Industry News: Sony Introduces the High-resolution A7R IV with World’s First 61.0 MP Back-illuminated, Full-frame Image Sensor

AlanF

Desperately seeking birds
CR Pro
Aug 16, 2012
12,355
22,534
He had images to back up his claim.



They are not 'the people' who test anything. They're just another site on the web. Their data must stand or fall on its own, and not on any claim to authority made by them, or by you on their behalf.

And it falls. They say they tested three lenses multiple times and chose 'the highest results' but they do not state which lens produced the highest results. I don't know what lens those data points are from, or if they're even all from one lens!

More importantly, they don't tell you want lenses were used on the other cameras. We know they're not adapting the two Canon lenses to the K-1, nor the RF lens to anything. So they weren't using the same lens on each body. How were the other cameras tested? What lenses used, and which ones produced 'the highest results'? Where does each square on the graph come from? I don't know and neither do you. Which means the graph is useless.

I can post sample images that agree with Tony's claim, but you'll hand wave them as "eyeballing." Problem is you can shop the web for a graph proving anything you want, including that men never landed on the moon and that the Earth is flat. Between a 'guess which lens produced this data point' graph and my lying eyes, I'll trust my lying eyes. Heck, I'll trust Tony's lying eyes over sharpness graph bingo.
For the Matrix Tests (ie the sensor) quote:
The resolution of the matrix is based on the MTF50 function, and the measurements are standard on non-focused RAW files, which we have previously converted to TIFF format using the dcraw program. To escape from optical aberrations, we measure MTF50 values only for the f / 4.0-f / 16 aperture range, in which the main factor limiting the lens's performance is diffraction. It is also worth reminding that on each aperture we make from a dozen to several dozen pictures (both with autofocus and with manual focus), then we choose the best ones. In this part of the test we used three lenses: Canon EF 100 mm f / 2.8 L Macro IS USM, Sigma A 35 mm f / 1.4 DG HSM (attached by the adapter) and RF 24-105 f / 4.0. The highest results obtained with the use of these lenses are presented in the graph below.
In other words, they are minimising the effect of optical aberrations by using small apertures where diffraction is limiting the lens resolution and even then checking further by using different lenses to make sure that there are not residual aberrations. This is far more rigorous than using the images produced by people where the lens itself is contributing to the overall resolution.
 
Upvote 0
Success at being third? Was that even an improvement over Minolta, the company they bought?

Sony really couldn't compete in DSLRs. They made a bit of a splash with APS-C mirrorless. Their main accomplishment was offering the only FF mirrorless bodies for a few years. The A7's and A9 are what they are known for. You hear 'Sony, Sony, Sony!' so much from certain sites and 'influencers' that it's difficult to imagine the EOS M series out sells Sony in the only market where mirrorless out sells DSLRs, Japan. Yet it does. The mocked, derided, limited lens catalog, 'old Canon sensor' M series kick's Sony's butt in the most important MILC market in the world.

Neuro nails it every time he points out that the things we think about and debate on this forum are not the things that concern the average consumer or probably even the average professional. "DR, DR, DR!!!" yet what percentage of photographs are exposed ETTR RAW, processed for maximum DR, then printed to 16x20 or larger? 1%? 0.1%? 0.01%? I can count on one hand the times a freaking 11 stop 7D failed to capture what I wanted in a single frame, yet thanks to DxO and dpreview sensors are now judged ONLY by DR and Sony is 'so far ahead.' Give me a break. The R is sharper/more detailed than the A73 despite having an AA filter (something even Tony Northrup pointed out). This impacts every image taken. But 'Sony is ahead' because of a DR difference that would be nearly impossible to exploit. DxO biases their score to this tiny DR difference while completely ignoring resolution and sharpness.

A number of Sony sensors, including the A9, have less DR than the 5D4/R but 'Sony is ahead.' The A72 had particularly poor high ISO for a FF sensor, well behind a 6D/6D2/RP, but 'Sony is ahead.' The first and second generation A7's suffered from off sensor flare, compressed RAW artifacts, and eating stars but 'Sony is ahead.' Both Canon and Fuji 4k footage grades far better than Sony's due to Sony's 100 Mbps limit but 'Sony is ahead.'

