How will advances in digital Technology change how exposures are made ?

[Pookeyhead]

It all depends on what you're doing, and what equipment you invest in. I'm finding print costs around the same now as they were then. Inks for a Epson 4880 are around £60 for each cartridge, and it takes 8 cartridges... 9 if you regularly swap the blacks around. A2+ papers are not cheap either. I reckon I spent less per month when I shot film than I did when I was still working flat out with digital. My costs have dropped dramatically now as I don't work commercially anywhere near as much these days... but it's still a hefty monthly expense.

 

[Terrywoodenpic]

I see exactly why you made that mistake Steve.

Terry - Would a different title, to reflect the matter under discussion, be useful?

If someone could come up with a suitabe title. I am sure a mod could change it.

But I think it is nearing the end of its run....

 

[simon ess]

You can change it yourself Terry.

Go to "edit" in your first post, click on "advanced", change away.

And I very much doubt it's nearing the end of its run. It's just the start.

 

[Mikesphotaes]

For the most part people who shoot digital have never shot film and don't care about film in any way shape or form.

Indeed, an accurate post at long last!

 

[simon ess]

Indeed, an accurate post at long last!

Well, partly perhaps.

 

[Terrywoodenpic]

Well I usually take around 10,000 pics a year and I use a very simple editing program which cost me £15.00 about 6 years ago.

I have just replaced my old computer with an i7 self build which cost about £600.00.

I seldom print a paper copy, except for those I wish to hang on my walls so after approx 6 years of taking pics I think I am well in.

Back in the days when film reigned supreme, film, cameras, lenses, darkrooms, darkroom equipment and all the usables like chemicals etc, were probably a lot more than cameras and computer equipment are now after adjusting for inflation etc.
.

If you go back to just before Digital took off, Film and material were cheap .
It is hard to quantify the price of cameras and other equipment as they were far less High tec and lenses were less sophisticated but more robust. having films processed and printed was cheap as chips.

Professional prices have always been at a different level, but costs are certainly much higher today.

Some people who want to scrimp can keep costs very low, but serious work costs..... as always.

 

[Terrywoodenpic]

You can change it yourself Terry.

Go to "edit" in your first post, click on "advanced", change away.

And I very much doubt it's nearing the end of its run. It's just the start.

I am stuck in a groove....any suggestions would be welcome.

 

[simon ess]

I am stuck in a groove....

Vinyl - cassette - CD - MP3 - Flac streaming - ?

Glass plate - film - DSLR - ?

 

[Terrywoodenpic]

For the most part people who shoot digital have never shot film and don't care about film in any way shape or form.

`

An interesting point....

However I doubt it is true.

Digital has been going in a serious way for about Five or six years.
Many photographers start with their first half decent camera in their early teens. I have been taking photography seriously for 67 years.

the chances are something between the two, say 25 years of photography is a reasonable average.

This would mean that photographers that started on film would make up a large percentage of todays users.

Very very few who started on film gave up photography when Digital came along.

 

[artyman]

Whilst perhaps from a technical point film ISO and grain does not directly equate with Sensor ISO and noise, but coming from the distant past with developing trays and such esoteric items, it is a good enough analogy for me. Basic exposure principles haven't really changed, grain, I mean noise is improving all the time in the way film improved from it's early beginnings.

I now can shoot at ISO's I could only dream of when I was a lad, happy days.

 

[Terrywoodenpic]

Whilst perhaps from a technical point film ISO and grain does not directly equate with Sensor ISO and noise, but coming from the distant past with developing trays and such esoteric items, it is a good enough analogy for me. Basic exposure principles haven't really changed, grain, I mean noise is improving all the time in the way film improved from it's early beginnings.

I now can shoot at ISO's I could only dream of when I was a lad, happy days.

Not wrong... But...
We see it that way, because the equivalence you see has been deliberately put in place to make it "Easy" to transfer skills from one medium to the other.

That was far easier a marketing strategy than making people cope with the changed situation. In any event, the early sensors found it hard to cope even with a restricted range of tones. at any ISO setting.

