This topic is in the 2013 archive. You can read it and like its posts here. It is read only.
Archive 2013NEX7 v A7r Image Quality Field Testing
p.2 #1NEX7 v A7r Image Quality Field Testing
This comparison only makes sense if you are "telephoto-limited." In the real world, you'd keep the framing the same...
Dec 23, 2013 at 05:49 PM
p.2 #2NEX7 v A7r Image Quality Field Testing
I've seen other reviewers compare the same lens on different sensor (size) cameras and I never understood the reason for it, then or now. When I'm out taking photos I stand in the place that gets me the near/foreground relationship I want, then I chose the lens that will give me the FOV that I want. I do NOT stick a longer lens on and move back x feet to recover the FOV. If you do this Max then you are an odd duck and may well have a few marks on your forehead from walking into walls. Camera bodies and lenses should be treated as a system when making these comparisons.
So by all means compare the Nex7 to the A7r, but chose the best lenses that will give you the same FOV, perhaps the 35 and 55 respectively with some MINOR shifts to give comparable FOV. And if you want to have this taken seriously ensure that both shots are perfectly focused, otherwise advertise the thread as a focus test, NOT a res test.
Mike
So by all means compare the Nex7 to the A7r, but chose the best lenses that will give you the same FOV, perhaps the 35 and 55 respectively with some MINOR shifts to give comparable FOV. And if you want to have this taken seriously ensure that both shots are perfectly focused, otherwise advertise the thread as a focus test, NOT a res test.
Mike
Dec 23, 2013 at 06:04 PM
p.2 #3NEX7 v A7r Image Quality Field Testing
Max, I can offer you some advice which can assist both more empirical thinking and discipline on your part and to increase the likelihood of you convincing others of your conclusions.
You need a null hypothesis to start with. This is statement of exactly what you are trying to show but expressed in such a way that the results obtained can be used for forming conclusions, once the methodology is regarded as a valid approach. Your NH might be here: 'the a7r is better than the NEX in reolution' for example.
Doing it this way enables you to then explain your methodology which should show what factors are important and which are 'adjusted' for in the experiment, to avoid ecological fallacy.
This methodology in turn can be assessed by any reader for its viability and applicability for proving/disproving the null hypothesis with valid results. The methodology must be fully explained and its effectiveness as evidence producing must be established to the general satisfaction of readers.
Professional warning signs are all over your work above. You begin with an exclamatory conclusion plus some poorly explained 'pseudo method' - in fact I read no further to be honest, that is how badly it was expressed.
Now I don't want to be offensive, but I do want to point out these things to help you make some sense from your efforts and encourage some discipline in your thinking in experiment design; and to stand at least a fighting chance of persuading rational viewers, rather than merely those with a similar emotional viewpoint to your own.
I can recommend some good texts on the subject if you wish to read further on these matters.
You need a null hypothesis to start with. This is statement of exactly what you are trying to show but expressed in such a way that the results obtained can be used for forming conclusions, once the methodology is regarded as a valid approach. Your NH might be here: 'the a7r is better than the NEX in reolution' for example.
Doing it this way enables you to then explain your methodology which should show what factors are important and which are 'adjusted' for in the experiment, to avoid ecological fallacy.
This methodology in turn can be assessed by any reader for its viability and applicability for proving/disproving the null hypothesis with valid results. The methodology must be fully explained and its effectiveness as evidence producing must be established to the general satisfaction of readers.
Professional warning signs are all over your work above. You begin with an exclamatory conclusion plus some poorly explained 'pseudo method' - in fact I read no further to be honest, that is how badly it was expressed.
Now I don't want to be offensive, but I do want to point out these things to help you make some sense from your efforts and encourage some discipline in your thinking in experiment design; and to stand at least a fighting chance of persuading rational viewers, rather than merely those with a similar emotional viewpoint to your own.
I can recommend some good texts on the subject if you wish to read further on these matters.