No, Sony is not ahead on every metric even when we look strictly at sensors.



It has a few very impressive specs but with Sony's typically poor ergonomics and poor weather sealing. It also has less DR than a 5D4 or 1DX II so clearly it's junk and Canon is ahead :LOL:



Its competition is the A73. It has better stills IQ. It produces video that grades better but is worse at high ISO (due to the crop factor; stills high ISO is on par). Despite the fact that we Canon users complain a bit about the ergonomics, it has better ergonomics than the A73 (showing how spoiled Canon DSLR users really are by Canon ergonomics). It's better weather sealed. It has better/worse AF depending on the situation. It lacks some of the cool features of the A73 (like IBIS).

And at the end of the day, the differences are not going to matter in the hands of competent photographers / videographers. The guy who really wants IBIS for adapting old glass, or maybe high ISO 4k footage, is going to get the Sony. The guy who wants weather sealing and a touchscreen is going to get the R. Some other guy will decide based on what's on sale that week.

Small differences in select sensor metrics don't matter nearly as much as you imagine.

It's funny, you take specific issues with individual sentences and debate them. Why don't you debate the second sentence too ("Canon has better support, a wider product line, but Sony's main advantages are related to the sensors.")?

How about commenting on the entirety of a post rather than just one sentence at a time with different assumptions/bases for each reply? Here I copied it below for you. So Sony's camera division is profitable because of something besides sensors? Sony is able to develop and grow their A7/A9 offerings because of something else besides their sensors? So Sony is able to fund development of there GMaster line because it is profitable due to something else besides sensors? Nikon and Fuji and using Sony sensors instead of their own -- why? Why don't you tell me what you think is the reason behind Sony's success BESIDES sensor technology?

If Sony doesn't win on sensors, what was its success built on? Canon has better support, a wider product line, but Sony's main advantages are related to sensors. It's true that Canon got to 50MP first, but Sony pushed DR, face/eye detect, and higher video specs. The A9 is a serious bit of engineering.

The R is a good camera, but it was not a Sony beater. I bought one knowing exactly what the limitations are, and I'm hoping that Canon's next bodies will be truly "pro". I don't need the best video specs, but I do want AF that is capable of BIF (not because I bird, but because that type of AF can keep up with kids on the soccer field, which I do shoot a lot of... with a 5D4). As an overall camera, I still like the 5D4 better than the R, but I fully expect the pro R to make the 5D4 obsolete as it should.
 
Upvote 0

dtaylor

Canon 5Ds
Jul 26, 2011
1,805
1,433
In other words, they are minimising the effect of optical aberrations by using small apertures where diffraction is limiting the lens resolution...

Lenses do not magically become identical at the so-called "diffraction limited aperture" because DFA is not a hard line. Diffraction does not take over at that line, it merely becomes apparent alongside all other factors. DFA is very much a misnomer. And btw, f/4 and f/5.6 are well below the DFA for a 30mp FF sensor.

Also: we were discussing sharpness, not resolution. The author(s) at that site confuse the two repeatedly, but we're not going to do that here. There's no question that a 30mp sensor can out resolve a 24mp one in the same format. All other things being equal one would expect a sharper image out of the 30mp sensor as well at the same view size.

So again, which lenses produced the data points on that graph? You don't know because they don't say for any of the cameras.

This is far more rigorous than using the images produced by people where the lens itself is contributing to the overall resolution.

They weren't testing for resolution, they were testing for sharpness (MTF50). Are you now claiming that the lenses used in their tests did not contribute to the overall sharpness results?
 
Upvote 0

dtaylor

Canon 5Ds
Jul 26, 2011
1,805
1,433
It's funny, you take specific issues with individual sentences and debate them. Why don't you debate the second sentence too ("Canon has better support, a wider product line, but Sony's main advantages are related to the sensors.")?

I reply to the stuff I disagree with. You want me to disagree with everything for the sake of disagreeing?

Why don't you tell me what you think is the reason behind Sony's success BESIDES sensor technology?

I already did. They bought Minolta (inertia and tech portfolio) and were the first and only FF MILC manufacturer for a few years. The former got them the #3 spot which they held onto, and the latter has gotten them a ton of press and good profitability in a subsegment of the market.

