Photo Forum / Film Photography / Darkroom / February 2007
How does tonal compression work?
|
|
Thread rating:  |
Tim Allen - 15 Oct 2006 13:58 GMT Hello,
My newbie questions concerns to tonal compression that happens when a negative (gamma of 0,7, exposure range of about 7 stops, max density of ~2.0D, brightness range of 100:1) is beeing printed on positive paper (gamma of 1,4, ~exposure range 4-5 stops, max density of ~2.0D, brightness range of 100:1). - I just do not understand, or cannot imagine, how the tones are compressed so that the 10 stops of subject brighness are squeezed into the 5 stops of the paper. What happens to the values that are not reproducible, are they lost? I would appreciate if someone could explain to me what´s going on in this process.
Thanks a lot! Tim Allen
Greg "_" - 15 Oct 2006 15:35 GMT > Hello, > [quoted text clipped - 10 lines] > Thanks a lot! > Tim Allen That's a very dense neg, with comparatively short tone range.
To answer:
The relative relation shrinks logarithmically so that it fits visually on the paper of choice.
Positive paper-using a negative?
 Signature Reality-Is finding that perfect picture and never looking back.
www.gregblankphoto.com
Al Denelsbeck - 15 Oct 2006 16:57 GMT "Tim Allen" <tim.allen@gmx.com> wrote in news:1160917100.221556.142590 @m7g2000cwm.googlegroups.com:
> Hello, > [quoted text clipped - 10 lines] > Thanks a lot! > Tim Allen The simplest explanation would probably be "rounding."
If you take a value of something that is expressed in decimals (money is a good example) and are restricted to using whole numbers (whole dollar amounts), you end up using whatever value is closest.
In photographic terms, the subtler gradations become pressed together into one. To oversimplify it, what might have been "lighter medium gray," "medium gray," and "darker medium gray" all now become simply "medium gray."
If you had subtle details within those ranges of the original negative, they will be lost, "rounded" into the narrower range of the paper.
I admit this is slightly misleading, in that it is not necessarily a straight progression. With various methods, you can choose to retain more detail in a particular range, at the expense of others - allow the highlights to blow out so you can retain shadow detail, for instance. But it illustrates the basic idea of total compression.
- Al.
 Signature To reply, insert dash in address to match domain below Online photo gallery at www.wading-in.net
j - 15 Oct 2006 18:03 GMT [...]
Excellent description, Al.
To the OP: If means of copeing with compression interest you, look into "contrast masks", "unsharp masking" and "split grade printing". (All wet-darkroom techniques, not Photoshop versions.) :)
David Nebenzahl - 15 Oct 2006 20:10 GMT j spake thus:
> To the OP: If means of copeing with compression interest you, look into > "contrast masks", "unsharp masking" and "split grade printing". (All > wet-darkroom techniques, not Photoshop versions.) :) Yes: now can someone explain how "unsharp masking" worked in pre-digital days? I've only found references that the term comes from wet photography, but no further explanations. How did they sharpen images optically in those olden days?
 Signature "In 1964 Barry Goldwater declared: 'Elect me president, and I will bomb the cities of Vietnam, defoliate the jungles, herd the population into concentration camps and turn the country into a wasteland.' But Lyndon Johnson said: 'No! No! No! Don't you dare do that. Let ME do it.'"
- Characterization (paraphrased) of the 1964 Goldwater/Johnson presidential race by Professor Irwin Corey, "The World's Foremost Authority."
David Nebenzahl - 15 Oct 2006 20:23 GMT David Nebenzahl spake thus:
> j spake thus: > [quoted text clipped - 6 lines] > photography, but no further explanations. How did they sharpen images > optically in those olden days? Wikipedia has this to say about the subject: http://en.wikipedia.org/wiki/Unsharp_masking
Is this accurate?
 Signature "In 1964 Barry Goldwater declared: 'Elect me president, and I will bomb the cities of Vietnam, defoliate the jungles, herd the population into concentration camps and turn the country into a wasteland.' But Lyndon Johnson said: 'No! No! No! Don't you dare do that. Let ME do it.'"
