Are cameras as good as they ever need to be?

[robfosters]

Looking at images nowadays, it seems as though many are so sharp and of very high quality, that it would virtually be impossible to improve on them.

I haven’t got a trained eye by any means, but when I look at images from cameras like the 5D, it just seems hard to imagine how they could be better in regards to equipment rather than photographer ability. I appreciate camera companies will always be developing new technology to keep their cameras selling, but how long before we’re buying cameras that in essence, produce no better images than the camera we bought 10 years ago?

It’s like megapixels. Even mid range cameras like the 800D have over 24 megapixels. That’s enough for very large prints and is absolutely ample for on screen viewing. Let’s face it, the huge 50 megapixel counts will start to trickle down to the mid range in due course too.

Surely it won’t be long before we can buy a camera safe in the knowledge that the image quality will not be bettered and will be everything you need to possibly last you a lifetime? We’re probably already in that place to some degree.

 

[Raymond Lin]

Image quality hasn’t been an issue for years, it kind of hit a plateau around the Nikon D800, what the technology now pushing towards now is stabilisation, focusing and size.

As a viewer the years goes back even longer, I don’t people who don’t take photos care, if they look at a magazine it all look the same for the last few decades since images turned colour. The difference and improvement is for the photographer and how easier it is to get that photo.

The difference between trying to get a correctly exposed and focus photo with a older DSLR vs a Sony A9 is staggering. There are a lot more tools and better tools inside the newer bodies than to help you acquire that photo.

And that process will continue to improve, it’s not perfect yet, the EVF can be better, lag can be improved, etc.

 

[SDK^]

Further development in the following areas will help improve image quality

- Focus systems on moving subjects
- More dynamic range
- High ISO noise
- In body stabilization - less camera shake

 

[Raymond Lin]

Further development in the following areas will help improve image quality

- Focus systems on moving subjects
- More dynamic range
- High ISO noise
- In body stabilization - less camera shake

Sony A7III is doing all that.

 

[SDK^]

Sony A7III is doing all that.

Yep - but it's not perfect, there will be more improvements to be made.

 

[Raymond Lin]

Yep - but it's not perfect, there will be more improvements to be made.

It'll never be perfect, look at the A7III and look at the Canon D30 (not 30D). It has come a long way. If you've never tried the AF on the new Sony A9/A7III camera, its hard to describe it.

 

[Andrew_McP]

Until my pocketable camera has a 100x zoom, can refocus and adjust DoF after I've taken the shot, takes decent shots by starlight, remove my thumb from the occasional shot, and can guess where I actually meant to point the darned thing in the first place, I won't be happy.

I'd rather not pay a million quid and wait 200 years for it either.

I guess at a push I'd settle for my DSLR's image quality in my existing pocket camera, preferably without phone style processing which magics suspiciously high quality out of a tiny sensor. Not entirely sure how physics will allow that, so I may have to settle for magic.

 

[SexyGreyFox]

I don't come in here often but it came up on my new threads so thought I'd share an experience I've just had.

A couple of weeks ago a lad came to one of my gigs and took 1,870 pictures, average size of 5 meg each (about 9.5gb all together) on a Canon 5 something camera which I was told cost around £3,000.
It was also obvious from the amount, that he had the camera on burst shooting for a large portion of them because if I keep my finger pressed down on my image viewer it's like a film.
At least 75% of them are out of focus which amazed me to be honest because I thought a camera of that quality would be point & click and be excellent on every take!
Or perhaps he had it set up wrong.

 

[Raymond Lin]

I don't come in here often but it came up on my new threads so thought I'd share an experience I've just had.

A couple of weeks ago a lad came to one of my gigs and took 1,870 pictures, average size of 5 meg each (about 9.5gb all together) on a Canon 5 something camera which I was told cost around £3,000.
It was also obvious from the amount, that he had the camera on burst shooting for a large portion of them because if I keep my finger pressed down on my image viewer it's like a film.
At least 75% of them are out of focus which amazed me to be honest because I thought a camera of that quality would be point & click and be excellent on every take!
Or perhaps he had it set up wrong.

Low light, harder to focus
People move
Low light, needs faster lens and high ISO and shutter speed or motion blur
He probably shot wide open so if you move it'll be out of the DOF and out of focus.

25% hit rate is pretty decent tbh and not sure why he would give you all of them, he should have not shown people the ones that is destined for the bin.

 

[Greebo]

There is one last major step forward in cameras that will make all current cameras obsolete.

Organic global shutter sensor.

30 dynamic range
No mechanical shutter
Built in ND filter
Pixel by pixel noise reduction

https://news.panasonic.com/global/press/data/2018/02/en180214-2/en180214-2.html

 

[Minstadave]

Be interesting to see how cameras integrate computational photography. I think this will be an area of development in the next few years - both frame a fake bokeh point of view but also multi shot HDR/hi-res ability, especially if anyone can get it working well without a tripod.

Organic sensors have been promised for years, interesting to see how much they’ll cost!

 

[voidshatter]

You'll never get too many number of pixels (provided that each pixel is sharp/clear enough to retain details, i.e. diffraction, dynamic range etc). Keep in mind that even a small iPhone has a function called "pinch zoom".

 

[DeathStorm]

Further development in the following areas will help improve image quality

- Focus systems on moving subjects
- More dynamic range
- High ISO noise
- In body stabilization - less camera shake

Given how well current cameras do these I'd rather they focused on improving other areas.

 

[i know nothing]

If you've never tried the AF on the new Sony A9/A7III camera, its hard to describe it.

Not sure if your comment is negative or positive, but I'll ask you to elucidate if I may as the subject interests me! How would you describe the AF? Is it good or not? I haven't had an opportunity to try it.

 

[Minstadave]

Not sure if your comment is negative or positive, but I'll ask you to elucidate if I may as the subject interests me! How would you describe the AF? Is it good or not? I haven't had an opportunity to try it.

It’s meant to be stonkingly good. The eye AF even in C-AF has got a lot of praise.

 

[i know nothing]

It’s meant to be stonkingly good. The eye AF even in C-AF has got a lot of praise.

I'll hold my hand up and say that if I were to consider switching from Nikon to Sony, 99% of the reason would be the eye AF. I won't be switching anytime yet as I've only had my current camera for a year, but I do like to keep abreast of what's going on.

 

[Greebo]

Be interesting to see how cameras integrate computational photography. I think this will be an area of development in the next few years - both frame a fake bokeh point of view but also multi shot HDR/hi-res ability, especially if anyone can get it working well without a tripod.

Organic sensors have been promised for years, interesting to see how much they’ll cost!

Well Panasonic will have their first camera out ready for the next Olympics. No idea when Sony will have theirs out but they announced their organic sensor the same time as Panasonic.

In fact all the major cameras makers are working on this as it is the future.

No doubt the first one will be a premium but they will come down in price. I bet in 5 to10 years even mobiles will come with organic sensors.

 

[SexyGreyFox]

25% hit rate is pretty decent tbh and not sure why he would give you all of them, he should have not shown people the ones that is destined for the bin.

At least I have an explanation now, it seems my Note 3 has a better hit rate.
The lad is a relative of the bass player and he just handed everything over, he obviously didn't want to go through 1800 shots.

 

[D.P.]

Be interesting to see how cameras integrate computational photography. I think this will be an area of development in the next few years - both frame a fake bokeh point of view but also multi shot HDR/hi-res ability, especially if anyone can get it working well without a tripod.

Organic sensors have been promised for years, interesting to see how much they’ll cost!

This is the main thing going forwards. At the lens and sensor level we are pretty close to physical limits of performance. For example, for the people asking about better high-ISO performance, with modern sensor each pixel is basically counting individual photons in the dark bands,. One pixel might get 5 photos and an adjact pixel 7, and that will be bumped to a higher shade of grey. Hpwever, there is a long way to go with computational photography, and deep-learning is only just being applied. Future high-iso improvements wont come form the sensor but form advanced algorithms doing noise removal and feature extraction.

Similar with resolution. Modern sensor are close to the diffraction limits of green light at f/11. While oversampling is useful, ultimately to get a higher resolution we will have to move to bigger sensor, or use computation. Either multiple frames can be combined in succession or deep learning can be sued to fill in details as a form of advanced interpolation.

the future is in merging data form multiple exposures to increase resolution, reduce noise , increase dynamic range, increase (or decrease) DoF, generate 3D depth maps, etc.




There are a few possible sensor improvements, but often these involve a trade-off. iF surface area is used for something other than direct capture of photos then you lower resolution/DR/noise performance. thus to really do some of these Justice you might want hot-swappable sensors:
1) Black and white sensors. Around 2/3rds of the photos hitting the sensor are thrown away at the colour filter. We get left with a lower color spatial resolution which then needs to be interpolated to full color, and when we want black and white photos converted all the way back. While capture color photos and then converting to B&W has advantages, a true B&W camera would offer much better high-ISO, DR and sharpness.
2) As above. Since each pixel is only measuring a single color channel, much light is wasted. sigma Foveon sensors try to capture the full color spectrum at each pixel. but the technology is just not that mature yet and has some significant drawbacks.
3) IR pixels. Sensors are very sensitive to IR light, so normally comes with an IR filter. For people that shoot IR it is ncie to gert the filter removed. Ideally this could be done on the fly. Moreover, Ir can be combined with visible light in very interesting ways. IR light tends not to be affected by haze, so can be used to give much higher contrast images in the summer,. E.g., distant mountains can be made sharper and clear if you sampled IR at certain pixel frequencies.
4) High DR: A big limit of digital sensor is the lack of latitude in the upper exposures. Highlights get clipped very easily, with less room than decent film. That is because film has a logarithmic response because there were smaller and smaller silver halide particles that were less and elss sensitive. With digital every pixel is equal and gets saturated quickly. his is easily solved by having smaller pixels and larger electron wells. Fuji uses to do this when they made DSLRs, having special pixels to capture highlight detail, the camera were about 2 stops better than canon/nikon. It would also be possible to have a logarithmic ADC.