Today that is no longer true and we are getting to the stage were an ISO free strategy is becoming realistic over 5 or so stops.
In this debate we have supposed that this trend will continue ( I am sure it will). This will lead to the point where that ISO free way of working will be so advantageous, that it will be worth the effort to take on board new working practises, and dump "Film" thinking altogether.

It is "Your Dream" of achieving virtually noise free high ISO's that makes this possible.

 

[petersmart]

If you go back to just before Digital took off, Film and material were cheap .
It is hard to quantify the price of cameras and other equipment as they were far less High tec and lenses were less sophisticated but more robust. having films processed and printed was cheap as chips.

Professional prices have always been at a different level, but costs are certainly much higher today.

Some people who want to scrimp can keep costs very low, but serious work costs..... as always.

Well I wasn't talking about a few years ago I was talking about 50 years ago.

At that time I was, I believe, earning something like £6.00 per week.

I decided to go for a better camera than I had and went to the local phot shop.

I was going for a Rolleiflex until I saw this really incredible camera - the Mamiya C2 - TLR, interchangeable lenses and incredible quality for the day.

The cost was £80.00 which I funded at £2.00 per month over about 2 years.

Now compare my wage with the cost of that camera and wages today - if we say the average wage is £350/week then that camera today would cost £3500 - and that was S/H!

And don't talk about Hasselblads - at comparable prices today they would be in the tens of thousands!

Prices are actually CHEAPER today than then in everything!

And as for being less high tech - well we didn't have all the buttons etc but then we didn't need them - I had a Weston Photo meter for exposure which worked excellently as long as you used it correctly and a film the ASA (not ISO) speed of which was built in.

I worked with 3 main films - Ectachrome for 2 1/4 inch slides, Pan-X for studio work and Tri-X for weddings.

Tri-X was rated at 400ASA but I overexposed by 2 stops for bright sunlight and cut the developing time by approx half to reduce the contrast and get full detail in Bride's white dresses and Groom's dark suits.

On dull days it was used at full rating and processed normally.

Nobody really worried about grain too much because with a normal 10x8 or 12x10 print there was no grain even after cropping.

And of course we were always taught to fill the frame as much as possible for the best quality.

And I must say that it was much easier that way because all the extra buttons and adjustments these days mean that it is more difficult to really learn to use a camera to the best.

Of course these days there are point and shoot cameras that do almost everything for you - but we had those as well - they were called Box Brownies and later on - Instamatics.

.

 

[Terrywoodenpic]

Now a little more “Anal”


I think a few posters are a little irritated when I bang on about “Photons”
I do it for a reason...
Digital is all about bean counting... and photons are the beans it counts.
OK, there is not just one variety of “Beans”. They come in a range or energies (Colours). Sensors resort to analogue colour filters in the form of Bayer screens, or the like, to sort them into their various energies.
No problem there then?
Well there is... any one who has used a camera filter knows that it absorbs light, and that you need to compensate by increasing exposure to regain the “lost” light. (modern TTL cameras sort that for you)

As yet I have heard of no digital sensor that can sort Photons by their energies(colours), with out an analogue filter.
Of course this only matters when the availability of photons are at a premium... when it is dark.

Our vision, and that of most other animals, has to a large extent sorted that problem out. We have in effect a dual layer sensor consisting of Rods and Cones.
At low intensity the Rods take over the function of vision. They do however work at a different energy level, which equates to the violet to yellow area of the spectrum with peak sensitivity in the Green. This is not by chance, as it matches the general colour of natural light at night. A rod can respond to a single photon.
So in this way our eyes take maximum advantage of the conditions, but at the cost of being able to recognise colour.

There are secondary effects, caused by the preponderance of cones being in the periphery of our vision. The main one being that night vision is not very acute ( you would need much higher density of rods at the centre, for it to be so) The second is that it is highly sensitive to noticing movement at the periphery.

What we actually see, after our eyes have adjusted to the dark, is very different to the effects seen with night vision glasses. We do not see “Speckles” (well most of the time).

The three types of cones take care of vision at normal light levels, each responding to photons in a different energy range.

The result is that at “any light level” we have an adaptation that gives us the best compromise of sensitivity, colour and acuity. And we do not even have to change an ISO setting.

