1. Welcome to Mu-43.com—a friendly Micro 4/3 camera & photography discussion forum!

    If you are thinking of buying a camera or need help with your photos, you will find our forum members full of advice! Click here to join for free!

Equivalence thread (including posts moved here from other threads)

Discussion in 'Back Room' started by dalto, Sep 6, 2016.

  1. DanS

    DanS Mu-43 All-Pro

    Mar 8, 2016
    Central IL
    That was his way of justifying this statement, which is straight up wrong.

     
  2. Turbofrog

    Turbofrog Mu-43 Legend

    Mar 21, 2014
    I'm sorry, it's not wrong.

    Assuming the same reproduction size, the total area is much more important than photosite size. This has been demonstrated again, and again, and again. The 42MP A7r II has better dynamic range and signal-to-noise ratio than the 12MP A7s until ISO 12800. It takes until 51200 before you can say "yeah, that 12MP sensor looks better" despite having photosites that are more than 3 times larger (even taking into account the BSI vs. FSI architectures).

    I wonder if the confusing part in all of this is that people don't seem to take into account both sides of the signal to noise equation which we take as a useful proxy for image quality.

    If you have bigger photosites, you certainly have less noise. But you also have less signal. Those additional photosites are capturing chroma and luminance information that enhances the image. If your image was made up of just one pixel, you'd have no noise. But also no signal.

    Once you get to the point where it's mostly garbage information (say, ISO >12,800) the pixel size starts to matter because those additional photosites are no longer large enough discriminate usefully. But until that point, the SNR is pretty much invariant with photosite size.
     
    Last edited: Sep 8, 2016
    • Agree Agree x 1
    • Informative Informative x 1
  3. bigboysdad

    bigboysdad Mu-43 All-Pro

    Aug 25, 2013
    Sydney/ London
    What a salty thread this is!
     
  4. DanS

    DanS Mu-43 All-Pro

    Mar 8, 2016
    Central IL
    I'm sorry, but it is. The final print argument only works when the final image size (print if you prefer) is smaller than the actual image size.
     
  5. Turbofrog

    Turbofrog Mu-43 Legend

    Mar 21, 2014
    So you'd prefer to stand at the same distance and compare the image quality of a 10x15" print and a 24x36" print of the same photo and call that useful? Whatever you like, I guess.
     
    • Like Like x 1
  6. dalto

    dalto Mu-43 Veteran

    302
    Jul 6, 2016
    Texas
    Sorry that I am late to the debate and I may be rehashing something that has been gone over many times in the past but I am struggling with understanding the practical significance of "per pixel noise".

    When I look at a photo in post I am looking at the impact that noise has the entire frame of the photo. If I take the same photo with two bodies and one has twice as many pixels as the other I don't blow them out to 1:1 and evaluate the noise. Even if I am cropping both images I don't base my crop on the number of pixels in the image. I base it on what I want the image to look like.

    Maybe I am misunderstanding but it seems like we are getting more caught up in the technical measure of the amount of noise instead of the real world impact it has on a photo.
     
    • Agree Agree x 2
  7. ijm5012

    ijm5012 Mu-43 Legend Subscribing Member

    Oct 2, 2013
    Pittsburgh, PA
    Ian
    Well, this thread has denigrated to discussions using typical equivalence buzzwords like "photosites" and "photons". How foolish of me to think I could create a post commenting on how large the O25 PRO is compared to a 50mm FF lens, and expect a somewhat civil discussion about m43 vs FF without it getting to the point that it has.

    Sigh, and I thought people here were better than over on DPR (though I will admit, we haven't gotten to the point of name calling of verbal insults, so that's a plus).
     
  8. Nathanael

    Nathanael Mu-43 Veteran

    462
    Oct 12, 2015
    Well if the two options are either "blissfully unaware" or "argue for endless hours about equivalence in online photography forums" then yea.. I would say most people would opt for the former. ;) 
     
    • Like Like x 1
    • Agree Agree x 1
    • Funny Funny x 1
  9. Jonathan F/2

    Jonathan F/2 Mu-43 Hall of Famer

    Aug 10, 2011
    Los Angeles, USA
    The moment you wrote "FF" in this thread it was doomed from the get-go!

    I have no doubt the Oly 25mm 1.2 Pro will be awesome, I think the important question to ask if it will be cost effective versus similar lenses of bigger sensor formats? If one is looking for low light performance, I think the Olympus 25mm will be the one to get (IBIS, fast aperture and superior micro-contrast is M43's forte). If you're looking for shallow DOF performance and better dynamic range, a cheap used A7 and the 50mm 1.8 FE will easily do the job.
     