Now they do sell sensors as a separate business. That business obviously relies on them producing sensors people want to buy. But the sensor metrics that are hyped to the moon are not all that important in terms of their position in the ILC market.
 
Upvote 0

AlanF

Desperately seeking birds
CR Pro
Aug 16, 2012
12,355
22,534
Lenses do not magically become identical at the so-called "diffraction limited aperture" because DFA is not a hard line. Diffraction does not take over at that line, it merely becomes apparent alongside all other factors. DFA is very much a misnomer. And btw, f/4 and f/5.6 are well below the DFA for a 30mp FF sensor.
I've been analysing data like those for years. Here are plots for good lenses on the 5DIII, measured by lenstip (top), on the 5DII by photozone (opticallimits) (middle) and 5DSR by ePhotozine (bottom) from my files. You can see how the different lenses MTFs converge as the f/number increases. There is still some variation and so Optyczne uses data from the higher f-numbers and the best values to measure the resolution of the sensor.

185736185737185738
 
  • Like
  • Love
Reactions: 1 users
Upvote 0
I reply to the stuff I disagree with. You want me to disagree with everything for the sake of disagreeing?



I already did. They bought Minolta (inertia and tech portfolio) and were the first and only FF MILC manufacturer for a few years. The former got them the #3 spot which they held onto, and the latter has gotten them a ton of press and good profitability in a subsegment of the market.

Now they do sell sensors as a separate business. That business obviously relies on them producing sensors people want to buy. But the sensor metrics that are hyped to the moon are not all that important in terms of their position in the ILC market.

No, but context is lost when you respond to individual sentences and take a different basis to address each sentence.

It's true that Sony bought Minolta and it was one of the earlier manufacturers of FF MILCs that were affordable, but I don't think that was the main reason for the A7/A9 series' success. The original A7 didn't have IBIS, had horrible battery life, subpar AF, horrible warranty/support, limited lens selection, etc. and yet people bought them. Back then, DSLRs did a lot better than Sony's A7, but Sony had advantages with their sensors. People adapted Canon and Nikon glass to Sony bodies because of Sony sensors. Without the sensors, I don't think Sony would have maintained its #3 spot nor would it be within striking distance of #2, Nikon.
 
Upvote 0

dtaylor

Canon 5Ds
Jul 26, 2011
1,805
1,433
I've been analysing data like those for years. Here are plots for good lenses on the 5DIII, measured by lenstip (top), on the 5DII by photozone (opticallimits) (middle) and 5DSR by ePhotozine (bottom) from my files. You can see how the different lenses MTFs converge as the f/number increases.

They're not even fully converged at f/20! The original graph you posted had plot points for f/4, f/5.6, and f/8. Stop pretending that the lens doesn't matter, it absolutely matters. And they didn't state which lenses they used. Which means their graph is unusable and does not support your claim in the least.

There is still some variation and so Optyczne uses data from the higher f-numbers and the best values to measure the resolution of the sensor.

Whether you're testing resolution (MTF10) or sharpness (MTF50, what they were testing) you cannot equalize lenses this way.

And you want to criticize Tony Northrup? The man is a paragon of scientific integrity by comparison to your sleights of hand here.
 
Upvote 0

dtaylor

Canon 5Ds
Jul 26, 2011
1,805
1,433
It's true that Sony bought Minolta and it was one of the earlier manufacturers of FF MILCs that were affordable, but I don't think that was the main reason for the A7/A9 series' success. The original A7 didn't have IBIS, had horrible battery life, subpar AF, horrible warranty/support, limited lens selection, etc. and yet people bought them. Back then, DSLRs did a lot better than Sony's A7, but Sony had advantages with their sensors.

The most often cited Sony sensor advantages are DR, FF 4k, and fast on sensor AF. The A7 only had one of those, good DR, and it was behind Nikon on that metric. And yet people bought them, indicating sensor metrics were less of a factor than simply having a FF mirrorless camera, EVF, lens adaptability, etc.

People adapted Canon and Nikon glass to Sony bodies because of Sony sensors.

The Nikon D8x0 sensors have always been excellent and have set the bar for DR. It was only the generation 3 A7's that finally stood toe-to-toe with the D8x0 series on DR. So why would someone adapt Nikon glass to a lower spec'd Sony sensor? Again, it's more likely they just wanted mirrorless and whatever features the Sony had that the Nikon didn't (i.e. video related; exposure preview; etc).

Also, when you say 'people adapted their lenses' there was no mass movement. Sony's "success" has been in holding #3. But who was the challenger to them from below? Pentax? Fuji? Would they really have lost the #3 spot had they been shipping cameras with sensors made by someone else?
 
Upvote 0

AlanF

Desperately seeking birds
CR Pro
Aug 16, 2012
12,355
22,534
They're not even fully converged at f/20! The original graph you posted had plot points for f/4, f/5.6, and f/8. Stop pretending that the lens doesn't matter, it absolutely matters. And they didn't state which lenses they used. Which means their graph is unusable and does not support your claim in the least.



Whether you're testing resolution (MTF10) or sharpness (MTF50, what they were testing) you cannot equalize lenses this way.

And you want to criticize Tony Northrup? The man is a paragon of scientific integrity by comparison to your sleights of hand here.
No they have not fully converged, that is why optyczne take the consensus highest values, which I have plotted. Yours and Northrups images are fully dependent on lens quality, whereas optyczne attempts to minimise the contribution of the lens. I have no wish to discuss this any further and will not be doing so.
 
Upvote 0
Jul 21, 2010
31,099
12,863
Increased profitability. Sony is in a better position now than was a decade ago. It's lens lineup is better and the ecosystem is better fleshed out and it is able to sustain future development (unlike Pentax). Without their sensors, do you think Sony would have been able to maintain the status quo. I don't think so -- it would have fallen behind Fuji into Ricoh/Pentax status.
So ‘success’ = not falling behind. Okidokee.

Can you really say that sensors increased profit, any more than say reducing personnel or manufacturing expenses?

Incidentally, two of the drivers of market share are brand loyalty/familiarity and popularity. I have Brand X so I’ll buy another one of them when the time comes. Particularly true when accessories (e.g. lenses) help lock in users. Plus if my friends use it, maybe I should too. Point being, barring a paradigm shift (Sony sensors aren’t, nor is removing the mirror from an ILC) or a manufacturer truly screwing things up, market share shifts slowly. Sony was (and remains) 3rd, well ahead of Fuji/Pentax/etc. They didn’t need great sensors to maintain that lead over Fuji/Pentax.

If a better sensor actually had the impact on the market that you seem to think it does, Sony would have gained some market share over time, as would Nikon for putting Sony sensors in their bodies. Instead, Sony stagnated and Nikon lost while Canon (without ‘great sensors’) gained.
 
Upvote 0

dtaylor

Canon 5Ds
Jul 26, 2011
1,805
1,433
No they have not fully converged, that is why optyczne take the consensus highest values, which I have plotted.

You simply cannot eliminate lenses as a factor this way. It's stupid to try when the obvious course would be to test the same lens on all bodies. Regardless I will believe my 'lying eyes' over some graph on a random website.

Edit: to be clear, I don't think the MP difference or sharpness difference (due to MP) is any reason to choose one or the other. I think the hairsplitting over sensor metrics has reached the point of being ridiculous. I brought it up in reply to yet another post proclaiming how far ahead Sony is in sensors.
 
Last edited:
Upvote 0

ashmadux

Art Director, Visual Artist, Freelance Photography
Jul 28, 2011
578
145
New Yawk
photography.ashworld.com
Given Canon's many innovations over recent years I would say the opposite is obvious.



What are we discussing then? The selective set of metrics where Sony is ahead? Because that's what it feels like any time someone cries about Sony innovation on this forum. Should I go to a Sony forum and insist that we only discuss weather sealing, DPAF, color science, and super fast lenses? In that case it's obvious that Sony has been dragging their feet for years.



So in a basketball game the team that scores 18 points "drop kicks" the team that scores 40?



What's the critical thing that's missing again? A fraction of a stop of DR at ISO 100?



Sorry. I've shot with an A9. It was not a game changer. I can see how it would be very nice for a newbie. I cannot say that its keeper rate overall would be better than the keeper rate of someone who knows how to use a DSLR's AF system. I felt like it was less work, but I also felt less confident in it. Kind of like an auto versus manual transmission.



If the A9's AF was truly a game changer...like the difference between AF and no AF, or Canon USM lenses vs in body AF motors in the 90s...you would have already sold the 5D3 for an A9. Intelligent subject tracking is a 'nice to have' feature. The R has it, but it's not as fast as Sony's. By the time your 5D3 goes belly up it probably will be.


Even going point by point, you're replies are a mix of wrong, wth, and just straight on nonsense.

Just like this here: "you would have already sold the 5D3 for an A9."

This is so dumb, I'll just leave it here to mention that its dumb. Because you do know how I make my hardware decisions. :rolleyes:

and this keeper: "So in a basketball game the team that scores 18 points "drop kicks" the team that scores 40?"

You're trying to make a point I'm not even arguing. Critical thinking, look it up. Its the deflectors in here that want to tall about market-share...red-herring. :unsure:

"I can see how it would be very nice for a newbie."

Newbies? Wow. A9. Really...wow. Where's the canon equivalent......(crickets). And I'm a canon fan. But I dont live under a rock. The A9 "isn't(wasn't) a game changer.." Because YOU said so, riiiight (n)

At the end of the day, Canon is taking too many punches in the gut (more like b*tch slaps up and down the street). They have the worst bodies on the market. Years old nikons DSLR's are still more capable than recent canons. The old d500...canon likely wont even reach that level. Years of ho-hum products are supposed to produce big sales? Of course not. They made this grave, and they dont have the tech to dig themselves out.
 
Upvote 0
Apr 25, 2011
2,510
1,885
The original A7 didn't have IBIS, had horrible battery life, subpar AF, horrible warranty/support, limited lens selection, etc. and yet people bought them. Back then, DSLRs did a lot better than Sony's A7, but Sony had advantages with their sensors.
No, back then Sony was selling an idea that mirrorless is "cool" because it's small. A7 was a fashion item.
 
  • Like
Reactions: 1 user
Upvote 0

dtaylor

Canon 5Ds
Jul 26, 2011
1,805
1,433
This is so dumb, I'll just leave it here to mention that its dumb. Because you do know how I make my hardware decisions. :rolleyes:

A "game changer" means you're missing shots that the other guy is getting and it's entirely due to equipment. When that actually happens you typically upgrade. That kind of a change has been rare in photographic history, despite the hype associated with every change.

You're trying to make a point I'm not even arguing. Critical thinking, look it up.

If you looked it up you might realize you have completely failed to support your claim that Sony is 'drop kicking' Canon. Just because there is a feature Sony has that you really want does not mean that in the grand scheme of things, all technology considered, A is drop kicking B.

Newbies? Wow. A9. Really...wow.

Yes, my opinion is that intelligent tracking features (like eye AF or animal AF) are most beneficial to newbies. They may reduce the workload of an experienced shooter, but they're not necessarily the difference between getting/missing the shot for experienced shooters.

The A9 "isn't(wasn't) a game changer.." Because YOU said so, riiiight (n)

But it is a game changer because you said so?

At the end of the day, Canon is taking too many punches in the gut (more like b*tch slaps up and down the street).

At the end of the day you have a lot of emotional rhetoric ('b*tch slaps', 'made this grave') and empty assertions ('worse bodies on the market'). And that's all.
 
Upvote 0

AlanF

Desperately seeking birds
CR Pro
Aug 16, 2012
12,355
22,534
You simply cannot eliminate lenses as a factor this way. It's stupid to try when the obvious course would be to test the same lens on all bodies. Regardless I will believe my 'lying eyes' over some graph on a random website.

Edit: to be clear, I don't think the MP difference or sharpness difference (due to MP) is any reason to choose one or the other. I think the hairsplitting over sensor metrics has reached the point of being ridiculous. I brought it up in reply to yet another post proclaiming how far ahead Sony is in sensors.
Optyczne.pl is the mother site of lenstip, one of the most respected photographic sites on the web. The lenstip reviews are the English translation of the optyczne ones. If you just write them off as a “random site” because their considered reviews and careful measurements conflict with your eyeballing, then there is no point in having discussions with you. You should not be accusing others of emotional rhetoric.
 
