Post processing images and HDR

[Sagitar]

So basically point the camera at subject, take photo, and that's it.

 

But that conveniently ignores all the in-camera processing that goes on between pressing the button and seeing the image on the viewing screen. I think I understand well what Tony is saying and I have tried very hard to show that he is setting up a false dichotomy, but clearly I have failed.

 

 

the world's best photographers are the best because they can pretty much get an amazing shot straight away (I'm sure they still play with them but the initial image will still be stunning).

 

I think you could not be more wrong on this. Great artistic photographers undoubtedly "see" differently that the rest of us and recognise more readily the potential in a subject. But they also work a lot harder at what they do than most of us are prepared to. To suggest that they don't give enormous attention to the processing of their pictures is simply not right. I don't know how much you have read about the methods of working used by famous photographers, but if you haven't dipped into the topic, I suggest a look at the pictures and the writings of Ansell Adams. I think you might be surprised at the level of attention that he gave to the production of prints of his photographs.

[Rich]

So basically point the camera at subject, take photo, and that's it.

 

But that conveniently ignores all the in-camera processing that goes on between pressing the button and seeing the image on the viewing screen.

 

Exactly, it seems we're still going around in circles though...I don't know what's difficult to understand about it. With a point and shoot camera the image produced isn't what you took, it's what the camera has processed for you based on how it's been designed to do it. With a DSLR and shooting in RAW that will be the 'real' photo as the camera hasn't processed it.

 

 

the world's best photographers are the best because they can pretty much get an amazing shot straight away (I'm sure they still play with them but the initial image will still be stunning).

 

I think you could not be more wrong on this. Great artistic photographers undoubtedly "see" differently that the rest of us and recognise more readily the potential in a subject. But they also work a lot harder at what they do than most of us are prepared to. To suggest that they don't give enormous attention to the processing of their pictures is simply not right. I don't know how much you have read about the methods of working used by famous photographers, but if you haven't dipped into the topic, I suggest a look at the pictures and the writings of Ansell Adams. I think you might be surprised at the level of attention that he gave to the production of prints of his photographs.

 

You need to see what goes on in a production studio when it comes to processing photos, alot of it goes on. Not so much in newspapers due to time constraints and the fast turnaround needed. Your monthly magazine will be a different story though, same for photos used in adverts. I've not heard of Ansell Adams but going to have a look later.

 

I've noticed Tony hasn't replied again seeing that he started this debut!

[parthiban]

Think I'm going to duck out of this one before embarrassing myself any further, way out of my depth in this one

 

I do agree with you that there is no such thing as an unprocessed photo (even going back to film cameras) - a photo will always be an interpretation but then you can go even deeper and say our eye does the same thing so what exactly is the real image.......

 

Yep you can see a slight difference between RAW and the processed JPEG but it's very slight and I'm sure a large proportion of people wouldn't even notice (I have friends who are adamant they can't tell any difference between VHS and DVD )

 

Learning some interesting stuff though so keep it going!

[Rich]

Think I'm going to duck out of this one before embarrassing myself any further, way out of my depth in this one

 

You haven't embarrassed yourself at all mate, I suppose it's the way you look at it that can make it a complex subject.

 

Yep you can see a slight difference between RAW and the processed JPEG but it's very slight and I'm sure a large proportion of people wouldn't even notice (I have friends who are adamant they can't tell any difference between VHS and DVD )

 

The photos I posted weren't a good example TBH, I'll find another to use. A good example of what I mean is with my old bridge camera. If you remember the camera was processing the reds as pink? Didn't matter what I did to the settings or what mode I took the photo in all reds were pink or not very vibrant. As it was a bridge camera it didn't save the RAW data so I had no choice but to accept the camera processed image....this wasn't the real photo I took, an interpretation of what my eyes saw, it's what the camera did for me. I could've processed them myself to get the reds out but doing this on Jpeg's can be hard work and TBH I couldn't be bothered!

 

I will post some examples later of Tony's Lotus taken with my old camera and my new one to show how different cameras process the images.

 

I lol at the VHS comment!!

[hms]

There is some free third party software for Canon point and shoots, that loads off the memory card and leaves no trace on the camera, that will allow you to save in RAW format. Also adds loads of other features as well. I may have posted about this software earlier.

 

FeaturesProfessional control - RAW files, bracketing, full manual control over exposure, Zebra-Mode, Live histogram, Grids, etc.

Motion detection - Trigger exposure in response to motion, fast enough to catch lightning.

USB remote - Simple DIY remote allows you to control your camera remotely.

Scripting - Control CHDK and camera features using ubasic and Lua scripts. Enables time lapse, motion detection, advanced bracketing, and much more.

 

CHDK, the wiki is here:

http://chdk.wikia.com/wiki/CHDK

Have it on my camera but dont use it as it's an older one with a small screen, so I cannot read the menu!

h

[parthiban]

The photos I posted weren't a good example TBH, I'll find another to use. A good example of what I mean is with my old bridge camera. If you remember the camera was processing the reds as pink? Didn't matter what I did to the settings or what mode I took the photo in all reds were pink or not very vibrant. As it was a bridge camera it didn't save the RAW data so I had no choice but to accept the camera processed image....this wasn't the real photo I took, an interpretation of what my eyes saw, it's what the camera did for me. I could've processed them myself to get the reds out but doing this on Jpeg's can be hard work and TBH I couldn't be bothered!

 

I will post some examples later of Tony's Lotus taken with my old camera and my new one to show how different cameras process the images.

 

I lol at the VHS comment!!

I meant between the RAW and the JPEG on our DSLRs, there's a difference but it's not immediately noticeable (well to me anyway ). Do remember the stuff about your bridge though, and there definitely was a difference there but without seeing a RAW file we don't know whether that was the processing or the sensor that was creating the colour problem.

 

Will be interesting to see the pics of Tony's car though off both cameras

[Sagitar]

Here are five images taken within seconds of each other in bright sunshine in my garden this morning.

 

The camera was on a tripod, the lens aperture was set to f/6.3 throughout. I used the same setting (AV) for all the images, so I set the aperture and the camera decided on the exposure.

 

The images are just as they came out of the camera apart from reducing their file size to make them easy to load.

 

They are each obviously differently exposed. i.e. the in camera processing system has made a different exposure decision for each of them. Why do you think that is? And given the comments leading up to this post do you think that any one of the images is more "real" than the others?

 

[Rich]

Brilliant example Sagitar...

 

Would the camera have processed them on the basis of how much light was entering the camera even though they were taken seconds apart?

[Sagitar]

Would the camera have processed them on the basis of how much light was entering the camera even though they were taken seconds apart?

 

Thanks E_R, but not many other takers then?

 

The amount of light entering the camera was as near to constant as it would ever get. All the pics were taken under the same conditions - full sun and a cloudless sky, inside a period of less than a minute.

 

The reason for the differences between them lies in the camera metering system. It allows spot metering and I set it up to read in turn from each of the five colours on the target.

 

1) The first pic is based on metering on the white ring. It results in the camera setting 1/2000 sec as the necessary exposure. This is much less than is really needed and is partly because the metering algorithm assumes that the meter is seeing a mid-tone whereas it is actually looking at a high-tone. The histogram for this picture shows that the top third of the exposure range is missing completely, because white has been pushed down towards the centre of the histogram.

 

2) The second pic is metered on the black ring. It results in the camera setting 1/640 sec as the necessary exposure; more than three times as much as the previous pic. It is clearly over-exposed and the histogram is missing about 15% from the bottom end, but the distribution is much better than in (1). The effect is not so marked as for the white ring because the black ring is not a really dense black.

 

3) The third pic is metered on the blue ring. It results in the camera setting 1/1250 sec as the necessary exposure. The histogram is better spread but it still loses about 20% from the top end because of the mid-tone metering assumption.

 

4) The fourth pic is metered in the red ring. It results in the camera setting 1/1000 sec as the necessary exposure. The histogram is now much better spread, because the red ring is very close to being the mid-tone for the target but there is still some 6% to 7% missing at the top end of the histogram.

 

5) The final pic is metered on the gold ring. It results in the camera setting 1/1600 sec as the necessary exposure. The histogram spread is more limited and is missing about 25% at the top end because the gold is a high-tone but is being pushed into the mid-tone region by the assumptions of the metering algorithm.

 

My purpose in carrying out this experiment is to show that images produced by the camera are not just a function of the light that enters the lens. For these five pictures, the light entering was always the same. It is the software inside the camera that makes judgements and adjusts the nature of the output from the camera. That software was designed by people and in choosing to use the untouched output from the camera we are depending upon the skill and judgement of those people rather than upon our own abilities.

 

In the five images that I have shown, I would judge that not one of the images is perfectly exposed. Further, because of the range of tones and the different ways in which they reflect light, one could argue that it is never going to be possible to find a single exposure level that will "see" the target as the eye and brain see it. The "red" image is closest to giving an ideal histogram, but I would want to reduce the exposure a bit and adjust the distribution at the dark end of the histogram. It is a question of judgement, perception and preference.

 

What I find odd to the point of perversity, is that anyone should accuse me of cheating, because I want to use my own skill and knowledge to change the imperfect output that the camera creates.

[parthiban]

That is quite interesting with the 5 photos and how they look so different without much actually changing.

 

The question though is which one best portrays what your eye was seeing? I think that's all Tony was trying to say......

[Tango]

That is quite interesting with the 5 photos and how they look so different without much actually changing.

 

The question though is which one best portrays what your eye was seeing? I think that's all Tony was trying to say......

 

Whose eye? That's the other side of this very interesting subject. We each have a network of light sensors in our eyes that send electrical signals to our brains, and our brains process these signals into sensations of sight. This means that colours only really exist within the brain of the viewer, and as I've no doubt we're not all built exactly the same, there will be variation in each part of the system in more or less the same way as Sagitar's expanation for variation in the software and hardware in a DSLR. I'd suggest that if proof were needed of variation, the condition of colour blindness provides it, although the term 'colour blind' is a bit of a misnomer as almost all colour deficient individuals can distinguish at least some colours depending on the number of colour-sensing cones that are missing or less responsive than normal to incoming light.

 

Light travelling from objects to our eyes may well be transmitting/reflecting a different set of wavelengths of light or White Light; but what essentially defines a 'colour' as opposed to a 'wavelength' is created within the brain. We use Colour Names to describe how we want to reproduce a given colour, but are we all seeing or interpreting the same? Unlikely.

[parthiban]

I meant his own eye - so the truest representation compared to what he was seeing.

 

 

But yes absolutely, this subject could get very deep if you start thinking about what it is we all actually see, as no one will ever know if we are actually seeing the same thing.......

[Rich]

Would the camera have processed them on the basis of how much light was entering the camera even though they were taken seconds apart?

 

Thanks E_R, but not many other takers then?

 

The amount of light entering the camera was as near to constant as it would ever get. All the pics were taken under the same conditions - full sun and a cloudless sky, inside a period of less than a minute.

 

The reason for the differences between them lies in the camera metering system. It allows spot metering and I set it up to read in turn from each of the five colours on the target.

 

1) The first pic is based on metering on the white ring. It results in the camera setting 1/2000 sec as the necessary exposure. This is much less than is really needed and is partly because the metering algorithm assumes that the meter is seeing a mid-tone whereas it is actually looking at a high-tone. The histogram for this picture shows that the top third of the exposure range is missing completely, because white has been pushed down towards the centre of the histogram.

 

2) The second pic is metered on the black ring. It results in the camera setting 1/640 sec as the necessary exposure; more than three times as much as the previous pic. It is clearly over-exposed and the histogram is missing about 15% from the bottom end, but the distribution is much better than in (1). The effect is not so marked as for the white ring because the black ring is not a really dense black.

 

3) The third pic is metered on the blue ring. It results in the camera setting 1/1250 sec as the necessary exposure. The histogram is better spread but it still loses about 20% from the top end because of the mid-tone metering assumption.

 

4) The fourth pic is metered in the red ring. It results in the camera setting 1/1000 sec as the necessary exposure. The histogram is now much better spread, because the red ring is very close to being the mid-tone for the target but there is still some 6% to 7% missing at the top end of the histogram.

 

5) The final pic is metered on the gold ring. It results in the camera setting 1/1600 sec as the necessary exposure. The histogram spread is more limited and is missing about 25% at the top end because the gold is a high-tone but is being pushed into the mid-tone region by the assumptions of the metering algorithm.

 

My purpose in carrying out this experiment is to show that images produced by the camera are not just a function of the light that enters the lens. For these five pictures, the light entering was always the same. It is the software inside the camera that makes judgements and adjusts the nature of the output from the camera. That software was designed by people and in choosing to use the untouched output from the camera we are depending upon the skill and judgement of those people rather than upon our own abilities.

 

In the five images that I have shown, I would judge that not one of the images is perfectly exposed. Further, because of the range of tones and the different ways in which they reflect light, one could argue that it is never going to be possible to find a single exposure level that will "see" the target as the eye and brain see it. The "red" image is closest to giving an ideal histogram, but I would want to reduce the exposure a bit and adjust the distribution at the dark end of the histogram. It is a question of judgement, perception and preference.

 

What I find odd to the point of perversity, is that anyone should accuse me of cheating, because I want to use my own skill and knowledge to change the imperfect output that the camera creates.

 

I didn't think about the metering, it's not something I've used much ATM but I have a book for my camera now so will read up on it. Makes sense though, you meter a specific part of the frame and as the camera is set to aperture priority the camera will adjust the exposure differently everytime like you have demonstrated.

 

That's a good point Tango makes about our eyes seeing colours differently from each other, I never thought about that.

[Sagitar]

That's a good point Tango makes about our eyes seeing colours differently from each other, I never thought about that.

 

Yes indeed, a very deep question and beloved of philosophers along with others such as, "if a tree falls in the forest and there is no-one there to hear, does it make any sound".

 

Though it's an interesting question, I doubt if it makes much difference to photography. If I see red differently than you do, it still boils down to how my eyes and brain respond to a particular part of the visible spectrum, whether generated by the subject or a picture of the subject. So we are both likely to see an "accurate" photograph as having the same colour as the subject. There are effects that cause confusion, metamerism for example. You may have noticed that two adjacent areas of colour on your car that normally appear perfectly colour matched, will sometimes appear not to be matched. It's because the two areas respond differently to changes in the illumination source. My understanding of the colour vision deficiencies that people call colour blindness is that they generally result from the absence of function of particular colour receptors in individuals. I don't think anyone suggests that the individual receptors produce different outputs in different individuals. Given the same cellular and chemical structures, I think it is reasonable to judge that we all normally see pretty much the same thing, but I doubt if we can ever be certain.

 

I think the best that the photographer can do is to manage the various elements of colour reproduction to get the closest possible match of subject and image. For the "pro" who needs to do it consistently on a daily basis this means using a good monitor, scanner and printer capable of reproducing an extended colour gamut and taking care to calibrate them regularly using a colorimeter and appropriate control software.

 

There is little point in arguing about the accuracy of an image if it is being displayed on a monitor calibrated differently from the one that the photographer used in creating it. Similarly, a printed image will never look the same as the monitor image if they are not matched through calibrated colour profiles.

 

It's a complex topic and in my book, pressing the camera release button is definitely part of the beginning rather than the end.

[Rich]

Have you calibrated your monitor? I've looked into it but seems expensive and for a hobbyist doesn't seem worth it. That is another good point though about what we see on the screen as one will display colours differently from another.

 

I've not actually printed any photos out yet or sent some off to be printed but I will do soon as I'm interested in seeing how they come out....do you print your own? I've heard that DSCL are very good and their prices are too but I'm also looking at getting a canvas print made to see how that comes out.

[phipck]

haha, i always do this, i asked some questions in your other thread that may or may not get answered in the next few comments you make here :-)

[Sagitar]

Have you calibrated your monitor? I've looked into it but seems expensive and for a hobbyist doesn't seem worth it. That is another good point though about what we see on the screen as one will display colours differently from another.

 

I've not actually printed any photos out yet or sent some off to be printed but I will do soon as I'm interested in seeing how they come out....do you print your own? I've heard that DSCL are very good and their prices are too but I'm also looking at getting a canvas print made to see how that comes out.

 

I used (and still have) a system called Monaco Optix with EZColor software to calibrate my monitor and produce colour profiles for my scanner and printer in the days when I did a lot of serious photography, but I haven't updated the software for some time. One of the really serious problems with this sort of stuff is that it goes out of date quite quickly and tends not to get updated for new operating systems.

 

If you have a decent monitor that is used by the graphics fraternity there is often access to profile information that you can use to set up your monitor and the standard profiles available for good printers and scanners are excellent. It isn't as satisfactory as a proper calibration, but as you say, the gear is quite expensive for the occasional user.

 

I do all my own printing, though a lot less than I used to, on an A3 Epson printer using Lyson inks and the profiles provided by the supplier of the inks.

[Rich]

haha, i always do this, i asked some questions in your other thread that may or may not get answered in the next few comments you make here :-)

 

I replied to your question in the other thread but no I haven't calibrated it, apart from maybe using some software for the contrast...black box on white background test? What Sagitar has said about profiles I'm going to try that as I'm not even sure what mine is set to.

[Sagitar]

haha, i always do this, i asked some questions in your other thread that may or may not get answered in the next few comments you make here :-)

 

I replied to your question in the other thread but no I haven't calibrated it, apart from maybe using some software for the contrast...black box on white background test? What Sagitar has said about profiles I'm going to try that as I'm not even sure what mine is set to.

 

I don't know what O.S. you are using, but in Win7 the following gets to your profiles:

 

Start - Control Panel - Appearance and Personalisation - Adjust screen resolution - Advanced settings - Color Management - Color Management tab.

 

You will see any and all profiles associated with the display and an option to mark one of them as the default profile. The one that you have marked as the default will load on startup.

[parthiban]

I do all my own printing, though a lot less than I used to, on an A3 Epson printer using Lyson inks and the profiles provided by the supplier of the inks.

Is Lyson good quality then? I've got an Epson A3 as well (Photo 2100) and been thinking about finding an alternative recently as the Epson cartridges are getting ridiculously expensive!

[Sagitar]

I do all my own printing, though a lot less than I used to, on an A3 Epson printer using Lyson inks and the profiles provided by the supplier of the inks.

Is Lyson good quality then? I've got an Epson A3 as well (Photo 2100) and been thinking about finding an alternative recently as the Epson cartridges are getting ridiculously expensive!

 

Lyson isn't cheap, but it costs a fair bit less than Epson.

 

It works well for me on my R1800 but YMMV.

 

Have a look here

 

A lot depends upon how much printing you do. I believe Marrut do a continuous ink system for the 2100 but I don't know anyone who uses it so I can't comment.

[Rich]

Start - Control Panel - Appearance and Personalisation - Adjust screen resolution - Advanced settings - Color Management - Color Management tab.

 

You will see any and all profiles associated with the display and an option to mark one of them as the default profile. The one that you have marked as the default will load on startup.

 

Interesting, I don't actually have any profiles associated with the display, I will have to play about with them and see what happens.

[parthiban]

Lyson isn't cheap, but it costs a fair bit less than Epson.

 

It works well for me on my R1800 but YMMV.

 

Have a look here

 

A lot depends upon how much printing you do. I believe Marrut do a continuous ink system for the 2100 but I don't know anyone who uses it so I can't comment.

Cheers, don't do a lot of printing to be honest but I might give that a go when I run out of the cartridges I've currently got.