    • Agree Agree x 1
  10. DanS

    DanS Mu-43 All-Pro

    Mar 8, 2016
    Central IL
    that's not what i'm talking about. print a 16mp image out as a 10x15", then take the exact same image and crop it for any number of reasons and print it out at 10x15". The cropped image will look worse, because you couldn't scale the noise away as much as you could with the 16mp image. A good example of this would be taking a full body portrait, and then having to pull a head shot from it after the fact.

    That is why the "total light" argument is flawed, it assumes that you always have more MP than you need, and thus can scale away issues with the image.


    it's not about equivalence, i could care less about equivalence, its about understanding how your gear works.
     
  11. Turbofrog

    Turbofrog Mu-43 Legend

    Mar 21, 2014
    So the assumption is that everyone with FF gear is just shooting all the time with ultrawides and then cropping out the center, so that they get the same image quality as we do with M4/3? I'm trying to spin this argument around in my head to find a way that it makes sense and isn't just specifically cherry-picked mental gymnastics.

    It's clear that we all understand the concepts involved, we just see the practical aspects of them very, very differently.
     
    Last edited: Sep 8, 2016
    • Agree Agree x 1
  12. DanS

    DanS Mu-43 All-Pro

    Mar 8, 2016
    Central IL
    As far as I'm concerned the conversation at this point is not about one system vs the other, it's about understanding how our gear works. Imo, when you understand how your gear actually works, it makes you better, because you can get more out of it, because you are aware of it's strengths and weaknesses.
     
  13. fredlong

    fredlong Just this guy...

    Apr 18, 2011
    Massachusetts USA
    Fred
    My gear works pretty good ;) 
    Fred
     
    • Agree Agree x 1
  14. pellicle

    pellicle Mu-43 Hall of Famer Subscribing Member

    Feb 10, 2010
    Killarney, OzTrailEYa
    pellicle
    I am beginning to think that I'm wrong here, not in statement, but in reason.

    Increasing pixel density must reduce pixel area. Thus the major effect of reduced pixel area is increased signal to noise ratio. I have found (in reading more carefully other sources that) this to be much more significant than the probability of photo strike reduction.

    So it remains that smaller areas generally result in higher signal to noise ratios.

    So Freds example of taking a slice out of a FullFrame is still correct in that no more noise is apparent. However the points on magnification will essentally cause the image to be degraded in other ways (outright pixel count for a start). I'm not entirely sure about noise.

    This is a difficult subject to be sure, and in my view discussion where we attempt to understand and learn. (not just assert our beliefs and "prove ourselves right")
     
  15. Klorenzo

    Klorenzo Mu-43 All-Pro

    Mar 10, 2014
    Lorenzo
    A little late to the party but the big total light/per-pixel light, etc. debate started to make sense to me only when I started to considered the subjects that appear in the picture too.
    Because it is true that a FF lens gathers more light, but it also has to lit more surface. And if we assume comparable technology there are not many ways to get out of this fact.

    But, in a portrait for example, if you consider what is actually projected on the sensor you'll see that the face of the subject covers a certain absolute physical surface (e.g. 2mm sq.). In the bigger sensor system that same face is recorded by a surface four times that size.

    Each individual surface unit receives the same amount of light in both system, but one system uses more surface to record the same features. This is what I got, eventually and for me everything else falls into place: crop, apertures, total light, etc.

    Pixel density, sensitivity, BSI, etc. are relevant but just muddy the waters and would not account alone for the big noise difference we see in practice.
     
    • Agree Agree x 1
  16. Lunatique

    Lunatique Mu-43 Regular

    133
    Sep 14, 2014
    Lincoln, CA
    For me, I'm not going to be interested in any new lenses until they release ones that can achieve at least the equivalent DOF of full-frame f/1.8, with autofocus, and ideally in 50mm and 85mm full-frame equivalent focal lengths.
     
  17. Amin Sabet

    Amin Sabet Administrator

    Apr 10, 2009
    Boston, MA (USA)
    I've been on 4/3 and Micro 4/3 forums since the beginning of the format. Every "equivalence" discussion I have seen has turned into :026:.


    When it takes over other threads, I will try to move it to this one.


    Sometimes it is a tough call which stuff gets moved and which doesn't. I'll do the best I can.


    Addendum, please also see this thread: Posts moved to the equivalence bashing
     
    Last edited: Sep 29, 2016
    • Like Like x 1
  18. danelkins

    danelkins Mu-43 Top Veteran Subscribing Member

    664
    Mar 9, 2013
    Norhtern Illinois
  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.