ittfb@dcatla.UUCP (Thomas F. Blakely) (02/09/88)
A handy tool around the Black & White darkroom these days is a transmission densitometer. While relatively inexpensive ones are available (ca. $200) it might be an interesting project to build one. Has anyone had any experience along these lines? It should be pretty straightforward, unless there's something I missed. Since it's for monochrome, color sensitivity of the sensing element (photodiode?) shouldn't matter. How about connecting the output to your PC (everybody has one, don't they)? How about hearing from those of you with experience, or even ideas? (Or, for the sci. electronics folks, maybe we could start talking about mercury- filled speaker wires again :-). Tom Blakely (404)442-4866 gatech!dcatla!ittfb
larry@kitty.UUCP (Larry Lippman) (02/09/88)
In article <2659@dcatla.UUCP>, ittfb@dcatla.UUCP (Thomas F. Blakely) writes: > A handy tool around the Black & White darkroom these days is a > transmission densitometer. While relatively inexpensive ones are > available (ca. $200) it might be an interesting project to build > one. Has anyone had any experience along these lines? It should > be pretty straightforward, unless there's something I missed. > Since it's for monochrome, color sensitivity of the sensing element > (photodiode?) shouldn't matter. Here are a few suggestions: 1. You MUST use an incandescent white light source; for general purpose use, a single monochromatic light source such as an LED is unsuitable. The transmission spectrum of B&W film stock is not flat, and varies from manufacturer to manufacturer; consequently, the use of a single monochromatic light source is an invitation to trouble. 2. In order to achieve sufficient signal-to-noise ratio in measuring higher densities (< 10% T, > 1.0 absorbance), LED's will not provide sufficient luminous intensity with a photodiode or photovoltaic detector, anyhow. So, you really need an incandescent lamp. 3. The incandescent lamp MUST be powered from DC (use a 6 or 12 volt lamp), and MUST be regulated using a solid-state constant-current regulator. I cannot overemphasize the importance of having good regulation for the lamp supply. 4. You may wish to consider using fiber optics to direct the light source to the sensing head. 5. The calibration control for setting 100% T (0 absorbance) can either vary photodetector sensitivity or vary lamp current; some commercial instruments use the latter approach (I personally don't care for it, though). If you use a microprocessor, no potentiometer is necessary - just a switch to tell the uP to take a reference reading. 6. Calibration of a meter for % transmission is simple, because it is a linear relationship. If you want an absorbance readout, you will require a logarithmic scale; there are clever analog circuits for this, or you can do the conversion in a microprocessor. In this day and age, I recommend the latter. 7. I recommend a photovoltaic detector, but a photodiode will work - provided that it has decent visible light response. Forget about using a CdS photoresistive detector. A photomultiplier tube is used on most commercial densitometers, but this is probably too complex and frought with potential problems for a DIY project. 8. Use a Kodak or equivalent density wedge to verify operation of your completed densitometer. 9. Beware of ambient light interference resulting from poor optics design. > How about connecting the output to > your PC (everybody has one, don't they)? Not a bad idea - as long as you have some type of analog input board. You could then do the absorbance calculation (if you want to deal in absorbance units) on the PC. > How about hearing from > those of you with experience, or even ideas? This is the kind of stuff I have been designing for the past 18 years, so you now have the benefit of some of what I've learned The Hard Way. :-) <> Larry Lippman @ Recognition Research Corp., Clarence, New York <> UUCP: {allegra|ames|boulder|decvax|rutgers|watmath}!sunybcs!kitty!larry <> VOICE: 716/688-1231 {hplabs|ihnp4|mtune|utzoo|uunet}!/ <> FAX: 716/741-9635 {G1,G2,G3 modes} "Have you hugged your cat today?"
beyer@houxs.UUCP (J.BEYER) (02/09/88)
R eferences: <2659@dcatla.UUCP> There is an ANSI standard on transmission densitometry that you should look up. While you could make a densitometer in ignorance of the standard, and while you could calibrate it to serve your personal needs, your numbers will not agree with anyone elses unless your machine conforms fairly closely to the standard. The light color is fairly important as is the sensitivity curve of the detector. Of greater importance is the angle of acceptance of the detector and the angle of illumination of the source. If you don't get these right, you will not be measuring diffuse density, but something else. THis will confuse you and everyone else.
korfhage@CS.UCLA.EDU (02/10/88)
Recently I was wondering about using my color meter as a transmission
densitomiter. I have the low end Beseler model, and it uses a photomultiplier,
seems quite sensitive and has a very small metering aperature.
The general idea, it seems to me, is to compare the unknown density with
some known references until you find a match, or until you get close enough
that you can make a good estimate. The question is how do you get your
reference? A calibrated Kodak step wedge runs about $81 at the local store,
and I didn't care to spend this much money on a whim. An uncalibrated wedge is
less (about $45, as I recall), but then you need to find a densitomiter you
can use to calibrate it. Even cheaper would be to make a series of negatives
of slightly varying exposure and calibrate those. You do have to check for
uneven density in the negatives, but you really need to calibrate only a
very small spot on the negative.