Upvote 0

dtaylor

Canon 5Ds
Jul 26, 2011
1,805
1,433
Optyczne.pl is the mother site of lenstip, one of the most respected photographic sites on the web. The lenstip reviews are the English translation of the optyczne ones. If you just write them off as a “random site” because their considered reviews and careful measurements conflict with your eyeballing,

No, I wrote off the graph you copied here because it does not prove what you think it proves. I made a snarky general comment about my eyes and web sites, but my dismissal of your graph has nothing to do with that.

And if 'careful measurements' repeatedly conflict with simple, direct observations, then it's time to question the more complex process. Both your understanding of it and what it means, and the actual measurements themselves if need be.
 
Upvote 0

PVCC

Arts & Engineering
Jul 5, 2019
97
23
Optyczne.pl is the mother site of lenstip, one of the most respected photographic sites on the web. The lenstip reviews are the English translation of the optyczne ones. If you just write them off as a “random site” because their considered reviews and careful measurements conflict with your eyeballing, then there is no point in having discussions with you. You should not be accusing others of emotional rhetoric.

I like Lenstip reviews.

The only problem I see with this and several other review sites, is the fact that they only review one copy of the lens.

This is VERY important to keep in mind.

If they get a good (or bad) copy the results will be completely different from other ones.

I noticed it with a lens review they posted long ago (don't remember the lens).

So, for me, only sites that test many lenses copies are useful when considering a review of expensive lenses.

Otherwise it's kind of lottery...
 
  • Like
Reactions: 1 user
Upvote 0

Aussie shooter

https://brettguyphotography.picfair.com/
Dec 6, 2016
1,183
1,817
brettguyphotography.picfair.com
Jist had a quick look at the TN image quality video. At one point he is doing the whole 5 stop pushing BS and comparing the quality between the a7r3 and a7r4. To my eyes the 4 was far far worse but he straight up said the 3 was worse. Either I need to go to the eye doctor or TN is on drugs. Anyone else seen it?
 
Upvote 0
Mar 2, 2012
3,187
543
If that isn't what you meant at all, you shouldn't cherry pick other people's posts, right (read the entire post I copied below)? Nikon doesn't get the best of Sony's tech -- at best is a delayed version. Sony led with the 40+ MP sensors in their own cameras for years before it was available to the D850. That is what happens when you're not in charge of a key component in your product. And a "win" in sensors doesn't mean that it has to dominate because, as you imply, there are other factors that influence the choice of camera system. But a win can translate into a competitive advantage that helps sales and profitability. Do you really think something else besides sensor technology is the primary advantage for Sony and a reason why Nikon isn't really hurting (imagine the D850 with a 5DR sensor?!).

You responded with "Their strategy of diminutive cameras that much of the market likes?" I don't think so. If it were true then this Sigma FP is going to be a best-seller, but I don't think it will... If camera size was the number one factor, then u4/3 and APS-C cameras would be doing better.

I didn’t cherry pick; I responded to a post without quoting it all.

Yes, I suspect sony’s camera size has affected its sales over the past several years. Part of the market craves small. I am not among that market; I owned an a7rii for a while and didn’t like using it - I prefer 5D size. However the market is there and for some time Sony alone was selling small 135-format ILCs. I’ll also note Sony leadership listed the ability to be small as the primary advantage of their system. If it were sensor first, I do indeed think nikon would substantially outsell Sony even if there were a year-ish delay between good and slightly better. There are multiple brands now selling small 135-format ILCs, and I expect that will impact Sony camera sales more than if Canon or Nikon put an a9 sensor into a 1D or D6 form factor.

I also question the assertion that Sony withholds sensors from customers like nikon. I don’t really follow APSC, but as far as I know the last time Sony and Nikon both used the same full frame sensor, Nikon was first to market.
 
Last edited:
Upvote 0

sdz

CR Pro
Sep 13, 2016
262
209
Pittsburgh, PA
Jist had a quick look at the TN image quality video. At one point he is doing the whole 5 stop pushing BS and comparing the quality between the a7r3 and a7r4. To my eyes the 4 was far far worse but he straight up said the 3 was worse. Either I need to go to the eye doctor or TN is on drugs. Anyone else seen it?

I saw it too.
 
Upvote 0