Dec 23, 2013 at 06:29 PM
p.2 #4NEX7 v A7r Image Quality Field Testing
Another one of these threads. When did this become the Max forum? Really, this is getting tedious.
Dec 23, 2013 at 07:12 PM
p.2 #5NEX7 v A7r Image Quality Field Testing
I've already decided to hold off on jumping to full frame in this round, and I'm quite happy with the NEX-7... but this "test" is meaningless. Zooming in farther with the full frame camera makes it a joke. Max, if you want this comparison to mean anything, you have to compare with comparable field of view, and comparable output size.
Dec 23, 2013 at 07:20 PM
p.2 #6NEX7 v A7r Image Quality Field Testing
Max,
What you are showing is that the NEX 7 has higher pixel density not that it has higher IQ. If you want to do this test right move your feet and use the whole frame from both cameras that way you can use the same lens and get similar framing and use the whole resolution from each camera. To me that is the much more relevant test. I am not going to use a FF camera and a crop camera from the same spot with the same lens and then crop to get the same framing. Instead I will create the same or as similar a framing as possible and crop neither shot.
What you are showing is that the NEX 7 has higher pixel density not that it has higher IQ. If you want to do this test right move your feet and use the whole frame from both cameras that way you can use the same lens and get similar framing and use the whole resolution from each camera. To me that is the much more relevant test. I am not going to use a FF camera and a crop camera from the same spot with the same lens and then crop to get the same framing. Instead I will create the same or as similar a framing as possible and crop neither shot.
Dec 23, 2013 at 07:49 PM
p.2 #7NEX7 v A7r Image Quality Field Testing
bcaslis wrote:
Another one of these threads. When did this become the Max forum? Really, this is getting tedious.
Agreed.
When did forum become dpreview? The usual paucity of these types of ridiculous threads here is why I much prefer FM over dpreview - Max - please, take this silliness there.
Another one of these threads. When did this become the Max forum? Really, this is getting tedious.
Agreed.
When did forum become dpreview? The usual paucity of these types of ridiculous threads here is why I much prefer FM over dpreview - Max - please, take this silliness there.
Dec 23, 2013 at 08:12 PM
p.2 #8NEX7 v A7r Image Quality Field Testing
The test is a valid comparison.
Furthermore, the test is about comparing IQ. Not DOF of FOV or one len's field curvature against another (there are far many more criticisms for using different lenses then using the method I did)
It's all right there on the same screen - side by side, no processing, gimmicks or tricks.
http://www.josephjamesphotography.com/equivalence/index.htm
Given four cameras, one with...
...an mFT (4/3) sensor,
...another with a 1.6x sensor,
...another with a 1.5x sensor,
...and another with a FF sensor...
...and...
...a photo of a scene from the same position with the same focal point and the same settings (e.g. 25mm f/1.4 1/200 ISO 400) with all cameras,
...the photos cropped to the same framing as the photo from the mFT (4/3) camera,
...and the photos are displayed at the same size...
...then the resulting photos will be Equivalent.
(edited to reflect the scenario I did)
Given two cameras, one with...
... a 1.5x sensor,
...and another with a FF sensor...
...and...
...a photo of a scene from the same position with the same focal point and the same settings (e.g. 50mm f/5.6 1/200 (or equal for each scene) ISO 100) with both cameras,
...the photos cropped to the same framing as the photo from the APSc camera,
...and the photos are displayed at the same size...
...then the resulting photos will be Equivalent.
Maybe Sony will fix these issues - and in 6 months I'd be willing to buy the camera again. It's a lot more intuitive than the NEX7 - even a week of using it I felt a comfort and control (mainly the EV and mode dials) that the NEX makes awkward. But right now the IQ isn't there and I am not willing to put up with a camera I can't trust.