If anybody sees anything wrong with this scheme, I would like to know.
It would not be as accurate as a real densitomiter, but it would seem like
an inexpensive way to figure out 0.1 above film base + fog, or whatever,
with accuracy good enough for a hobbyist. And it would not require finding
a place to put another machine.
Willard Korfhage
ARPA : korfhage@cs.ucla.edu
UUCP : {ucbvax,ihnp4,randvax,trwrb!trwspp,ism780}!ucla-cs!korfhagedonl@foist (02/10/88)
In article <2659@dcatla.UUCP>, ittfb@dcatla.UUCP (Thomas F. Blakely) writes: > A handy tool around the Black & White darkroom these days is a > transmission densitometer. While relatively inexpensive ones are > available (ca. $200) it might be an interesting project to build > one. Has anyone had any experience along these lines? It should > be pretty straightforward, unless there's something I missed. I use an enlarging meter to measure density, under the enlarger light source. I either project the negative or lay it over the photocell. The meter gives linear values (a function of time and paper speed), so i take the log myself. There is an outfit (which seems to be basically one man) called Science & Mechanics Instruments or something like that, with plans for densitometers. The basic scheme is that they will sell you one of a couple of different meters, and plans for a light source, which is simply a wooden box with a bulb and brightness control, with a little stand for the meter and an arm for the probe. I agree that density measurements don't have to be very complicated, and it should be easy to build a densitometer in any of several configurations. The meter i use, which is also handy for general printing, is a Beseler i bought for about $75. It has proven itself plenty accurate and repeatable for my purposes. It's a little clumsier to use than a densitometer, because i have to multiply the ASA dial and time and then take the log, but for the money, it serves me just fine! - donl {sun,pyramid,adobe,decwrl,ucbvax,allegra,research}!sgi!donl donl@sgi.com
donl@foist (02/10/88)
In article <11237@shemp.UCLA.EDU>, korfhage@CS.UCLA.EDU writes: > Recently I was wondering about using my color meter as a transmission > densitomiter. I have the low end Beseler model, and it uses a photomultiplier, > seems quite sensitive and has a very small metering aperature. I recently bought a used Cosar Mornick color analyzer to reduce testing in color printing. I was quite excited that it had a density scale on it as well as the time and filtration scales, though i am not yet sure how accurate it is. The density scale is basically just the log of the time scale, and color filtration numbers are (supposed to be) density x 100. If the density scale were not present, the log of the exposure time scale would do fine. > The general idea, it seems to me, is to compare the unknown density with > some known references until you find a match, or until you get close enough > that you can make a good estimate. The question is how do you get your > reference? A calibrated Kodak step wedge runs about $81 at the local store, > and I didn't care to spend this much money on a whim. An uncalibrated wedge is > less (about $45, as I recall), but then you need to find a densitomiter you > can use to calibrate it. [...] There are different sizes of Kodak step tablets; the #2 is about 3/4 inch wide, and about 5 inches long. Mine is uncalibrated, and i think it was about $23. I've mounted one of these in a sheet of 4x5 film so that i can project it when i want, which makes for quick paper-testing exposures. > If anybody sees anything wrong with this scheme, I would like to know. > It would not be as accurate as a real densitomiter, but it would seem like > an inexpensive way to figure out 0.1 above film base + fog, or whatever, > with accuracy good enough for a hobbyist. I think there is a tendency to be too worried about how accurate the readings are in this sort of thing. I suspect that shutter and aperture variations, processing variations, metering problems, differing amounts of flare, light falloff in the enlarger, etc. all add up to a much larger variation than the inaccuracy of using an inexpensive meter as a densitometer. Reasonable tolerances suggested by the master are .09 to .11 for film speed, which is 10%, which should be pretty easy to meet. If you can show that the meter you are using doubles and halves pretty accuratly, the density readings you obtain with it should be more than adequate. I have found the uncalibrated step tablet to be accurate enough for my needs; i have found steps that are a little high or low using a rather inexpensive ($75) black-and-white enlarging meter as densitometer, but the overall slope is plenty close. - donl {sun,pyramid,adobe,decwrl,ucbvax,allegra,research}!sgi!donl donl@sgi.com
pwc@mitre-bedford.ARPA (Patrick W. Connors) (02/10/88)