Other animals have adaptations that have come to different compromises. Compare eagles with deep-water fish.

So what is this to do with Digital sensors or Photography?
It shows that there might well be a different solution to both wide tonal ranges and poor low light quality.
Even if it is a micro biological one.

 

[petersmart]

The result is that at “any light level” we have an adaptation that gives us the best compromise of sensitivity, colour and acuity. And we do not even have to change an ISO setting.

We do, actually, it's called an iris.

.

 

[Terrywoodenpic]

Well I wasn't talking about a few years ago I was talking about 50 years ago.

At that time I was, I believe, earning something like £6.00 per week.

I decided to go for a better camera than I had and went to the local phot shop.

I was going for a Rolleiflex until I saw this really incredible camera - the Mamiya C2 - TLR, interchangeable lenses and incredible quality for the day.

The cost was £80.00 which I funded at £2.00 per month over about 2 years.

Now compare my wage with the cost of that camera and wages today - if we say the average wage is £350/week then that camera today would cost £3500 - and that was S/H!

And don't talk about Hasselblads - at comparable prices today they would be in the tens of thousands!

Prices are actually CHEAPER today than then in everything!

And as for being less high tech - well we didn't have all the buttons etc but then we didn't need them - I had a Weston Photo meter for exposure which worked excellently as long as you used it correctly and a film the ASA (not ISO) speed of which was built in.

I worked with 3 main films - Ectachrome for 2 1/4 inch slides, Pan-X for studio work and Tri-X for weddings.

Tri-X was rated at 400ASA but I overexposed by 2 stops for bright sunlight and cut the developing time by approx half to reduce the contrast and get full detail in Bride's white dresses and Groom's dark suits.

On dull days it was used at full rating and processed normally.

Nobody really worried about grain too much because with a normal 10x8 or 12x10 print there was no grain even after cropping.

And of course we were always taught to fill the frame as much as possible for the best quality.

And I must say that it was much easier that way because all the extra buttons and adjustments these days mean that it is more difficult to really learn to use a camera to the best.

Of course these days there are point and shoot cameras that do almost everything for you - but we had those as well - they were called Box Brownies and later on - Instamatics.

.

Sounds like you are nearer my generation of photographers... Though I started working with glass plates and mamiya's were a long way in the future.

The biggest difference between then and now was, Labour was cheap and things expensive. It is very much the other way round now.

But strange to tell, I never had any trouble managing on my first wage of a generous £8 a week. (1952) before that I worked part time for free.

 

[Terrywoodenpic]

We do, actually, it's called an iris.

.

Not so.... the Iris is more equavent to aperture, though our Iris adjusts more to contrast than light level.

We do not have the equivalent to ISO. or film speed.

 

[petersmart]

Sounds like you are nearer my generation of photographers... Though I started working with glass plates and mamiya's were a long way in the future.

The biggest difference between then and now was, Labour was cheap and things expensive. It is very much the other way round now.

But strange to tell, I never had any trouble managing on my first wage of a generous £8 a week. (1952) before that I worked part time for free.

Well I'm 68 and my first wages were about £4.00+

And I bought my Mamiya when I was 18.

.

 

[Terrywoodenpic]

Well I'm 68 and my first wages were about £4.00+

And I bought my Mamiya when I was 18.

.

I was ten years ahead of you...... and on about £20 or so by then.
Seemed good money at the time.

Though as a photo student working at Bultins as a "brown coat" photographer in the holidays (1956), I came home with over £600 in a five week stint, all in white fivers.

 

[HoppyUK]

I've been following with interest, but I'm not sure where this thread is heading. So auto-ISO is a good thing, and it's getting better. Well yes, excellent. We're already using cameras in this way and no doubt ISO will find its way to greater prominence as a direct access control quite soon. It just needs another knob.

I see no revolution there, nor any new way of working any time soon. I don't accept the premise that somehow digital photography is stuck in a rut created by a heritage of film. This seems to be fundamental assumption of the thread which is somehow holding things back. I disagree. We use cameras the way we do, largely in the same way as film, because early digital was rubbish at higher ISOs so there was no alternative. It's still a good way, and further developments and improvements will make things better, but not turn everything on its head.