- Characterization (paraphrased) of the 1964 Goldwater/Johnson presidential race by Professor Irwin Corey, "The World's Foremost Authority."
John - 15 Oct 2006 23:31 GMT >Wikipedia has this to say about the subject: >http://en.wikipedia.org/wiki/Unsharp_masking > >Is this accurate? Yep. That's the basics of it and about all I had the time to learn about it. Works pretty well for some subjects however you need a pin registering jig to make it work well. Hollyweird has used the effect in a few movies to enhance the graphic look of some scenes. I believe Gladiator is one example.
== John S. Douglas Photographer & Webmaster Legacy-photo.com - Xs750.net
Greg "_" - 16 Oct 2006 00:15 GMT > >Wikipedia has this to say about the subject: > >http://en.wikipedia.org/wiki/Unsharp_masking [quoted text clipped - 11 lines] > Photographer & Webmaster > Legacy-photo.com - Xs750.net Howard Bond offers reasonably priced workshops on the technique as applied to conventional printing.
 Signature Reality-Is finding that perfect picture and never looking back.
www.gregblankphoto.com
Tom Phillips - 16 Oct 2006 03:59 GMT Greg \"_\" wrote:
> > >Wikipedia has this to say about the subject: > > >http://en.wikipedia.org/wiki/Unsharp_masking [quoted text clipped - 14 lines] > Howard Bond offers reasonably priced workshops on the technique as > applied to conventional printing. I've seen sone of Howards masked prints (including a 50 inch enlargement of the Tetons.) It was pretty stunning. Course for his workshops you have to travel to MI and they tend to fill up fast.
Matt Clara - 16 Oct 2006 23:42 GMT > Greg \"_\" wrote: >> [quoted text clipped - 21 lines] > stunning. Course for his workshops you have to travel > to MI and they tend to fill up fast. http://www.apogeephoto.com/howard_bond.html One class left in November...
professor352003 - 13 Feb 2007 06:50 GMT Check out Lynn Radeka site on masking at maskingkits.com. I attended a workshop of his and learned about all kinds of contrast masks. Amazing how much there is to it. I think Howard Bond only does unsharp masks and maybe dodge masks. Lynn's workshop is more comprehensive. Highly recommended.
Professor John --
> > Greg \"_\" wrote: > >> [quoted text clipped - 24 lines] > http://www.apogeephoto.com/howard_bond.html > One class left in November... professor352003 - 13 Feb 2007 06:56 GMT Check out Lynn Radeka's site on masking at maskingkits.com. He also teaches workshops and I highly recommend them. Howard Bond only teaches unsharp masking and dodge masking but this workshop is much more comprehensive and teaches all kinds of useful masks. I used to think unsharp masking was the only kind but now I learned how to make all sorts of masks that control other aspects of print contrast.
Professor John
--
> > Greg \"_\" wrote: > >> [quoted text clipped - 24 lines] > http://www.apogeephoto.com/howard_bond.html > One class left in November... Claudio Bonavolta - 16 Oct 2006 08:50 GMT David Nebenzahl a écrit :
> David Nebenzahl spake thus: > [quoted text clipped - 24 lines] > presidential race by Professor Irwin Corey, "The World's Foremost > Authority." This is one type, the most common, of silver masking but there several others. Here are few pointers: - regretted Barry Sherman posted on rec.photo.darkroom his method for LF Ilfochrome printing, I made a copy available on http://www.bonavolta.ch/hobby/fr/photo/masking101.htm - http://www.largeformatphotography.info/unsharp/ - http://www.maskingkits.com a commercial site but with a good description of the various mask types. - Post Exposure by Ctein (ISBN 0240802993) - there was an old Cibachrome booklet, the french version was named "Ilford Cibachrome II - Masques Correcteurs" and gave many infos on contrast and color correction masks.
Claudio Bonavolta http://www.bonavolta.ch
Nicholas O. Lindan - 16 Oct 2006 14:39 GMT > > Yes: now can someone explain how "unsharp masking" worked in pre-digital > > days? I've only found references that the term comes from wet > > photography, but no further explanations. How did they sharpen images > > optically in those olden days? An old post on the subject:
From: "Nicholas O. Lindan" <nolin...@ix.netcom.com> Subject: Unsharp masking: What & How Date: 1997/12/30 Message-ID: <34A9CE46.6FC5@ix.netcom.com> X-Deja-AN: 311579667 Content-Transfer-Encoding: 7bit Mime-Version: 1.0 Reply-To: nolin...@ix.netcom.com Content-Type: text/plain; charset=us-ascii Organization: Netcom X-NETCOM-Date: Tue Dec 30 5:48:19 PM PST 1997 Newsgroups: rec.photo.darkroom
Unsharp masking accentuates the detail in the picture. It does not make the picture sharper. It is the photographic equivalent of turning up the treble control on the stereo.
The results can be quite striking, though, given the right subject matter. Rocky landscapes become surreally bold while the sky and clouds look unchanged.
To explain the process I will use a picture of a light gray square on a dark grey background. Thrilling subject, no? And I am going to take a line through the middle of the picture on which to make graphs of either reflectivity or negative density.
In the beginning we take a plain old B&W picture:
Subject reflectance:
Lite grey _______________________ | | | | | | Dark grey _________| |_______________
Negative density:
Lite grey _________ _______________ | | | | | | Dark grey |_______________________|
So far, so good. If we make a print from the negative we would, if all things are done right, get the same tones as the subject.
Instead we make an unsharp mask by contact printing the negative onto another piece of film, called the mask, and doing it badly. We place a material between the negative and the mask when we expose the mask. This makes the mask into a fuzzy positive image of our subject. Since the mask is fuzzy, the technique has come to be called 'unsharp masking'. We also expose and develop the mask so that it's contrast is less than the negative.
If we plot the density of the mask it looks like:
Very lite grey ______________________ / \ Lite grey ________/ \_____________
Notice that the transitions between the grey tones is sloping. In real life the slope will look more like an S-curve.
If we now sandwich the negative with the mask the densities of the two will add as shown below:
Negative density:
Lite grey _________ _______________ | | | | | | Dark grey |_______________________|
Mask density:
Very lite grey ______________________ / \ Lite grey ________/ \_____________
Resulting density:
Grey Dark grey ________/| |\______________ | | Very dark grey | _____________________ | Almost black |/ \|
The masked negative we have created has a lower contrast for the continuous regions than the original negative. However the contrast at the transitions between the two greys have the same contrast range as the original negative.
To compensate for the loss of contrast we print the masked negative on a higher contrast grade of paper with the following result.
Print reflectance:
White |\ /| Lite grey | \___________________/ | | | | | | | Dark grey _______ | | _____________ \ | | / Black \| |/
Which you can compare with the original subject reflectance:
Lite grey _______________________ | | | | | | Dark grey _________| |_______________
In the print the transition from dark grey to lite grey goes first to black then takes a flying leap to white before settling down to the lite grey tone.
This effect is sometimes called 'sharpen' in photo/paint computer programs, but actually no sharpening takes place. What does happen is that edges are made more prominent.
Note that I cheated in the drawing of the print reflectance in the final step. The horizontal dimensions of the overshoots should be half of what is shown, but ASCII only has one slope available for the /\ characters.
 Signature Nicholas O. Lindan, Cleveland, Ohio Darkroom Automation http://www.nolindan.com/da/index.htm n o lindan at ix dot netcom dot com
David Nebenzahl - 16 Oct 2006 17:38 GMT Nicholas O. Lindan spake thus:
> This effect is sometimes called 'sharpen' in photo/paint computer > programs, but actually no sharpening takes place. What does > happen is that edges are made more prominent. So it really ought to be called "edge enhancement", or some such.
 Signature "In 1964 Barry Goldwater declared: 'Elect me president, and I will bomb the cities of Vietnam, defoliate the jungles, herd the population into concentration camps and turn the country into a wasteland.' But Lyndon Johnson said: 'No! No! No! Don't you dare do that. Let ME do it.'"
- Characterization (paraphrased) of the 1964 Goldwater/Johnson presidential race by Professor Irwin Corey, "The World's Foremost Authority."
prep@prep.synonet.com - 24 Oct 2006 17:37 GMT > Yes: now can someone explain how "unsharp masking" worked in > pre-digital days? I've only found references that the term comes > from wet photography, but no further explanations. How did they > sharpen images optically in those olden days? You expose a film spaced from the original so the fine detail is gone. You normally also expose it so it is quite thin. Expose through both, and the large detail is subtracted from the original, but the fine detail, lost from the mask, is left at full strength.
 Signature Paul Repacholi 1 Crescent Rd., +61 (08) 9257-1001 Kalamunda. West Australia 6076 comp.os.vms,- The Older, Grumpier Slashdot Raw, Cooked or Well-done, it's all half baked. EPIC, The Architecture of the future, always has been, always will be.
professor352003 - 13 Feb 2007 07:04 GMT Sorry for the repeat message. Still getting used to these newsgroups.
Professor John
--
> > Yes: now can someone explain how "unsharp masking" worked in > > pre-digital days? I've only found references that the term comes [quoted text clipped - 13 lines] > Raw, Cooked or Well-done, it's all half baked. > EPIC, The Architecture of the future, always has been, always will be. Gregory Blank - 13 Feb 2007 11:50 GMT > Sorry for the repeat message. Still getting used to these newsgroups. > > Professor John Sure thing "Lynn".
 Signature George W. Bush is the President Quayle we never had.
Nicholas O. Lindan - 15 Oct 2006 19:20 GMT "Tim Allen" <tim.allen@gmx.com> wrote
> I just do not understand, or cannot > imagine, how the tones are compressed so that the 10 stops of subject > brighness are squeezed into the 5 stops of the paper. You have the squeezing done by the film and the squeezing done by the paper. Each is controllable - the gist of the Zone system is to take the scene brightness and translate it into the paper tones so it comes out the way you want it.
Under normal conditions no squeezing is done: a scene has a 7 stop range of brightness and paper has a 7 stop range of reflectance. By manipulating exposure, development and paper grade this range can be varied either to accommodate the subject or to convey the 'visualized image'.
For the short explanation see A. Adams Vol. 2 - "The Negative", and Vol. 3 - "The print"
For the long explanation see C. E. K. Mees - "The Theory of the Photographic Process", a book that spends 1100 pages answering your question.
Sorry to give a 'smart-arsed' answer, but the subject is either: really simple [it's all 7 stops / 2.1 OD and don't worry about it]; to moderately confusing and not-quite-right [A. Adams]; to mind boggling/numbing and like the Bible accepted on faith by most of us [Mees]
The gist of the whole thing can be shown be drawing HD curves, the sort of S-shaped curves that pop up everywhere in photography, and reflecting eye to brightness to negative to print to eye.
These curves must be somewhere on the web but I can not find them, so as part of a review of Mees* and the fair-use provision of the copyright act I have posted them at:
http://www.nolindan.com/UsenetStuff/Mees001.jpg
http://www.nolindan.com/UsenetStuff/Mees002.jpg
*Review of Mees, The Theory of the Photographic Process, Revised Edition: Damn good book, buy it. abebooks.com
 Signature Nicholas O. Lindan, Cleveland, Ohio Darkroom Automation http://www.nolindan.com/da/index.htm n o lindan at ix dot netcom dot com
Nicholas O. Lindan - 15 Oct 2006 19:46 GMT > "Tim Allen" <tim.allen@gmx.com> wrote
> > tonal compression ... negative ... brightness range ...paper...?
> The simplest explanation would probably be "rounding." > [quoted text clipped - 6 lines] > negative, they will be lost, "rounded" into the narrower range of the > paper. I am not sure of the original question: if it is about range in brightness then this effect may be secondary, it is however, scene _information_ is lost or compressed. The effect has to do with the grain size of the film and the size of the detail in the subject.
If the grain is coarse then a small area [say the size of 4 grains] can only have one of 4 shades of grey depending on how many grains have been developed. If the grain size is fine [say the same area has 64 grains] then 64 shades of grey can be shown.
If large areas of the film are measured the two films will show no difference in the number of shades of grey they can produce: 2,000 shades of grey looks no different than 20,000 shades of grey.
The effect is often refereed to as 'micro-contrast'. The reason given for the effect is usually wrong - lenses from a certain German firm being touted as having high 'microcontrast'.
When a resolution test is performed on films it can be shown that two films can resolve the same number of lines/mm and visually the two films have the same grain. However, if the same test is run using step tablets as the target the finer grained film will show a progression of square-shaped tones and the coarser grained film will show a white to black blur.
The 'look' of large and medium format is due to this effect. Even though grain may be invisible in a 35mm picture and in a medium format picture of the same resolution the MF picture will look, well, 'medium formatish' and better. $5 Zeiss Nettar trounces $5,000 Leica.
It is also why TechPan can [could] make large/medium format looking prints. TMax100 developed in Microdol has the same RMS grain as TechPan and the same resolution but TMax100 pictures will never be confused with anything large or medium format: they look distinctly 35ish. It is not the RMS visual grain or resolution that matters but the number of grains per square mm, and in this measure TechPan trounces TMax.
It is the same reason all printers may a resolution of 300 pix/inch but ones that put down more and finer dots look so much better.
 Signature Nicholas O. Lindan, Cleveland, Ohio Darkroom Automation http://www.nolindan.com/da/index.htm n o lindan at ix dot netcom dot com
Tim Allen - 16 Oct 2006 23:23 GMT Thank you Al for this great comparison. Now I got it!
Tim Allen
Floyd L. Davidson - 15 Oct 2006 23:23 GMT >Hello, > [quoted text clipped - 10 lines] >Thanks a lot! >Tim Allen You have (probably) two process where three characteristics are significant.
The three characteristics which are manipulated are the black point, the white point, and the shape of the curve used to go from black to white as brightness increases.
In an original scene there is a range of brightness, which may or may not be about the same as whatever is used to record it with (film or digital, for example).
In the first process (recording the scene), with film it is possible to manipulate the chemical development stages to adjust for different lighting in the original scene. With digital, post processing of the image date accomplishes the same task.
The second process is when the image is transfered to paper. Again, as with film, paper that is chemically developed can be manipulated while in the digital world the data is.
Hence, by using different film and processing, a scene that has shadows say at 5 fstops lower brightness than the brightest white levels, the shadows can be recorded with those shadows totally black, or significantly brighter. At least one fstop of variation is easy, and more is not difficult. The same is true of what is recorded as white or near white.
With film (as opposed to digital) these totally white or totally black points of recorded brightness are not abrupt, but instead represent a significant change in the slope (rate of change) rather than an absolute value. Hence a "linear" curve may exist for approximately 7 fstops of brightness, and either above or below that the curve no longer has the same slope. Everything above the upper "toe" of the curve is "white", and everything below the lower toe of the curve is "black".
The same is true of paper that is chemically developed.
Ideally a scene would match the film and the resulting negative would match the paper, and the print would *look* exactly like the original scene looked. But often one part or another does not match. If the scene has too large a brightness range to be recorded on film, there simply is no way to reconstruct it. But a narrow range can be expanded on the final print by using film that also has less range. Or conversely a wide range film can be used and a narrow range printing paper can be used. Different paper "grades" are used, basically to increase or decrease the contrast ratio available on the film in an attempt to match the final results (the print) to the original scene.
With digital it is a matter of looking at "curves" with an editing program, and moving the white or black levels.
Changing the shape of a curve is relatively difficult with chemically developed film or paper (different film or paper is selected to change), and extremely easy with digital. Hence that "linear" curve covering 7 fstops need not be linear at all. That is where "gamma correction" comes in. With digital processing it is very easy to dial in preset amounts of "gamma correction". It is nothing more or less than a non-linear transfer function that causes near white areas to be shifted more, or less, than near black areas.
 Signature Floyd L. Davidson <http://www.apaflo.com/floyd_davidson> Ukpeagvik (Barrow, Alaska) floyd@apaflo.com
|
|
|