What is "true to life" in a coin photo?

[moonmoth]

This thread was stimulated by some comments in another thread, "Your advice needed for cutting and cropping coins", to the effect that some people prefer an unedited photo because it better represents the coin as it is in the hand, in real life.

Some people might prefer an idealised image, that gives the best possible view of their coin.  Some might even go further and fix up blemishes, but for this thread I will leave alone that ignoble concept.

But what is an unedited image?  Even the purists must make many choices to get any image onto  a screen.  Things that will affect how the photo looks include:

- Choice of camera
- Choice of background and setting
- Choice of lighting type and position.
- Choice whether to use flash.
- Choice of exposure time, focus setting, aperture, white balance, simulated film speed.
- Choice of file type: JPEG, RAW, or other.
- Choice of quality in a JPEG.
- Many choices when converting a RAW image.
- Choice of software to display the image.

And the condition and type of the coin also makes a big difference.  All this is before any thought is given to editing. 

My own view is that an unretouched photo does not usuallly represent a coin at its best.  By its best, I mean showing all the flaws and blemishes, all the details of tone and patina, every good and bad point about it.  I do not think that the beauty of a coin is in artificial perfection, but in its real self.  I will happily:

- edit highlights so that the coin surface can be seen instead of a white glare.
- Similarly edit black areas with no visible detail.  I will still aim to preserve a full tonal range of white to black if possible.
- Apply levels gradients to compensate for one-sided lighting.
- Sharpen enough (using "unsharp masking") to compensate for the inherent softness of digital images.
- Adjust saturation levels to compensate for odd camera effects.
- Allow the final image to be lighter than the coin seen under normal lighting conditions, such as a room's incandescent bulbs, so that detail is made apparent.  The coin looks like that under very bright light anyway.

If I think I have made the result too artificial, I will scrap it and start again, using less bold changes.

Here's one example.  The first photo is the one I saw when I bought the coin.  The second is the unretouched flash images of the obverse and reverse, pasted together and shrunk for this display, but otherwise unchanged.  The final one (next message) will be the version I ended up with.

[moonmoth]

OK .. here is my final version.

Now, I do not think this image is untrue to the coin.  No-one can say it artificially eliminates flaws, for example.  In fact I think it is more true to the essence of the coin in my hand than either of the two previous images.

Any comments?  .....

Bill

[DruMAX]

It depends on the camera you are using. A great digital camera will not make soft images thus no need to sharpen...I have never had to use sharpen filter on any coin.

Lighting if set up right will not leave areas that are dark or very bright thus there should be little need for selective adjustment, maybe a bump in the levels or a bump on brightness and contrast at the most.

I started taking photos of coins and using all the above listed techniques to adjust and fix what I was lacking in correct camera settings or poor lighting.

I spent quite some time messing with shutter speed, lighting placement, light filtering and finally found a setting that produces about as close to the original coin color, detail, lighting, etc...as I think is possible.












All these coins came out almost as perfect as can be. I did not use any form of PS adjustment, not even a bump on the levels...optio5i digital with a down light (about 2 feet florescent at a bit of an angle)

I guess when I said that I wanted to do as little in editing as possible is that IMO most good cameras (digital or otherwise) are perfectly able to capture an image almost life like in most aspects if the settings are just right (which takes a lot of time and testing) thus eliminating the need for further editing. Unless its just a bump here or there on a slider.

Also I cant stress enough the difference the angle of the light makes...I was going to post a comparison of a coin that drastically showed the effect the angle of light makes. It was a rather flat Constantine coin with the right angle it can help bring out the edges with a little shadow (like you would do if it were in hand) and in the case of this coin, it caused the back figure to look DRASTICLLY different in relation to what side the light is coming from. One way made the figure look almost like a skeleton....the other angle brought out the details much more and was a dead perfect reproduction of the figure as it is in real life, filled out the face and the cloth on the body...

What I am mainly looking to achieve is the clearest possble detail, the best possible match in color, without lighting glare (as if holding it well away from a light that makes a heavy shine) in the original photo itself doing as little post adjustment in PS.

[maridvnvm]

Since migrating to a digital camera I have been struggling with getting a setup that gives the sort of images that reflected reality of my coins.
I can only concur that finding a lighting setup that you are happy with is challenging.
My collection contains silver, silvered (ranging from fully silvered through partially silvered to slight silvering remaining) and bronze coins, with their associated range of toning range. I have worked towards a setup that produces images that I am happy with that reasonably closely represent the coin in hand when seen on the screen of my laptop.
I try to work with a minimum of software effort since I have too many coins to photograph to be able to spend time in software.
I take my photographs at a high resolution and then reduce them for general use, whilst keeping the high resolution images for the records.