The sensor is only part of the imaging chain, and modern cameras work the way they do because they've evolved into a marriage with the lenses we use - and they're not changing any time soon, not in the next ten years anyway, not fundamentally. Current DSLR sensor sizes combined with focal lengths and apertures give us the creative control we want and need. Also, we as enthusiasts love our cameras the way they are. Technology may change, sometimes fast, but old habits die harder. The current challenge is to get CSCs to focus-track moving subjects as well, or better, than DSLRs - that, and a properly solid state shutter. Crack those things and happy days all round for both manufacturers and consumers.

High ISO is great, and really usable ISO100,000 is surely not far away, but who needs more? We can already shoot in light almost too dark to see by, and then the quality of light from artificial sources is usually poor. Flash works, LEDs have great potential. If there is no market demand, ie no major consumer benefit, there will be no return on business investment and technological development will stall. I'm not seeing this upside.

 

[Terrywoodenpic]

I've been following with interest, but I'm not sure where this thread is heading. So auto-ISO is a good thing, and it's getting better. Well yes, excellent. We're already using cameras in this way and no doubt ISO will find its way to greater prominence as a direct access control quite soon. It just needs another knob.

This may be the route they take...Later they will remove the optional settings or simply introduce high range, normal and low range (with no mention of ISO)and give the ability to set any combination of aperture and shutter speed with in each of those ranges, for image rather than exposure control..

The High range would be noiseless as might the normal range. However in the Low range noise would be present but acceptable.

I see no revolution there, nor any new way of working any time soon. It will probably sooner than you think

It would seem different to me and more useful

The sensor is only part of the imaging chain, and modern cameras work the way they do because they've evolved into a marriage with the lenses we use - and they're not changing any time soon, not in the next ten years anyway, not fundamentally. Current DSLR sensor sizes combined with focal lengths and apertures give us the creative control we want and need. Also, we as enthusiasts love our cameras the way they are.

That sounds rater reactionary in a future gazing thread... what we are talking about is freeing up exposure so that we can choose any aperture or shutter speed to give more creative control not less.

Technology may change, but old habits die harder. The current challenge is to get CSCs to focus-track moving subjects as well, or better, than DSLRs - that, and a properly solid state shutter. Crack that and happy days all round for both manufacturers and consumers.

That is a fresh question, and quite interesting. The question of focus in a non DSLR is taking up the efforts of every manufacturer right now, and has improved beyond recognition in the last two years.
Shutters will certainly change. perhaps the they will not bother to stop the light, simply stop recording it after the required time.

High ISO is great, and really usable ISO100,000 is surely not far away, but who needs more? We can already shoot in light almost too dark to see by, and then the quality of light from artificial sources is usually poor. Flash works, LEDs have great potential. If there is no market demand, ie no major consumer benefit, there will be no return on business investment and technological development will stall. I'm not seeing this upside.

Some cameras can already take usable shots in moonlight, My previous posts show why the results will always be lower than optimal quality. There simply are not enough photons unless you are prepared to wait for a sufficient number of them to arrive.

The quality of using artificial light sources is not in question, The convenience of making a suitable set up certainly can be, so we make suboptimal choices.

 

[HoppyUK]

<snip>

That sounds rater reactionary in a future gazing thread... what we are talking about is freeing up exposure so that we can choose any aperture or shutter speed to give more creative control not less.

<snip>

That's what people always say when you don't share their vision

But I don't see how this differs from the concept of auto-ISO, and that's no more than an evolution of where we are now.

Where's the big deal? Where's the the fundamental change? What's the relevance of film and photon capture and the 'illusion' of ISO? What is the revolutionary consumer benefit? Freeing up shutter speed and aperture selection is great, no question, and a good step forward. But a shift of emphasis from where we are now, a shift that is already happening, rather than a complete change that will allow us to do something totally new and different.

 

[Terrywoodenpic]

That's what people always say when you don't share their vision

But I don't see how this differs from the concept of auto-ISO, and that's no more than an evolution of where we are now.

