How to properly compare RX-100 and NEX-6 kit high ISO noise?

Started Apr 29, 2013 | Discussions thread
2eyesee
Senior MemberPosts: 1,676
Like?
Re: How to properly compare RX-100 and NEX-6 kit high ISO noise?
In reply to CosmoZooo, May 11, 2013

CosmoZooo wrote:

The individual photosite size is not so relevant on modern sensors with gapless pixels - it's the total amount of light gathered by the lens that's the most important factor in determining noise levels/DR. I'm no expert but I have been corrected on this in another thread and upon further investigation it seems to be true for modern sensors.

Even though I will admit to not knowing much on the subject of gapless pixels I would cautiously disagree that individual photosite size is not relevant.

I've seen this mentioned in multiples forums that it very much does and even in dpreview articles - the tiny photosites on the smaller sensor is to my understanding the main reason they do not perform the same way as the bigger sensors. It has also been argued that it's the main reason NEX-7 doesn't have the same ISO performance as the NEX-5n/5R/6.

It's true that the NEX-7 doesn't have the same high ISO performance as an NEX-6 when you view at 100%. But If you reduce the 24mp NEX-7 image to the same size as the 16mp NEX-6 image the resulting NEX-7 image is no worse then the NEX-6 image. In fact it looks a bit better as the noise has a finer grain due to the higher resolution of the NEX-7 - I did this exercise myself with a couple of ISO6400 shots over at the Imaging Resource Comparometer. And of course the NEX-7 also enjoys a resolution advantage at low ISO.

So the theory is with modern sensors more megapixels is always better.

I have also seen others say as they did in the link to the post I provided earlier that a total amount of light is not relevant - that is essentially a crop issues...your field of view is very different...you see less in your frame. Think about it...just like looking through a big pipe/straw/pinhole or a little one...you see less area - but the smaller area is not less detailed...do you understand my logic here?

So I disagree with that total light hitting the sensor matters - that simply means there is more stuff in your field of view...it is the amount of light per photosite and how well that photosite performs that will decide the overall quality of the sum of all those photosites. Bigger photosites have more light land on them.

If you can explain how more overall light would impact the image quality I would like to hear it.

Here's is what I assert hypothetically:

If you took the RX100 lens at f1.8 and mounted it on the NEX-5R and took a picture you would only see a small crop of the sensor - the resulting image would be black all around except some area in the middle.

Say then you crop that area into a clean rectangle. You then take a 1.8 NEX lens take the same photo and crop out the same exact rectangle from the center of your image.

I assert that assuming both lenses are equally sharp and perform similarly on their respective systems the quality of those rectangles would be nearly identical. I can not think of any reason why that would not be the case.

I agree, but I'm not sure what this proves as you would just have a small crop of the NEX image.

Think about how legacy FF lenses work on APS-C. They do not produce lower quality images and their speed doesn't change...but you're cropping the image...there is less overall light hitting the sensor but not per photosite - that stays the same assuming the photosite on your FF and APS-C sensors are the same size. You loose bokeh and field of view in the process but not quality - not if the two sensors perform similarly.

But an APS-C sensor does produce lower-quality images than a full-frame sensor.

Now think about how the speedbooster works - it condenses the light of FF into an APS-C area actually increasing the amount of light hitting each photosite and making the lens faster. The overall amount of light coming through the lens is still the same but it is being concentrated into a smaller area - of course the difficult part here is not to distort the image or introduce another issue so you can only condense things so much...but you see my point here.

I say total light doesn't matter - light per photosite does matter and the performance of each photosite does matter.

I'll let others chime in if they disagree...

I too will welcome any comments from others. I'm certainly no expert - the only reason I'm on these forums is to learn, so I'd appreciate any contributions from others to help improve my understanding.

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow