Sony a7R IV added to studio test scene comparison

Ah... there is a rub. They are just measuring reproduction of one chart. Film manufacturers used to do all kinds of scientific measurements of film reproduction and surely all camera companies do extensive studies of this. But they also shoot real world scenes and test viewer preferences of a range of "looks."

What film companies learned long ago is that supposed "accurate" reproduction did not always correlate with what people preferred. Even in the early b/w testing they learned that a 45 degree D/log E reproduction of a scene lost out in preference to a curve that had less contrast in both the shadows and the highlights.

Maybe what people prefer can change as we are viewing on monitors now and seem to accept even extreme HDR imagery.

Years ago my stock agency told me to shoot only Velvia. Those jumped out on the light box and were selling better.

Since we are using digital cameras we can abandon any thought of being stuck with the choice that the manufacturer gives us.

For fun I clipped out the best chart - Fuji and put it up with the worst chart - Canon RP. I can't see anything that would concern me. It seems some Canon colors are a bit more vibrant. I think you'd have to be very specialized color reproduction work for these differences to matter. And in real world photography, how could this matter? Besides, you can apply a "corrective" profile to any file.

https://www.dpreview.com/galleries/3401543788/photos/3973843/fuji-canon-color-test

I find the Northrup experiment ok as a brief study. Those who think color is something absolute should evaluate their claims once more.

Of cause those will not rethink, since they already know that they are right, and that the Northrups and a lot others are wrong! :-)

To the paranoid Jake. I really doubt that all places are somehow paid. That was a study on peoples perception genius. I really doubt if all of them were Sony fanboys. Especially when the same test was given WITH and WITHOUT
the pictures tagged with the camera that created it and the results changed.

THE first test was without any knowledge of what camera was used. That is what the word "BLIND" means genius. So there couldn't be any bias DA, Try looking (and more important following) the video before commenting.

Funny how it seems to Canikon fanboys that everyone is in a conspiracy when their fav camera doesn't measure up. DXOmark, Lensrental with MTF (even though they use equipment to measure, they are still biased to some fools).

Here is another, not testing peoples opinion but using other means:

https://www.pdnonline.com/gear/cameras/the-best-cameras-for-color-reproduction-ranked/

That video was in reality a low point for Tony. That was incredibly bad for an educated person. That's not how you do studies. The only valid point was that people have confirmation bias, which everybody already well aware of.

Regarding the best colors/tonality/skin tones, that experiment was a complete failure. For my own purposes, I test cameras side by side with equivalent lenses at equivalent settings, taking hundred or so pictures of identical subjects and portraits in different light. Then I process raws as I usually do. And then compare them side by side in LR. There is never a camera that is always best in every image, but statistically you can see which camera produces more pleasing images. You can also estimate how much time the processing takes.

@ forpetessake: If everybody was well aware of confirmation bias, why all the weird or false claims about color and other issues in these forums?!?

And how do you pick "best colors/tonality/skin tones? For you? As a general fact? Or taste? Who's taste?

What do you think your hundreds of pictures tell, more than the pretty simple study than the Northrups did - and actually published?

@ forpetessake: You can measure color, you can reproduce color, but you can't tell exactly how others perceive color. And taste is beyond rational thought.

How is that they faded so speedily? There was a dozen cameras tested for the last year, but you don't see those bad colors on others. And how is that fading effected jpegs and raws so differently? I would take any of those explanations with a grain of salt. I don't really like the idea of showing a "controlled" test and then putting a green disclaimer that it's basically meaningless.

How? They used Sony photo paper for the prints.

forpetesake "I would take any of those explanations with a grain of salt. I don't really like the idea of showing a "controlled" test and then putting a green disclaimer that it's basically meaningless."

There are other colors on the test chart that have not faded. Try using those. I think the problem is that DPR has a lot of historical test data they don't want to just throw away. As they've noted themselves, they're trying to find a solution.

For me personally, I always shoot RAW. I could care less about JPEGs or OOC files. I prefer color accuracy over a camera company dictating their color choices. In the film days, we didn't have much choice. With digital, we do.

Feel free to actually put it in the comparison chart, and actually compare at ISO 100-6400... before you comment next.

Edit:
The A7III is holding up better than the A7RIV.. with the A7RIII just about on par with the 5DSR up to ISO400/800.
At ISO 1600 the A7RIV & Z7 are starting to show more noise than 5DSR & A7RIII.
At ISO 3200 the A7RIV, 5DSR and Z7 seem to be around the same.. with the A7RIII more than likely implementing a bit more noise reduction.

@cbphoto123

The A7rIII is the previous model and thus is always included in the comparison tool.

The other 2 cameras are PEERS of the A7rIV. The point is that the viewer is able to judge the A7IV against comparable cameras that they might also be considering BUYING.

The Canon 5DSR cannot realistically be assumed to be an alternative choice for a potential buyer of the A7rIV.

In the first place it is a DSLR, not a mirrorless, secondly it is a very old camera with a legacy mount. Also its performance and specifications are not competitive with either the Nikon z7 or the Panasonic S1r - it has no IBIS, only 61 focus points etc.

That is why DPR did not chose it as a default camera. Nevertheless there is nothing stoping you selecting it from the pull-down list and measuring the performance against the Sony or any other camera.

I did. It's even worse than the S1R. Stop lying!
https://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=lowlight&attr13_0=sony_a7riv&attr13_1=canon_eos5dsr&attr13_2=panasonic_dcs1r&attr13_3=nikon_z7&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=6400&attr16_1=6400&attr16_2=6400&attr16_3=6400&attr126_0=1&attr126_1=1&attr126_2=1&normalization=compare&widget=715&x=0.021416257999689027&y=0.0070858164969545825

@kristo93.. based on your response, I can only conclude that either:

https://www.psychologicalscience.org/news/full-frontal-psychology/is-reading-comprehension-the-hidden-disability.html

..or that you are born in 1993 and have fallen victim to the far too often poor standards of US public education...Maybe a bit of both?

@cbphoto123 I had done what you suggested prior to posting. I went back again just to be sure of my impressions. I will agree that there is more contrast with the Canon (something I think is easily adjusted in PP) but I didn't see any scenario where I thought the 5DSR showed less noise than the A7IV or Z7. They're pretty close but if I had to pick a winner it would not be the 5DSR.

I think you're not only blind, but have reading comprehension issues with your own text as well. But each to their own. Contrast has more to do with the lens and post processing than the sensor.

I'll also have to add kinda stupid to the list now. You clearly said the 5dsr it's better iso 100-6400. Do you want me to quote you?

"Feel free to actually put it in the comparison chart, and actually compare at ISO 100-6400... before you comment next"

It's actually worse at any iso, just not very visible at 100. But start pulling those shadows and it will fall apart immediately.

As for post processing... There's always post processing, LCD screens and web browsers can't display raw sensor readout. First it goes through the AD converters, then gets processed in camera. Then camera raw applies a camera profile and demosaics it. Loads of processing where the contrast changes. The Sony sensors have way higher dynamic range so the lower contrast is a good thing, helps with editing.

What a piece of work... The 5dsr is softer and noisier than the R4 at any iso. That's a fact, it's right there in the comparison shots. No idea why you think you can keep lying about it... As for pulling shadows at iso 100? It's kind of a huge deal in photography, has been since film. That's what matters at iso 100.

If you're going to look at colors, look at the color checker. To my eye, all the colors there look nearly identical. Also, the colors in the tubes of paint look just fine. As others have already noted, the problem with the colors in the faces is the fact that they are older pictures and are starting to fade.

@JPatrick
Not likely, rather someone at sony loaded old (pre-A7III) color profiles and white balance in the firmware, the ones that give all the faces liver disease. Very similar to the same crappy ones that S1R is using.

@photomedium I don't know why the faces look so terrible but if you look at the color checker and the tubes of paint you get a completely different story. Just a guess, but I suspect the color checker squares as well as the paint tubes are more color fast than the face images. Your theory about color profiles and white balance does not hold up when the colors of those objects are essentially the same for all the camera brands.

Read the green information below charts. Read the green information below charts. Read the ....

Color perception and preferences are very complicated and tricky subjects.