Where's the big deal? Where's the the fundamental change? What's the relevance of film and photon capture and the 'illusion' of ISO? What is the revolutionary consumer benefit? Freeing up shutter speed and aperture selection is great, no question, and a good step forward. But a shift of emphasis from where we are now, a shift that is already happening, rather than a complete change that will allow us to do something totally new and different.

The fact that we are even having to talk about ISO or auto ISO is the problem.
It is very liberating to divorce shutter speed and aperture from the light gathering process. It is only the concept of a sensor having a variable sensitivity (when it does not) that ties us to thinking in terms of ISO.

Many beginners ( and I have taught many) have great difficulty linking shutter speeds, apertures and "Film speeds" as an interrelated whole.
In future they will not need to.
Shutter speed will only relate to stopping power, and aperture to depth of field (+sweet spot). This having to increase one if you reduce the other nonsense will be things of the past.

 

[POAH]

The fact that we are even having to talk about ISO or auto ISO is the problem.
It is very liberating to divorce shutter speed and aperture from the light gathering process. It is only the concept of a sensor having a variable sensitivity (when it does not) that ties us to thinking in terms of ISO.

Why? How else to you define the gain used in amplification. It's part if the system we use so why not talk about it? If you can't get your head round the simple maths involved in exposure you should go back to primary school.

 

[simon ess]

Crikey! You do seem to have a talent for unnecessarily confrontational language.

Is that deliberate, or just unfortunate?

 

[Terrywoodenpic]

Why? How else to you define the gain used in amplification. It's part if the system we use so why not talk about it? If you can't get your head round the simple maths involved in exposure you should go back to primary school.

Not every one has a distinction in additional math (I did) for some simple math is difficult.
we have covered amplification earlier.
The ISO in a sensor is not the product of variable gain. It is a product of Firmware settings on the digital output of that sensor. The sensor has a single sensitivity.

(I suspect, but don't know)some early sensors might have tried variable gain, but it is excessively noisy. (more like night vision glasses.)

 

[HoppyUK]

Not every one has a distinction in additional math (I did) for some simple math is difficult.
we have covered amplification earlier.
The ISO in a sensor is not the product of variable gain. It is a product of Firmware settings on the digital output of that sensor. The sensor has a single sensitivity.

(I suspect, but don't know)some early sensors might have tried variable gain, but it is excessively noisy. (more like night vision glasses.)

Sorry to be blunt Terry, but I think you're making too much of all this and quoting some quite esoteric science and frankly irrelevant film history to make a simple point about improving high ISO performance.

This is just semantics: "The ISO in a sensor is not the product of variable gain. It is a product of Firmware settings on the digital output of that sensor. The sensor has a single sensitivity."

Agreed. But what's the point? The signal is modified ('gain' is a simply understood term that works for me) and since no system is ever 100% efficient there are downsides too. IQ downsides, cost, physical size, operating speed. All that is happening now is that those downsides are being reduced. There's nothing new in that, but as things improve we just draw the line a little higher up with each new generation.

PS Just to add, I always enjoy your posts, including this thread. You've just lost me a bit this time

 

[simon ess]

I agree with pretty much everything you say there Richard.

I just can't help feeling though that there's something more to what Terry's talking about.

I'm going with the flow at the minute to see where it goes.

 

[srichards]

One way they could take it is to increase the sensitivity of the sensor so that its base iso equivalent was more like 6400. You could then use artificial nd filters to reduce the light at 'lower' isos ie brighter conditions. If the base iso equivalent was 6400 then you have the opportunity to gain up to levels where just about any shutter speed/aperture combination was permissible.

That's where I want sensors to go. This whole notion of iso ie amount of noise you're going to get is a pain. The vast majority of people want iso 100 levels of noise at higher iso equivalents. If you want grainy and noisy images then you can diddle it with photoshop. It's much harder to de noise then it is to add noise in later.

The fuji has some ND filter thingy within it so the technology already exists. You could even have partial ND filters so that it could be the equivalent of grad filters or even ones that filtered bright parts to balance exposure automatically.

I think there are a lot of possibilities if you get rid if the idea of 'iso' and remove that leg of the exposure triangle altogether.