Furthermore, the test is about comparing IQ. Not DOF of FOV or one len's field curvature against another (there are far many more criticisms for using different lenses then using the method I did)
It's all right there on the same screen - side by side, no processing, gimmicks or tricks.
http://www.josephjamesphotography.com/equivalence/index.htm
Given four cameras, one with...
...an mFT (4/3) sensor,
...another with a 1.6x sensor,
...another with a 1.5x sensor,
...and another with a FF sensor...
...and...
...a photo of a scene from the same position with the same focal point and the same settings (e.g. 25mm f/1.4 1/200 ISO 400) with all cameras,
...the photos cropped to the same framing as the photo from the mFT (4/3) camera,
...and the photos are displayed at the same size...
...then the resulting photos will be Equivalent.
(edited to reflect the scenario I did)
Given two cameras, one with...
... a 1.5x sensor,
...and another with a FF sensor...
...and...
...a photo of a scene from the same position with the same focal point and the same settings (e.g. 50mm f/5.6 1/200 (or equal for each scene) ISO 100) with both cameras,
...the photos cropped to the same framing as the photo from the APSc camera,
...and the photos are displayed at the same size...
...then the resulting photos will be Equivalent.
Maybe Sony will fix these issues - and in 6 months I'd be willing to buy the camera again. It's a lot more intuitive than the NEX7 - even a week of using it I felt a comfort and control (mainly the EV and mode dials) that the NEX makes awkward. But right now the IQ isn't there and I am not willing to put up with a camera I can't trust.
Dec 23, 2013 at 08:45 PM
p.2 #9NEX7 v A7r Image Quality Field Testing
Last but not least, if there was a skew in the test it was skewed toward the A7r. I picked the lens that scored highest on the A7r and shot all scenes at its highest Imatest score (f5.6)
The A7r would be a great B&W camera. It's resolution on B&W with the FE55 is perhaps very close to the 800e and the Otus.
Almost every issue I have with the A7r is about color and tonal transitions.
Why doesn't anyone want to discuss why the roof shingles are monotone with the A7r ?
It's clear as hell there's a difference.
Why doesn't anyone want to discuss the disturbances where the roof meets the sky ?
How does one fake that or how is it caused by the difference between an APSc and a FF sensor ?
What about the leaves ? A better sensor and yet the leaf doesn't look nearly as much like a leaf as the NEX7 example -
If anything the bigger pixels of the A7r should have done way better with that leaf.
I ran these same RAW through Capture 1 and there was no improvement. It's all in the garbage that the A7r delivers.
The A7r would be a great B&W camera. It's resolution on B&W with the FE55 is perhaps very close to the 800e and the Otus.
Almost every issue I have with the A7r is about color and tonal transitions.
Why doesn't anyone want to discuss why the roof shingles are monotone with the A7r ?
It's clear as hell there's a difference.
Why doesn't anyone want to discuss the disturbances where the roof meets the sky ?
How does one fake that or how is it caused by the difference between an APSc and a FF sensor ?
What about the leaves ? A better sensor and yet the leaf doesn't look nearly as much like a leaf as the NEX7 example -
If anything the bigger pixels of the A7r should have done way better with that leaf.
I ran these same RAW through Capture 1 and there was no improvement. It's all in the garbage that the A7r delivers.
Dec 23, 2013 at 08:57 PM
p.2 #10NEX7 v A7r Image Quality Field Testing
The irony is that if you performed the test as you describe with film (maintaining the same emulsion), you'd be correct. The images would be equivalent.
The problem is that the same is not true of digital when you are talking differing pixel densities. Your test is the same as comparing a NEX-7 and NEX-6 with the same lens and same effective magnification. You're functionally cropping the A7r image to 2/3rds the pixels of the NEX-7 image. Of course the noise is more noticeable, given roughly similar per-pixel noise the NEX-7 pixels are smaller so the noise is smaller and less obtrusive.
If you want an accurate comparison of IQ you need to shoot the same FoV between cameras, not the cropping down to the same effective sensor size (which is what you did here). Your comparison is biased wildly in favour of the camera with the higher pixel density. The reason this test worked on film is because the emulsion had the same basic characteristics in all formats (effectively the same pixel density). What you're proved here is that higher pixel density cameras resolve more detail and per-pixel noise has a smaller impact on the overall image the higher the pixel density gets.
If you wanted to do a valid test of the A7r vs an APS-C NEX body, cropped down to an APS-C FoV, you should be using a 16MP body, then you'd have something approaching an accurate comparison between the two as the A7r is a ~15MP body when cropped to APS-C (which is close enough for light testing).
If you really want a good test of IQ, I'd compare the D800E to the A7r with the same lens (say the otus you own). Then for a good comparison of what you are doing here, compare the A7r to a D610, also with the Otus (which would have the A7r in the same situation as the NEX-7 is in your test).
The problem is that the same is not true of digital when you are talking differing pixel densities. Your test is the same as comparing a NEX-7 and NEX-6 with the same lens and same effective magnification. You're functionally cropping the A7r image to 2/3rds the pixels of the NEX-7 image. Of course the noise is more noticeable, given roughly similar per-pixel noise the NEX-7 pixels are smaller so the noise is smaller and less obtrusive.
If you want an accurate comparison of IQ you need to shoot the same FoV between cameras, not the cropping down to the same effective sensor size (which is what you did here). Your comparison is biased wildly in favour of the camera with the higher pixel density. The reason this test worked on film is because the emulsion had the same basic characteristics in all formats (effectively the same pixel density). What you're proved here is that higher pixel density cameras resolve more detail and per-pixel noise has a smaller impact on the overall image the higher the pixel density gets.
If you wanted to do a valid test of the A7r vs an APS-C NEX body, cropped down to an APS-C FoV, you should be using a 16MP body, then you'd have something approaching an accurate comparison between the two as the A7r is a ~15MP body when cropped to APS-C (which is close enough for light testing).
If you really want a good test of IQ, I'd compare the D800E to the A7r with the same lens (say the otus you own). Then for a good comparison of what you are doing here, compare the A7r to a D610, also with the Otus (which would have the A7r in the same situation as the NEX-7 is in your test).
Dec 23, 2013 at 08:59 PM
p.2 #11NEX7 v A7r Image Quality Field Testing
michael49 wrote:
Agreed.
When did forum become dpreview? The usual paucity of these types of ridiculous threads here is why I much prefer FM over dpreview - Max - please, take this silliness there.
I think that is where he came from..
But for some reason all "MaxBerlin" threads I saw over there linked to this text:
"Forum thread deleted or withdrawn
The thread that you are looking for has been deleted by the moderators or withdrawn by the original poster. If you have a second, we would appreciate if you let us know in case you have reached this page via some existing and genuine link. We will investigate the problem as soon as possible."
bcaslis wrote
Another one of these threads. When did this become the Max forum? Really, this is getting tedious.When did forum become dpreview? The usual paucity of these types of ridiculous threads here is why I much prefer FM over dpreview - Max - please, take this silliness there.
I think that is where he came from..
But for some reason all "MaxBerlin" threads I saw over there linked to this text:
"Forum thread deleted or withdrawn
The thread that you are looking for has been deleted by the moderators or withdrawn by the original poster. If you have a second, we would appreciate if you let us know in case you have reached this page via some existing and genuine link. We will investigate the problem as soon as possible."
Dec 23, 2013 at 09:04 PM
p.2 #12NEX7 v A7r Image Quality Field Testing
Max, you've been given multiple substantive explanations for the problems in your comparison, yet you insist its valid. Maybe it's best to just leave it at that, and let us all be wrong.
I hope you find a camera system that makes you happy, whatever your metric.
I hope you find a camera system that makes you happy, whatever your metric.
Dec 23, 2013 at 09:13 PM