If it were easy, all manufacturers would have figured out long ago how to interpret colors the exact same way in every circumstance, we would not adjust them at all, and everyone would be happy with the results.

E.g. do you want that golden hour glow warming the colors, or do you want the image to look like it was shot at 6000K?

Do you want your dusk shots of a building to look cold or warm? I sometimes can't decide.

@ AlanG. True. And the perceptual and psychological impact of color has been discussed long before Johannes Itten, Wassily Kandinsky, Josef Albers and others at the Bauhaus school took a more scientific approach on the exploration of color theory.

Photomedium. I did watch the video. Good information but it certainly did not attempt to explain much beyond biases.

There are other issues at play...color and brightness constancy, surface characteristics, order that images are presented and displayed, personal preferences and prejudices depending on subject (e.g. warmer skin tones vs. cooler), color vision quality of the observer, what else is viewed nearby for comparison.

He concluded that respondents preferred the Sony image(s) but did not explain why or what about it made it better. Trying to make more likeable ooc jpegs is fine but is no interest to me. It certainly doesn't have anything to do with how an adjusted file can look.

What I wonder is if most people actually care about color. Not just color balance or the look of a photo, but color in general.

Take cars for instance. The overwhelming majority are white, grey black silver or may have a hint of color. Cars used to celebrate color. White kitchens dominate now. Homes are becoming neutral. Look for colorful tile at Home Depot. You won't find much. Where did all of the color go?

Only way that someone could establish a color preference is if side by side comparison is done.

science and northrups don't really go together.

I think Tony Northrup and science go well together. He is always very interested in what is behind the outcome.
In this video you see one part of scientific work: blind test, reasonably large sample of people, simple statistics.
Screening the preferences of 1,500 people showed that majority of them (1,336) considered photos from Sony to be the best. People liked them most.
And remember the result does not in any way depend on Tony Northrups opinion nor his likings.

And if put commercial business into this, it is only natural that Sony Corporation is pleased into something like this. This sells. Their effort was not in vain.

However, color perception is very personal matter. It is a learned ability.
White balance is of course fine if a white object is rendered white into the photo.
However, human mind does strange things: we may very well see an object as white, if we think it is white. Also in very dark, we may think we do see colors (looking at known objects), which is not possible.

@Magnar W
I admit I have no idea what is it that they mean saying "colour science", but looking at colour transforms calculated from DPR studio scene for ILCE-7RM4 and ILCE-7RM3 - the difference is non-existent. Including ILCE-7M3 just in case:

{"ILCE-7RM4",0, 0,
{5365,-1194,-299,-4504,11201,2859,-1305,2060,5609}},
{"ILCE-7RM3",0, 0,
{5394,-1121,-266,-4503,11223,2798,-1350,2179,6064}},
{"ILCE-7M3",-512, 0,
{5456,-1061,-238,-4396,11132,2774,-1421,2301,5894}},

Read the footnote on the article...

@ Iliah Borg: Too many are throwing out bold claims without no back up at all, like the OP here, appelton.

This "color science" phrase is the new diffuse argument that some people are using when they lack factual arguments.

even after down sampling its really bad.

Not surprised. The pixel density of higher resolution sensors are higher. The noise output of A7RIV is still impressive despite it being higher than Z7's.

Agree!

Look it over the A73 and A7R3...Something has gone wrong on this fourth iteration...

You read capabilities are supercrap
"Please note that since it was originally printed, the portraits on our test chart have faded. As such, they should not be used to draw comparisons between the color output of different cameras. We are aware of the issue and we are working on a solution."

Its pronounced "reading"
That could be it. The Z7 matches the color of the D850 and that was just a little while ago though, so seems kind of weird. That's what they say though so I guess we will have to go with it.
Also, color can be corrected in post so not a big deal.

I agree with Anulu, your reading comprehension skills are clearly lacking. Either that or you're just trying to find an excuse to dis Sony and purposefully ignored the disclaimer. So sorry you ended up buying a crap camera and taking it out on Sony. Cry me a river....LOL

I would also say the high ISO performance of the D850 is superior as well.