The following is an example of the image that I am now getting. Whilst I am not 100% happy yet I will keep tinkering util I have it about right.



And the picture below is an example of the detail from the full size image,

Regards,
Martin

[slokind]

I have fussed a great deal over this question, and I have no bone to pick with any of the above.  Defining what is True / vrai / vero / treu / alêthos or gnêsios in a coin photo is almost impossible; words won't hack it (consider how Keats couldn't really evoke a neo-Attic marble urn...).  I got to thinking that the negative approach is more useful.  Can't be exhaustive, but, for example, a true photo of a denarius doesn't look like cookie dough; a true photo of a stripped orichalcum sestertius doesn't look like gold, and gold is not at all lemon-colored; metal does not look like plasticine or plaster or plastic.  Though allowance for dim monitors must be made, out of consideration for one's fellow members who haven't been able to replace the things, no image of a coin ought to look like a photo through milk glass (or like an image taken through a puddle of molasses, either).  Do you agree?  There are as many true ways of seeing, so of photographing, as there are local and regional light situations, personal eyes and visual processing in the brain, and so on: no one is truer than the other.  But the false ones are false to practically everyone.  Right?  Pat L.
P. S. Your Probus breastplate has true probity (couldn't resist putting it that way), and it is also a great advertisement for digital cameras.
P.P.S.  I sometimes, considering the reflectivity of shiny silver, hold one of the red boxes I used to get Saflips in opposite the lamp, to obtain some warm highlights (choose your own color, but red cannot be mistaken for the local color of the coin); I did so on a very clean denarius I got the other day (attached)--maybe a bit much, but the neutral gray of the ground glass is unretouched.

[moonmoth]

Hi, DruMAX!

I think you have done an excellent job with the lighting, the exposure, and the focus, and you are producing some remarkable photos.  But here's where the question of taste comes in.  From my viewpoint, I'd say you have too much yellow in the Claudius and Gallienus coins to be true to life.  I would desaturate that considerably.  See below .. this also has the obverse turned clockwise slightly, a slight gradient applied to lighten the bottom left of the obverse, the black toned areas lightened very slightly, and 15% unsharp mask, radius 1 pixel, threshold 0. You don't need to make any of these changes.  But I would if it were my coin.  I sit there thinking "how can I bring out the true beauty of this coin?" and the software tools in Photoshop are all there for me.

However good your camera, and yours is clearly a stunner and well set up, all the variable factors I mentioned above will come into play somewhere.  And I forgot to mention that different computer screens, viewed in different lighting conditions, can make the same photo look very different.

I'm not saying there's any right or wrong here.  In fact I am saying there definitely is not any right or wrong here .. whether you get the right image with the camera or with the software or both, it's just as artifical and just as real.

Bill

[moonmoth]

Hi, Pat!

I got to thinking that the negative approach is more useful.  Can't be exhaustive, but, for example, a true photo of a denarius doesn't look like cookie dough; a true photo of a stripped orichalcum sestertius doesn't look like gold, and gold is not at all lemon-colored; metal does not look like plasticine or plaster or plastic.  Though allowance for dim monitors must be made, out of consideration for one's fellow members who haven't been able to replace the things, no image of a coin ought to look like a photo through milk glass (or like an image taken through a puddle of molasses, either).  Do you agree?

I think people should use the approach that works best for them.  When I used film, I would have said the same as you about the negative approach.  You can't fix a poor film negative, not really.  But a digital image is not a negative.  Different approaches can be used .. images that seem to be through milky glass or treacle might or might not contain the necessary information.  You can't put it in if it's not there.  But often there is something to be stripped out: excessive brightness, or excessive yellow, for example.

You can always use the empirical test.  Does the end result look satisfactory?  If so, does it matter how you got it?

But .. what does "satisfactory"  mean?  This differs from person to person and I won't argue with taste or purpose.  In my case, I want to love the image as much as I love the coin.

Bill

[moonmoth]

P.P.S.  I sometimes, considering the reflectivity of shiny silver, hold one of the red boxes I used to get Saflips in opposite the lamp, to obtain some warm highlights (choose your own color, but red cannot be mistaken for the local color of the coin); I did so on a very clean denarius I got the other day (attached)--maybe a bit much, but the neutral gray of the ground glass is unretouched.

Pat, I can't be critical of your photos.  You know what you want, and you get it.  Any comments I might have made on other shots of yours were made in ignorance of their purpose and are withdrawn. 

In my case, as long as I can represent the roughness of an over-cleaned denarius, the toning on a long-stored one, the soft and greasy look of an old-fashioned fouree, and the bright false look of a white metal imitation, I won't care how I do it.  (Note that I am not yet claiming to be able to do all these things!)

Bill

[DruMAX]

Hi, DruMAX!

I think you have done an excellent job with the lighting, the exposure, and the focus, and you are producing some remarkable photos.  But here's where the question of taste comes in.  From my viewpoint, I'd say you have too much yellow in the Claudius and Gallienus coins to be true to life.  I would desaturate that considerably.  See below .. this also has the obverse turned clockwise slightly, a slight gradient applied to lighten the bottom left of the obverse, the black toned areas lightened very slightly, and 15% unsharp mask, radius 1 pixel, threshold 0. You don't need to make any of these changes.  But I would if it were my coin.  I sit there thinking "how can I bring out the true beauty of this coin?" and the software tools in Photoshop are all there for me.

However good your camera, and yours is clearly a stunner and well set up, all the variable factors I mentioned above will come into play somewhere.  And I forgot to mention that different computer screens, viewed in different lighting conditions, can make the same photo look very different.

I'm not saying there's any right or wrong here.  In fact I am saying there definitely is not any right or wrong here .. whether you get the right image with the camera or with the software or both, it's just as artifical and just as real.

Bill

I would have to say that the changes you made, at least in my opinion, made the photo look a bit less like the surface of the actual coin. This coin has a bright copper color that is almost identical to the picture I posted...By moving the light a bit father away I took some shine off it. I have held these scans up to pro calibrated Canon Color Laser Copier 5000 prints (almost photo realistic prints I make at work ) to the real coin and they are about as dead on as I can get them, if they need anything at all, maybe just a very minor color balance correction but I wont bother with that as its it such an extremely minor difference in tone. The changes you have made removed the vibrant copper color that the coin does indeed have in person.  Maybe its just the difference of perception each person has when they look at these coins and photos of coins. Either way, I feel very satisfied with my current set-up and technique. Not to say I wouldnt make changes if I felt they were needed....or that my way is the best way at all.

[Numerianus]

First: the use of a photo editor is absolutely necessary and inavoidable. There is a great difference
between monitors. I adjust all my photos to my 3-year old Sony notebook with 1600x1200 UXGA  monitor.
At the moment I am working with a brand new Think Pad 1400X1050 monitor which has  different properties.
My photos looks too dark and too contrast on it.  Editing includes a lot of subtleties and adjustment can be done
much more precisely than just using only camera features. E.g., enhancement depends on the desired size of the picture.
Even a rotation on 1 degree sometimes does matter.

Second: the coin is a dynamic object  and there is no sense to  separate it from the environment. We look at it under various lighting conditions and in various environment. By the way, never in the environment of the photo stand.
Moreover, different people may have different preferences. E.g., I like  photos made under natural light in front of a window
(under different  conditions) but recently I my preferences, at least, for some coins,  turned to electric light.    

[maridvnvm]

As much as I envy the images created by Numerianus, I have found that the natural lighting available in my home does not allow me to generate consistent results, which is one of the criteria that I strive for. As such I have had to resort to a desk lamp with a daylight bulb setup. I used to use a flatbed scanner but the images generated seemed completely out of my control.

I have included two images of the same coin, the first is from a flatbed scanner, the second was taken a few minutes ago just to illustrate the results that I could achieve with a bronze coin with a variagated patina.
Regards,
Martin

Flatbed:-


Digital Camera below:-

[DruMAX]

First: the use of a photo editor is absolutely necessary and inavoidable. There is a great difference
between monitors. I adjust all my photos to my 3-year old Sony notebook with 1600x1200 UXGA  monitor.
At the moment I am working with a brand new Think Pad 1400X1050 monitor which has  different properties.
My photos looks too dark and too contrast on it.  Editing includes a lot of subtleties and adjustment can be done
much more precisely than just using only camera features. E.g., enhancement depends on the desired size of the picture.
Even a rotation on 1 degree sometimes does matter.
   

I both agree and disagree. If you are talking about cutting, pasting, rotating and the like...certainly you must use PS. But I do not think that one must use it for adjustments to the image itself (color, contrast, levels, saturation) if the camera is set correctly. I automatically attempt to adjust the Brightness/contrast and level with everything but these days, 8 out of 10 times I look at it and realize that it isnt needed and leave it be.

It seems its just a matter of those who prefer to set the camera just right to those who prefer to take the picture and correct it after in PS. I just believe that if I get the camera set just right, the look will be more true than if I attempt to correct what I get from a camera not set precisely.

maridvnvm: the second one looks better to me. I dont like the look scanners give because of the lack of a localized light source, instead you get a complete saturation of light with no realistic shadowing. Of course we work with what tool we have.

[moonmoth]

One choice that has to be made when taking a digital photo is the file type.  This has a massive effect on how the image is produced.

I think all digital cameras will produce a JPEG.  This is a "lossy" format - it is compressed in such a way that some data is always lost.  The camera contains software that will take the raw image from its sensor, edit it to whatever colour temperature has been set (the "white balance"), adjust the brightness and saturation, and leave out some information to compress the result into a smaller file size.   The resulting edited data is then saved to a memory stick.

Some cameras offer the option of saving a raw image, containing all the data captured by the sensor.  This can then be converted or edited later using either the software provided with the camera or a third party editor such as Photoshop's RAW plugin.  This raw image is the nearest you will get to a negative using a digital camera.

The method used by Pat and others allows the camera to do the initial editing.  A raw image from the same camera would look much different.  But the advantages of a raw image are:

- you can edit for the correct colour temperature, brightness and saturation later, using better software than the stuff in the camera.
- You can go back and do it again later if still better software becomes available.
- You have all the data and can save the edited version in a loss-free format like TIFF if you prefer.  If you want smaller file sizes, TIFF has the option of using the loss-free LZW compression algorithm.
- You can edit and re-edit withut losing data.  If you edit a JPEG and save the result, for example to change the background colour, you are compressing something that has already lost some data through a compression algorithm.  A high-quality JPEG can stand that without affecting how the human eye perceives the result, but some can not.

Now, a good camera with good software will do a good job.  JPEG compression was defined by the Joint Photographic Experts Group to look good to the human eye.  If it gives the desired result, I would not quarrel with it.  It works for Pat, and I like it too.  But purists should be aware that letting the camera software do your editing for you, which all JPEGs straight from the camera do, is not a particularly better option than doing it yourself with different software.

The Photoshop raw plugin allows a lot of detailed control over how your image is converted ready for editing with Photoshop.  Here, as an example, is the nearest I can get to showing a raw image, and the JPEG version produced using Photoshop.  (The raw image is as shown by the Photoshop browser, captured off screen by other software.)  The raw version looks like cheese, because of the low colour temperature of the incandescent bulb I used as lighting, but contains all the information I need to produce a true image.

Bill

[Numerianus]

I completely agree. The settings of a digital camera do not change physical properties. They are  controls   varying parameters of some implemented algorithms of image processing applied to the array of numbers obtained from the sensor matrix, a kind of authomatical adjustment.   This means that if the camera store just this array, you can do the job on your computer manually with a number of algorithms and obtain much better results for an individual picture on your particular monitor. The information lost during the passage from yje  raw format  to (large) jpeg is not essential and one still can profit from the manual adjustment which is better than an automatical adjustment.
As I said, there are many subtleties. How do you chose the size of the pic? In fact, rescaling  means that pixels are aggregated 
and the outcome may vary quite a lot in dependence of size.  The perception depends also on the size. Should the size of the pic  depends on the coin size? 

[DruMAX]

I was getting the color you are showing in your 'raw' jpg when I started but was able to get a raw image that was far closer to the correst color by adjusting the light settings on the camera.

Well, I guess its all in how you approach the subject. I use both conventional cameras and digital cameras as part of my profession to generate graphics and photos for both print and the web. The way I see it is unless you are printing your end product will be 72 dpi jpg for web view (saving the large format before reduction). Much of the same options are present on my camera that I would use in PhotoshopCS: levels, brightness contrast, color balance, hue saturation in additions to light settings and other settings that pertain to the conversion and compression.  I would be hard pressed to think of much else I would utilize in PS  that wouldnt begin truly changing the image or actual physical properties of the coin (selective pixel changes, cloning or most filters.)

There is little need for sharpening the image if it is a large format shot in macro (they are as clear as a bell, you see every little mark before downsize with high mega pixel camera) nor would one blur it (since it has already been compressed and the point is to get the sharpest image possible). I see no need for most other image adjustment options if you can accomplish these tasks with he camera. I took the time and effort to program several settings for different levels of light that I use to best capture high relief and low relief.

I use Photoshop and Illustrator every day for photo restoration, manipulation, web graphics, graphics for offset print and the web and in conjunction with illustrator and know that one can achieve a "true life" image if you choose to go that route and it seems to me a question of whether you want to do it manually afterwards in PS or (if camera allows) when you convert the image.  Without doubt most images I bring into Photoshop (from a digital camera) I will end up making slight adjustments before reducing it for web view (mainly contrast and slight color tweaks) but in the end those small adjustment are all that is needed when the camera is set correctly for the conditions. This also is dependent on the camera one uses and the options said camera offers.

In the end I just want as accurate a representation of the actual coin as one can get in 72 dpi and 600 points width how ever one reaches that. I think many have said there is no one right way.

[Numerianus]

My wife is photographer  and infographist
(see her webpages http://kabanova.neuf.fr/). She is very sceptic about my work with Photoshops considering it as
non-professional. But many people say good words about my outcomes.

[moonmoth]

Then, as has been pointed out elsewhere, there is the question of reducing the size of the image.  The original is huge in web terms, pretty big in print terms, and will certainly need some work.  This can also be done in the camera, by selecting a small size of JPEG.  But again, you need to be aware that you are letting the camera's software do this for you, and it might not be using the best algorithm.  Photoshop uses a pretty good algorithm.

Many cameras apply sharpening to their JPEG images before saving them.  Reducing the size of the image also makes it look slightly sharper.  Here, as examples, are first, a section of the raw image I showed above, converted for colour temperature but left at its original size; then, the same section reduced in size by about the same amount as I used to show the final image above.  The second image shows the reduced image on the left, and on the right the same but with 20% extra sharpening applied.  Which looks most lifelike is up to the vision and taste of the photographer. 

Bill

[moonmoth]

Hi, DruMAX.

I was getting the color you are showing in your 'raw' jpg when I started but was able to get a raw image that was far closer to the correst color by adjusting the light settings on the camera.

That probably means that the camera's software is doing some pre-processing. 

In the end I just want as accurate a representation of the actual coin as one can get in 72 dpi and 600 points width how ever one reaches that. I think many have said there is no one right way.

Absolutely!  I don't mean to suggest that there is.  In fact, the opposite: I mean to suggest that there definitely is not a single right way.  I am merely setting out in more detail what some of the options for processing digital images really entail. 

It would be very boring if everyone used the same method and produced identical pictures.

Bill

[DruMAX]

But many people say good words about my outcomes.

and thats what I think we all are looking for. Can you post a photo of one of your coins? One you think is a stunning example as a better reference of the outcome of your process?

[Numerianus]

I displayed already an example of Maxentius coin which is not very simple case: brilliant black patina
combined with a desert one, green concretions.  It  is too dark on the screen of Think Pad I used at the moment.
So all my photos, probably, needs a brightness corrections to adjust for this type of monitors (of course, this can be done very quickly by Picasa). Now the photo which I like.  Philip  (electric light):

[Numerianus]

Now to compare. Galerius (electric light, 96 dvi)

[Numerianus]

The same coin under sunlight looks quite different.

[Numerianus]

Gallienus (electric light).

[DruMAX]

thanks!! they look very good indeed.

[slokind]

Everything you all say is basic knowledge and quite true, and external HD storage for all the raw files is now affordable.  But there is a difference between your accession file photos, also used for sharing / selling on line, and the photos you take for your great show of numismatic photography or for the great book to replace the now hard to obtain Hirmer volumes.  I mean, even techie photographers like Ansel Adems took Leica 1 snapshots in early Kodachrome without consulting the zone system.  Also, since digital cameras contain chips and some resident software, it is technically accurate to say that they do processing, but the auto-sharpening can be (ought to be) turned off, and settings such as White Balance / Color Balance amount to no more than interchangeable backs with different films--except that they are more nuanced.  That's not really 'processing'.  Also, though of course all but beginners' cameras allow Raw as a choice, there are alternatives other than taking a picture with lower-number (lossier) jpeg.  Take the largest file the camera can at the highest, finest jpeg camera setting (similar to Graphic Converter's 100% or Photoshop's 12), and store that one rather than the larger and much rawer Raw image--unless, as I said, you're a graphic designer or want for yourself total personal manipulability.  Work with a first copy of that file.  Thing is, when you use digital camera instead of the handy flatbed scanner, you want good files, but you don't want to spend ALL your time being a numismatic Ansel Adams or Renger-Patsch, unless you have no numismatic studies to do, too.  Pat L.