of light reaching sensors of these cameras (2/3?). It is obvious, that when less light reaches the sensor this inevitably means noisier images.

This generic observation can also be confirmed by a very simple formula which quantifies the noise: SNR = Signal/Noise. Assuming the same noise, reducing the signal by half means twice amount of noise to be visible.

Please note, that I've used K5 as an example only. There are also other cameras that receive less light then others. For example, D7000 is tested in conditions similar to K5 thus making its direct comparison with other cameras to be less then trivial.

I can come up only with the two possible explanations:

  1. DPR methodology is correct, the amount of light reaching the sensor is the same. It is vendors choice to label lower ISO values as 6400. For example, an internal value of ISO=4300 could be presented as ISO6400 in K5IIs case
  2. Vendors mark the ISO in their cameras within expected accuracy (ISO6400 is indeed ISO6400), however the DPR tests cameras with inconsistent amount of light measured by sensors, thus making the direct comparison of the samples less usable.

I wonder if you can address these concerns and help me to resolve this riddle once and for all (or at least till the next time someone will ask the same ).

Thank you
Alex

 il_alexk's gear list:il_alexk's gear list
Pentax K-5 Olympus PEN E-PM1 Olympus OM-D E-M10 Panasonic Lumix G 20mm F1.7 ASPH Pentax smc FA 50mm F1.4 +7 more
ANSWER:
This question has not been answered yet.
Nikon D7000 Pentax K-30 Pentax K-5
If you believe there are incorrect tags, please send us this post using our feedback form.
ForumParentFirstPreviousNextNext unread
Flat view
Post (hide subjects)Posted by
ForumParentFirstPreviousNextNext unread
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow