What Would Improve Image Quality?

[a_c_skinner]

Upgrading technique is the answer for almost every occasion. Keep what you've got but use it better.

(Mind you I'm still torn about an X-E5 and 16-50.)

 

[Cherry_F]

If I were able to alternatively use the viewfinder and the LCD without having to remove my glasses, I suspect that would definitely improve my images !

___
Photography is so easy, that's what makes it highly difficult - Robert Delpire

 

[Hal Wilson]

In the thread Sensor Evolution Stuck, several people have asserted that we haven't seen improvements in image quality over the past decade.

What needs to happen to improve image quality?

• Expand dynamic range - would it be noticeable in JPG?
• Improve noise reduction at high ISO?
• Improve color science?
• Replace Bayer sensors with an "all colors per pixel" technology?
• ?

If we expand pixel count, will diffraction be a problem?

Do you actually need better IQ? Great chance if your photo is not to your liking, it's your lack of skill and not the camera. Almost all digital SLRs are capable of exceptional IQ.

I often shoot at very high ISOs (12,800+), in extremely dark places with fast moving action. Whist today's cameras can be extremely capable, and AI NR very good indeed, there's still some room for improvement. I personally would benefit from having cameras that can shoot better IQ in such conditions, although I appreciate this is a pretty extreme type of scenario. I do agree with the sentiment that a lot of 'failed' photos really aren't down to the equipment; I try to do the very best with what I've got, and enjoy the challenge the limitations of the equipment offers. Quite often, those limitations can be used to an advantage; I've learned how to use 'noise' to convey atmosphere; that fuzzy, grainy look can be quite appealing in certain contexts. So I do kind of wonder that if camera tach got so good the noise was all but eliminated, I'd lose a lot of that aesthetic. I was a big fan of pushing film to achieve a contrasty, grainy look, so I'd be sad to lose that really. Swings and roundabouts I suppose...

 

[ChuckB]

I am asking in the context of people complaining that there has been no improvement in image quality in the past decade.

Ignore them.

 

[a_c_skinner]

"I am asking in the context of people complaining that there has been no improvement in image quality in the past decade."

It isn't true. But it nearly is and is true for almost all ordinary applications - normal shooting conditions and the ordinary display modes.

 

[ahaslett]

Cameras released in 2015 include the Canon 5DS with its terrible DR, the Sony A7R2 with lesser resolution, lower DR and much less effective AF system than the A7R5, the Nikon D810A with worse DR than the A7R2…

In the last ten years, high performing lenses have become lighter and more affordable in real terms…

It’s not clear what is meant by “IQ”, but the ability to capture good images in different genres has advanced significantly.

Stacked Quad Bayer sensors with computational photography ex camera and proper RAW exposure aids would enable various approaches to improve light capture and avoid demosaicing artefacts.

Andrew

 

[RetCapt]

For me, AFAIK nothing.

My only photographic interest is scenic, which means sufficient light. Even night photography is dependent on bright highlights.

My sole products are the prints I make, up to the size limit of my printer (13X19").

I can't recall buying anything additional for improved image quality. A few times I have bought to gain additional features. This has been very seldom. Even at that I have still kept and used the older camera or body, because I did not need the additional features all the time.

I have no interest in buying gear on spec, hoping it will somehow improve image quality. Looking at my prints, from varying ages of gear, I can't see any differences in image quality. Thus I have not been moved to acquire newer gear based on an expectation of improved image quality.

I like to think I am getting the most out of my gear. I have been so complimented here and in other quarters where other photographers have looked at what equipment I used for a given image. My reaction to that is if someone is using the same or similar equipment and getting lesser results that is pilot error.

For years I have read countless threads about 'updating'. The qestion in my mind, because it is not included in the post, is to what end. If the poster cannot articulate that, what, other than GAS, is the point? Over the last several years, I am seeing fewer such posts, and in those, subsequent posters have challanged the OP to state specifically what they expect to gain from such an update.

For me, with my equipment, I am where the buck stops.

 

[a_c_skinner]

The 810A wasn't for general photography was it? The others would all produce pretty comparable pictures. I did say nearly too.

--
Andrew Skinner

 

[ahaslett]

The 810A wasn't for general photography was it? The others would all produce pretty comparable pictures. I did say nearly too.

Honestly, Canon lost landscape photographers to Sony when the A7R2 launched. I agree the D800E was OK, but still only 36Mpix.

In terms of capturing the shot, the A1 ii (for example) is better than any ten year old body. The A9iii excels in particular uses.

Whether any of this matters depends on what OP meant by IQ.

A

 

[J A C S]

• Improve color science?

Yes (but I do not like the term "color science").

 

[AstigmaticOwl]

• Improve color science?

Yes (but I do not like the term "color science").

Would you care to expand on that?

 

[GeoffRG]

This whole discussion is meaningless unless one is prepared to change cameras every time a new sensor is offered. In practical terms, nobody is going to"

• Expand dynamic range - would it be noticeable in JPG?
• Improve noise reduction at high ISO?
• Improve color science?
• Replace Bayer sensors with an "all colors per pixel" technology?

With an existing camera these things are set and cannot be changed except possibly for firmware tweaks.

In digital audio we already have a dynamic range capability that exceeds the range of human hearing. I suspect we aren't far off doing that with cameras having a dynamic range beyond human sight.

Many people have already said that noise reduction at the processing stage is more than adequate.

I'm not even sure what you mean by colour science, if you really mean improve the accuracy of colour rendition, you can do that with software already.

A sensor with different layers each sensitive to different colours sounds a reasonable way to go. However such things have been tried before and nobody was able to produce a sensor that, when properly processed, produced a colour rendition that was universally acceptable. The difference being that one could change sensors as the situation required. Sigma's sensor has its limitations. I think there's a long way to go on this one.

However, the question remains, when are you going to buy a new camera? The sensor you have is likely to be behind the development curve, My cameras are certainly several generations old but I'm not about to change them.

 

[a_c_skinner]

I find it difficult to distinguish people genuinely looking for the best images from those pursuing a theoretical best result which is why I generally say for ordinary subjects in normal display - say prints up to two feet wide or so and not black cats in dark settings.

I can see there is satisfaction in pursuing the ultimate IQ as well as the aesthetics of the finished image but if one's main interest is the latter not the former I stick with my view that pretty much anything made in the last 10 years will be satisfactory.

I can easily see the difference between my D200 images and those from later, but I'm more hard pressed to see differences between D800 and X-E2, X-T2 and X-H2 images. They are there but you have to look a deal more closely than most.

 

[Philip Eihuyar]

This whole discussion is meaningless unless one is prepared to change cameras every time a new sensor is offered. In practical terms, nobody is going to"

• Expand dynamic range - would it be noticeable in JPG?
• Improve noise reduction at high ISO?
• Improve color science?
• Replace Bayer sensors with an "all colors per pixel" technology?

With an existing camera these things are set and cannot be changed except possibly for firmware tweaks.

In digital audio we already have a dynamic range capability that exceeds the range of human hearing. I suspect we aren't far off doing that with cameras having a dynamic range beyond human sight.

Many people have already said that noise reduction at the processing stage is more than adequate.

I'm not even sure what you mean by colour science, if you really mean improve the accuracy of colour rendition, you can do that with software already.

A sensor with different layers each sensitive to different colours sounds a reasonable way to go. However such things have been tried before and nobody was able to produce a sensor that, when properly processed, produced a colour rendition that was universally acceptable. The difference being that one could change sensors as the situation required. Sigma's sensor has its limitations. I think there's a long way to go on this one.

Years ago, Fuji produced a different type of sensor which I felt had great promise, the 'dual photodiode' per pixel sensor in their S4 and S3 model. This could be a way to increase dynamic range for each color channel. The surrounding technology has caught up and could be Full Frame inplimentation. Any thoughts on this?

However, the question remains, when are you going to buy a new camera? The sensor you have is likely to be behind the development curve, My cameras are certainly several generations old but I'm not about to change them.

 

[a_c_skinner]

It is interesting and I was never wholly sure what it did either. Currently I imagine using a sensor that is part of a big run of production modified as needed is viable (think X-trans on a ??Sony sensor) whereas a small run of a novel sensor is not.

 

[Guy Parsons]

If I were able to alternatively use the viewfinder and the LCD without having to remove my glasses, I suspect that would definitely improve my images !

___
Photography is so easy, that's what makes it highly difficult - Robert Delpire

For me the answer was tri-focal glasses. Top area for infinity view and for the EVF, bottom or middle area for the screen. Some buy varifocal glasses with continuous variation to do the same.

 

[Hal Wilson]

• Improve color science?

Yes (but I do not like the term "color science").

Would you care to expand on that?

It doesn't actually mean anything, does it? It's just a marketing term dreamt up to make things sound more interesting than they actually are. Adds an air of mystique to another otherwise very boring aspect of the technology. In relation to selling cameras anyway.

 

[Dick Dastardly]

In the thread Sensor Evolution Stuck, several people have asserted that we haven't seen improvements in image quality over the past decade.

What needs to happen to improve image quality?

Image quality in digital cameras has been improving for 3 decades and we are past the era of any meaningful upgrades.

The same can be said about personal computers and every other household item.

If anything, the industry is reversing course in some cases and producing worse quality products than before, for example software.

 

[sportyaccordy]

The typical modern sensor is within 1/2 stop of the theoretical max for DR, and I imagine isn't far behind in any other sensor related metric.

IMO people need to figure out how to get the most from IQ as it is today, vs continually dialing up the speed on the camera tech hedonistic treadmill. A great way to guarantee happiness is to practice gratitude. A great way to guarantee misery is to practice dissatisfaction from things you can't control.

 

[tbcass]

• Improve color science?

Yes (but I do not like the term "color science").

Agreed. The term "color science" is wrong when it pertains to the color produced by camera JPEGs. Fortunately I see it used less these days. In my opinion it should have been "improve the color of camera produced JPEGs". If you shoot RAW the color is under your control.