 

[LCPete]

Everyone talks about noise at high ISO but its the loss of dynamic range that's important
you can deal with noise but if the colours are not there you can't put them back
Having trouble explaining what I mean but if I go from ISO 800 to 3200 on a shot of a tiger for example the one at 800 looks nicer and it is not the noise that's the problem
Also if the light is not nice , eg you are in shadows it will never look the same even if the camera has a magic sensor

 

[Terrywoodenpic]

Everyone talks about noise at high ISO but its the loss of dynamic range that's important
you can deal with noise but if the colours are not there you can't put them back
Having trouble explaining what I mean but if I go from ISO 800 to 3200 on a shot of a tiger for example the one at 800 looks nicer and it is not the noise that's the problem
Also if the light is not nice , eg you are in shadows it will never look the same even if the camera has a magic sensor

Quite right
what we are talking about is light (photons). If there are not enough of them
you can't be invented them, the result is inevitably a degraded image.
Of course in the right context that is not a problem.. you do not expect shadows to contain bright colours. so the image looks right.

However in an under exposed image (one where there are too few photons to make up any of the tones fully) the whole image looks degraded.

Exactly the same result is obtained if the the illumination of the subject is too low or the Shutter speed and aperture did not pass enough enough photons to the sensor.

visible Noise is the result of under exposure , Noise is there in every exposure, but it is too small a proportion of the whole in well exposed shots to be noticeable.

The ISO is a red herring for your Tiger example

At ISO 800 it might be telling you to set 1/500 at f4 so at ISO 3200 you would set say 1/000 at f5.6 ( two stops difference.)

The sensor is then under exposed by two stops. But because the firmware knows that you have set ISO3200 it adjusts the output curve to compensate.

The sensor has still received two stops under exposure and you get all the consequences that go with that, lack of saturation and extra noise. The firmware kicks in and does it best to improve the situation by increasing the brightness and contrast and removing noise.

If you are shooting Raw the raw developer does most of this after reading the instructions in the picture file. You then have the option to take these adjustments further.

The important thing to think about is these ISO settings do nothing at the sensor level, they are all later adjustments, either in firmware or during raw processing.

Modern sensors are becoming very good indeed and capture a greater range of tones than are needed in a normal situation (as much as five stops more in some cases) this means that in the best cameras little difference is shown in the results in a wide range of ISO settings as they are making use of the extra "Headroom"
They are making use of the full straight line of the Raw curve. To take full advantage of this happy state of affairs, you must always expose to the right Just avoiding clipping the highlights.

In our Future gazing we have suggested that sensors and this head room, will get even greater, and reach the point that they can encompass the full range of tones from moonlight of full sun. If that is the case and with in certain limits, it will not matter what shutter speed or aperture you set, you will still obtain an optimal exposure. At that point the concept of ISO is meaningless.

 

[Terrywoodenpic]

These and other questions might be answered in the reply to LCPete above.

Sorry to be blunt Terry, but I think you're making too much of all this and quoting some quite esoteric science and frankly irrelevant film history to make a simple point about improving high ISO performance.

I have tried to keep away from science and talk on photographic terms (perhaps I have failed ) and this has not been about ISO performance at all.

This is just semantics: "The ISO in a sensor is not the product of variable gain. It is a product of Firmware settings on the digital output of that sensor. The sensor has a single sensitivity."

Agreed. But what's the point? The signal is modified ('gain' is a simply understood term that works for me) and since no system is ever 100% efficient there are downsides too. IQ downsides, cost, physical size, operating speed. All that is happening now is that those downsides are being reduced. There's nothing new in that, but as things improve we just draw the line a little higher up with each new generation.

Whilst there probable is some signal amplification (Gain) taking place in the sensor it is not for the purpose of ISO adjustment..

"when we adjust the ISO on an exposure meter, we simply move a scale against a reference point it does nothing physical. Firmware does much the same and then adjusts the S Curve"

PS Just to add, I always enjoy your posts, including this thread. You've just lost me a bit this time

I agree with pretty much everything you say there Richard.

I just can't help feeling though that there's something more to what Terry's talking about.

I'm going with the flow at the minute to see where it goes.

May be the answer to LCPete will help

One way they could take it is to increase the sensitivity of the sensor so that its base iso equivalent was more like 6400. You could then use artificial nd filters to reduce the light at 'lower' isos ie brighter conditions. If the base iso equivalent was 6400 then you have the opportunity to gain up to levels where just about any shutter speed/aperture combination was permissible.

That's where I want sensors to go. This whole notion of iso ie amount of noise you're going to get is a pain. The vast majority of people want iso 100 levels of noise at higher iso equivalents. If you want grainy and noisy images then you can diddle it with photoshop. It's much harder to de noise then it is to add noise in later.

The fuji has some ND filter thingy within it so the technology already exists. You could even have partial ND filters so that it could be the equivalent of grad filters or even ones that filtered bright parts to balance exposure automatically.

I think there are a lot of possibilities if you get rid if the idea of 'iso' and remove that leg of the exposure triangle altogether.

ISO sort of cuts exposure into segments, not real ones, but points of reference.
What we need from a sensor is the ability to capture, in any single exposure, a light input range from very bright to very dark.
Some cameras are not far off that ability. When they can, the ISO concept is redundant.

 

[POAH]

so the basic jist of this thread is to have hardware and software than allows you to set aperture and the shutter speed you want and the camera sets the gain on the sensor to give you the correct exposure with the lowest noise or auto ISO as its known at the moment.

 

[petersmart]

Unfortunately, with regards to noise, all electrical circuits generate noise and there is nothing that can be done about it, except to try and design the lowest noise circuits possible.

And as you increase the amplification factor (or gain) ALL the noise in any circuit is amplified not just the sensor noise.

To hear this just turn up your PC speakers or radio or CD player with no other signal coming through and you will hear that familiar hiss.

And sensors are no different, so I doubt if we will ever see noise free sensors.

In fact the more MPs we have the more noise is generated due to a reduction in size of the photodiodes used, but this can be offset by a reduction in magnification due to the larger size images.

So really all we can do is make the most of what we have.

Personally I have no problems with using 3200 ISO when I need to and then reducing the noise in Neat Image which gives very good results.

And as for sensors with vastly increased dynamic range, what good is that?

After all we can produce images already using HDR with much increased dynamic range but they then have to be tone mapped to bring them with the range of our monitors.

And even if we could vastly increase the dynamic range of all our equipment we still come up against our own biological failures - our eyes simply cannot accommodate such a vast range of tones in a single area.

for instance when we come indoors from a bright summers day our eyes take a few seconds to adapt to the darker interior and vice versa.

I have a 40" LED tv and I have noticed that in scenes with a bright area and darker areas my eyes close down a fraction to accommodate the bright area resulting in other areas appearing darker.

So unless and until someone comes up with a MKII or MKIII eye arrangement, in the final analysis it really would be a waste of time to spend vast sums trying to produce your ideal sensor.

.

 

[Terrywoodenpic]

so the basic jist of this thread is to have hardware and software than allows you to set aperture and the shutter speed you want and the camera sets the gain on the sensor to give you the correct exposure with the lowest noise or auto ISO as its known at the moment.

The Gain on the Sensor is never changed. The output might be.

 

[Terrywoodenpic]

From an pratical example on the DXO site

We will now make matters worse by simulating a wedding photographer shooting with only ambient light in a dimly lit castle. Say this requires boosting the ISO from say 100 to 3200 ISO (see Photo 3). This means that we are underexposing the sensor by 32×!

Boosting ISO settings on a digital camera merely underexposes
the sensor, and cranks up the resulting image/signal by analog
amplification or by digital multiplication.

They already told you that, right?

 

[Terrywoodenpic]

Unfortunately, with regards to noise, all electrical circuits generate noise and there is nothing that can be done about it, except to try and design the lowest noise circuits possible.

And as you increase the amplification factor (or gain) ALL the noise in any circuit is amplified not just the sensor noise.

To hear this just turn up your PC speakers or radio or CD player with no other signal coming through and you will hear that familiar hiss.

And sensors are no different, so I doubt if we will ever see noise free sensors.

