Friday, January 11, 2019

Bigger Sensor Less Noise And More Dynamic Range? Not Necessarily

With his permission, I photographed this man at the Taos Pueblo, Taos, New Mexico (click to enlarge)
Olympus E-M1 Mark II, 12-100mm f/4 PRO lens @ 54mm; 1/3200th sec. @ f/5.6; ISO 200
If you think a bigger sensor always means less noise, that may not always be true. For example with my three main cameras, according to these charts from the website "Photons to Photos," the Fujifilm X-T2 generates the same amount of digital noise at ISO 200 as does my Olympus E-M1 Mark II does at about ISO 1100-1200.  The Olympus sensor is 1/2 the size of the Fujfilm sensor.  Same with the X-T3 at ISO 200.  But the X-T3 has lower noise at its new base ISO, 160, and the noise comparison shows its noise about the same as the E-M1 Mark II's at ISO 1000.  Still a remarkable difference.  I've included the noise chart below so you can see the results of Claff's work for yourself.

I stumbled across this data when I was researching another topic.  I thought I would pass it along for what it may be worth to you.  Claff is a well credentialed scientist and his site is often cited by numerous photography websites and presenters as being factual and accurate.  However, I know his data is quite different from the other highly quoted website, DXOMark.  If you have two very reputable organizations coming up with different data, that doesn't necessarily mean one is wrong.  It may be because of different testing methods or other variables.  The answer is for you to test your own gear and see what YOU find.

Obviously, the laws of physics would back up the premise that the larger sensor has less noise but we have much more going on here besides just sensor size.  We also have to take into consideration the camera manufacturer's skill at creating a processing engine that maximizes image quality despite sensor size.  I think that may be some of what you are seeing here.

X-T2 in black, X-T3 in lime green and E-M1 Mark II in light blue.
As for Dynamic Range, look at the chart below.  Both Fujifilm cameras as well as the Olympus camera has very similar dynamic range.  Again, that is a surprise to me.  As above, it has been pounded into our heads that larger sensors have less noise and more dynamic range.  Obviously, we have more going on here than just sensor size as well.

Also, if you think it is okay to always photograph at a higher ISO than base ISO, look at how much dynamic range is lost as you raise the ISO.  For example, a stop and a half of dynamic range is lost between ISO 200 and ISO 800.  Here is a chart showing the dynamic range versus ISO for my three main cameras.

X-T2 in light blue, X-T3 in black and E-M1 Mark II in lime green.


I'm sure Claff's work is spot on.  After all he has advanced degrees in applied mathematics and teaches at Harvard.  However, instead on taking Claff's (or DXO's) tests as the last word on this, I always test for myself with my particular pieces of gear under circumstances I can control.  In the near future, I plan on testing my own cameras just to see what I find and what the "practical" difference is in noise generation by these two formats and camera brands, one being about double the size of the other.

One last thought.  We all are the beneficiaries of superb digital image editing programs.  I'll call them "The Great Equalizers."  Even if we get a variation in digital noise from different cameras, different sized sensors, different ISOs, the editing programs can largely, but not totally, equalize the playing field.  For most of us, we can use the tools in the editing programs to reduce or remove digital noise, adjust sharpness, etc.  There may not be a need to buy a full frame camera if you want low noise and good dynamic range.  These editing programs can probably produce the images you want without have to spend a lot more money on full frame gear.  And...it will only get better in the future.

In addition, another seemingly wonderful tool has been developed and released for our use.  In 2018, Topaz Labs created and released a program called AI Gigapixel.  Their new software uses artificial intelligence to give you the ability to upscale your images, they say, as much as 600% without losing image quality.   According to their literature, this upscaling is done in a completely novel way so as to increase detail and sharpness at the same time.  The downside is that it takes some very powerful computer hardware to utilize.  In fact, they tell you what type of computing power you need and what graphics processors will not work.  I'll be following this new development in the future.

As I have said in the past, we are all lucky to be photographing today and it is all good!

Join me over at Instagram @dennisamook or my website, www.dennismook.com

Thanks for looking. Enjoy! 

Dennis A. Mook 

All content on this blog is © 2013-2019 Dennis A. Mook. All Rights Reserved. Feel free to point to this blog from your website with full attribution. Permission may be granted for commercial use. Please contact Mr. Mook to discuss permission to reproduce the blog posts and/or images.

8 comments:

  1. Hi Dennis
    Very interesting - thank you, I wonder if you can help me - do you have a clear idea of what the “Low” setting on Olympus actually achieves ?
    All the Best
    Mike

    ReplyDelete
    Replies
    1. Michael, glad to help. However, I’m not sure which setting is the one to which you are referring? Can you be more specific?

      Delete
  2. Sure, sorry should have been clearer - was referring to the ISO settings, there is Auto- Low- 200 etc.
    I was wondering what it achieved and if it had any negatives - “heard” it affect Dynamic range but from your charts it doesn’t seem to

    ReplyDelete
    Replies
    1. The lower than ”base” ISOs are handy when your subject requires a slower shutter speed/aperture combination than is available. It may be just enough of a difference so you don’t have to attach a neutral density filter. I would suspect that Low is not often required. Also, I think it has advantages for the video people. They are restricted in which shutter speeds they can use so the video “looks” as it should and the ways to get that proper exposure might require a lower ISO even with the aperture wide open. For example, and I’m not a video person, a video shot at 60fps needs a shutter speed of 1/60th second otherwise the video will look choppy, so I’m told. You see how that could hamper proper exposure when outdoors.

      In the past, and I believe still, with certain cameras, setting the ISO below the “base” can reduce the dynamic range and increase digital noise. I think one would have to check each model to be sure if any particular camera is affected in that manner. On the other hand, I had read once that lower than base ISO on the E-M1Mark II actually improved detail. In my tests I didn’t see any noticeable difference than when my camera was set at ISO 200. I try not to take anything at face value if I can test it for myself.

      Thanks again for the question. Happy to help.

      Delete
  3. Dennis,

    Just a quick thought about the photons-to-photos charts - the dynamic range charts are suitable for comparing camera models, but there is a note on the noise chart which reads 'These raw values are not appropriate for comparing camera models because they are not adjusted for gain or area'. I believe this means it's not meaningful to compare noise across sensor size by using that particular chart? If so, that would fit my personal experience of these cameras - dynamic range is similar, but the m4/3 sensor is considerably more noisy even at base ISO than the Fuji sensor.

    All the best.

    ReplyDelete
    Replies
    1. James, thank you. I missed that note. I was looking at the charts on my iPad and didn't scroll down far enough to see the comments. I want to better understand what Claff means by his statement so I'm going to email him and ask for clarification. When I get a response I will add to the post. Thanks again and hope you are doing really well and have been very busy!

      Delete
    2. Hi Dennis, I'm well thanks, and hope you are too! Yes, busy, but all for the good - still greatly enjoying your blog and photos when I have a few moments. It would be great to get clarification on Claff's note - photonstophotos is an excellent site and it would be nice if we were able to compare noise across different sensors. I should also clarify, in my above comment the personal experience I was referring to was with a previous generation of sensor - the X-T1 and E-M1. The X-T2 is my daily camera now, but I never owned an E-M1 Mark II, and having done some research after reading your article, it seems Olympus have indeed closed the 'noise gap' quite dramatically. If so, this is a great thing - I still have a great deal of respect for the m4/3 system, which has superb IBIS and some distinct system size advantages in terms of lenses. All the best.

      Delete