In terms of IQ I would agree not so keen on the noise I'm seeing but there are AF tracking advances from the A9 supposed to be in there which is one good reason but not worth it if you don't use it for wildlife. I will give this one a miss looking forward to the rumored A9II but not it's estimated price.

Trolls don't read...

The jaundiced note? A test chart good up to 2 weeks, one month ago and NOW is not good anymore? Or the problem is somewhere else?

It's a bit ironic that the alert about the yellow/green issue is in green, but possibly that's because people see green so much more clearly.

Furlan likes his observation and wants everyone else to like it too: one post on jaundice isn't enough. We might ask whether the correct form is jaundice or jaundiced though.

Regardless, yes, the high ISO performance is disappointing. Everything else on the a7rIV is a reason to pass on the a7rIII and pay the extra money for the IV, but I was hoping for IQ that would allow the 26mp APS crop to be used regularly and the full 60mp only as needed. Don't see how that can happen now. The downsized 60mp image is not demonstrably better than my existing APS at ISO 3200-6400, so expecting to use the 26mp crop is not realistic.

I'll probably wait for a Fuji GF50 with IBIS and phase detect. The high ISO performance of the Fuji 50 is worth the extra $1500 to me. But not yet. Fuji needs to bring down the tech from the 100 to the 50.

Oh man, you have some issues, just like DPR chart...

“Please note that since it was originally printed, the portraits on our test chart have faded.”

Ok, I understand. But why isn’t that such an issue with the cameras which you have added not so long ago? The Fuji GFX100 springs to mind.

It is an issue with the GFX100, and the S1R, and any other camera tested recently. We've been weighing different options to address this issue, but the reason the green text was added to this particular article, and not previous articles, is because the previous articles for the Fuji or the Panasonic didn't devolve into a storm of comments about how the 'color science' of the particular camera tested has regressed ('green', 'jaundice colors', etc.).

Isn't that interesting? Because of the preconceived bias that Fujifilm colors are good (and they are), there were almost no comments about the skintones when we published the GFX 100 story. And the one commenter who did comment on it astutely said 'many cameras have a green bias in skin now'.

It seems 'Sony has poor colors' is a common bias, partially b/c years ago Sony's colors did need some work, and partially b/c it's just a favorite chant made by anti-Sony trollers...

... This bias has suddenly caused many readers to, in somewhat hysterical comments, point only to the inkjet prints of faces, ignoring the other color-steady portions of the scene that show no regression in color. Hence why we felt the need to add the note in green.

Even better use the the GFX 50S, 50R and 100 (or take sony's A7R, mk2, mk3, mk4) and one can easily see the fading of the prints.

Can you show us the last camera tested with the faded charts for comparison?

Sony evolution:
https://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=daylight&attr13_0=sony_a7r&attr13_1=sony_a7rii&attr13_2=sony_a7riii&attr13_3=sony_a7riv&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=100&attr16_1=100&attr16_2=100&attr16_3=100&attr126_2=1&attr171_1=1&normalization=compare&widget=1&x=-0.8919788944007425&y=0.1666800676245467

Fuji evolution:
https://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=daylight&attr13_0=fujifilm_xpro1&attr13_1=fujifilm_gfx50s&attr13_2=fujifilm_gfx50r&attr13_3=fujifilm_gfx100&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=200&attr16_1=200&attr16_2=200&attr16_3=200&attr126_2=1&attr171_1=1&normalization=compare&widget=1&x=-0.8885127664429125&y=0.14983856958095748

Comparison of zones with little or no fading (also pink/orange tones involved):

https://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=daylight&attr13_0=sony_a7r&attr13_1=fujifilm_gfx50s&attr13_2=sony_a7riv&attr13_3=fujifilm_gfx100&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=200&attr16_1=200&attr16_2=200&attr16_3=200&attr126_2=1&attr171_1=1&normalization=compare&widget=1&x=-0.35265360546093055&y=0.2265196190047927

We did originally anticipate this problem, and made many copies of the print for swapping out. What we didn't anticipate is how quickly we'd have to replace them. It's on us for not addressing this problem sooner, and our apologies for that. We're moving quickly to address this issue.