In fact the more MPs we have the more noise is generated due to a reduction in size of the photodiodes used, but this can be offset by a reduction in magnification due to the larger size images.

So really all we can do is make the most of what we have.

Personally I have no problems with using 3200 ISO when I need to and then reducing the noise in Neat Image which gives very good results.

And as for sensors with vastly increased dynamic range, what good is that?

After all we can produce images already using HDR with much increased dynamic range but they then have to be tone mapped to bring them with the range of our monitors.

And even if we could vastly increase the dynamic range of all our equipment we still come up against our own biological failures - our eyes simply cannot accommodate such a vast range of tones in a single area.

for instance when we come indoors from a bright summers day our eyes take a few seconds to adapt to the darker interior and vice versa.

I have a 40" LED tv and I have noticed that in scenes with a bright area and darker areas my eyes close down a fraction to accommodate the bright area resulting in other areas appearing darker.

So unless and until someone comes up with a MKII or MKIII eye arrangement, in the final analysis it really would be a waste of time to spend vast sums trying to produce your ideal sensor.

.

We are covering old ground...

The new point is Your HDR comment.
Every scene we view is high dynamic range.
Every time we take a photograph we make a selection from that wide tonal range, to enable us to capture that part of it, that seems important to us. We do this because of the limited output tonal range we are able to produce on our monitors or on our prints.

The rather Nasty HDR programs allow us to be selective and distort the out put, so that we can enhance areas of tone in an unrealistic way.

In contrast what I am suggesting is that the stopping power of the shutter and depth of field of the aperture need not be tied together as they are now. But that the tonal range captured by the sensor will be wide enough to allow these to be set freely.

I am not suggesting that we would be able to see or use this entire range as a final output, but we would be able to capture it for later selection.( as we do with our more limited range now.)

I have avoided all discussion about the types of Noise. Shot noise is the most random and unavoidable. We do not use variable gain like in a radio, and so signal noise is somewhat predictable and controllable with the existing algorithms. Noise control will improve but not by much.

See the DXO comment above.

 

[POAH]

I don't take any notice of DXO (not that I'm disagreeing)

From an pratical example on the DXO site

 

[petersmart]

Every scene we view is high dynamic range.

Not so - a scene on a dull day can easily be accommodated by current sensors - to see this simply view the histogram on your camera.

But that the tonal range captured by the sensor will be wide enough to allow these to be set freely.

Not so - the output from a photodiode will always be limited by its saturation level - the level at which no further output is possible no matter how much extra light is available.

And at the other end the output is so low that it cannot be distinguished from the ambient noise - which increases with increased temperature.

These are the two extremes which limit the dynamic range of a single photodiode and thus the dynamic range of a camera.

And if we go by DXO's definition then ALL camera sensors are underexposing ALL the time because the output from the photodiodes is so low it must be amplified before being digitised and stored in the camera's memory card.

.

 

[Terrywoodenpic]

Not so - a scene on a dull day can easily be accommodated by current sensors - to see this simply view the histogram on your camera.

Of course that is true of photographing anything with a limited tonal range... But it always captures as wide a tonal range as it can.

Not so - the output from a photodiode will always be limited by its saturation level - the level at which no further output is possible no matter how much extra light is available.

Of course it is, but we are future gazing and they will have far higher saturation capacities in the future. (all stated previously.)

And at the other end the output is so low that it cannot be distinguished from the ambient noise - which increases with increased temperature.

I have stated that many times ...It is the input which is the problem not the output... there are no photons to count....

These are the two extremes which limit the dynamic range of a single photodiode and thus the dynamic range of a camera.

And that as well....

And if we go by DXO's definition then ALL camera sensors are underexposing ALL the time because the output from the photodiodes is so low it must be amplified before being digitised and stored in the camera's memory card.

That is not what they are talking about... they are talking about input (exposure)
sensor output is always amplified. (But not variably) see all earlier posts.

What they say is quite correct. It is very similar to the way Professor Newman describes the process.

 

[Terrywoodenpic]

I don't take any notice of DXO (not that I'm disagreeing)

DXO are fairly good on science matters, and although I don't use them, their products are pretty good too... Just rather expensive for non pro use.