In article <11237@shemp.UCLA.EDU> korfhage@CS.UCLA.EDU (Willard Korfhage) writes: >Recently I was wondering about using my color meter as a transmission >densitomiter. I have the low end Beseler model, and it uses a photomultiplier, >seems quite sensitive and has a very small metering aperature. >The general idea, it seems to me, is to compare the unknown density with >some known references until you find a match, or until you get close enough >that you can make a good estimate. The question is how do you get your >reference? A calibrated Kodak step wedge runs about $81 at the local store, >and I didn't care to spend this much money on a whim. An uncalibrated wedge is >less (about $45, as I recall), but then you need to find a densitomiter you >can use to calibrate it. Even cheaper would be to make a series of negatives >of slightly varying exposure and calibrate those. You do have to check for >uneven density in the negatives, but you really need to calibrate only a >very small spot on the negative. > >If anybody sees anything wrong with this scheme, I would like to know. >It would not be as accurate as a real densitomiter, but it would seem like >an inexpensive way to figure out 0.1 above film base + fog, or whatever, >with accuracy good enough for a hobbyist. And it would not require finding >a place to put another machine. > > Willard Korfhage In the section on film testing procedures in _The Negative_, Ansel Adams states that a spot meter may be used as a densitometer. According to Adams, every 1/3 stop mark on the meter scale is equivalent to .1 density units. To find .1 above fb+f, set your negative on a light table and take a reading of an unexposed developed and fixed negative, and that is your fb+f density. All you need to find now is the exposure which reads .1 more. All readings are relative to the fb+f frame so no other calibration is neccessary. I have tried this with my Pentax Spotmeter V and it works fine. There is a book called _Beyond the Zone System_ (I forget the author), that gives a couple of other systems for controlling contrast with an incident meter. This book is very technical, but there is a companion workbook that has plans for turning a spotmeter into a densitometer. Pat Connors
kar@ritcv.UUCP (Kenneth A. Reek) (02/11/88)
In article <2659@dcatla.UUCP> ittfb@dcatla.UUCP (Thomas F. Blakely) writes: >A handy tool around the Black & White darkroom these days is a >transmission densitometer. While relatively inexpensive ones are >available (ca. $200) it might be an interesting project to build >one. Has anyone had any experience along these lines? I have built one of these using a CDS cell from Radio Shack. As was pointed out, the spectral sensitivity of the cell (and the lamp, too, for that matter) is of little interest to B&W workers. The problem is to get the answer in the right kind of numbers. I connect an ohmmeter to the cell and read its resistance in ohms; this value is determined by the density of the portion of the negative I'm reading. Unfortunately, it is proportional to opacity rather than density, so to use these values you need a step wedge of known densities with which to "calibrate" your ohmmeter (e.g. knowing that 42,500 ohms corresponds to .65 density on the step wedge and 49,350 ohms is .8 density on the wedge allows you to estimate pretty well the density that gives a reading of 46,000 ohms). I worked out my calibration in this way and got surprisingly good results, verified by readings on a Macbeth densitomiter here at RIT. Be prepared to recalibrate the thing each time you use it, however, unless you have a regulated power supply for the lamp, and have a lamp whose output changes little during its lifetime. The cell might be sensitive to temperature, too, and God knows what else. To get a readout directly in density units, you need an amplifier with a logarithmic characteristic to convert the resistance of the cell (proportional to opacity) into density (log of opacity). I've not yet done that. -- Kenneth A. Reek Rochester Institute of Technology {allegra,seismo}!rochester!ritcv!kar
beyer@houxs.UUCP (J.BEYER) (02/11/88)
When I bought a T21 scale, it was about $6. The steps are supposed to be about 0.15 density units apart with the 'clear' patch at about 0.06 density units (should these be called Driffields? :-) ). I measured mine with a MacBeth TD901 densitometer and I feel the densities listed are close enough for most work. I'd worry about exposing film and hoping the densities come out as expected. The reason I bought my densitometer is to find out if they are as expected. After over 10 years, I finally get what I expect sometimes ;-). You might try using Neutral Density filters, but, unless you get them somewhere much cheaper than my local dealer, a useful set would cost more than a calibrated wedge. The T21 is about 1/2 inch wide and 2-1/2 inch long.
donl@foist (02/12/88)
In article <663@houxs.UUCP>, beyer@houxs.UUCP (J.BEYER) writes: > R > eferences: <2659@dcatla.UUCP> > > > There is an ANSI standard on transmission densitometry that you should > look up. While you could make a densitometer in ignorance of the standard, > and while you could calibrate it to serve your personal needs, your > numbers will not agree with anyone elses unless your machine conforms > fairly closely to the standard. I disagree. I think that it's fairly easy to calibrate against full-stop increases and decreases, and that if these are right and you can show that the meter is pretty linear, you should be pretty gosh-darn close to any body else's numbers. > The light color is fairly important as is the sensitivity curve of the > detector. Of greater importance is the angle of acceptance of the > detector and the angle of illumination of the source. If you don't > get these right, you will not be measuring diffuse density, but something > else. THis will confuse you and everyone else. Sounds like you're talking about reflection densitometers, which would definitely be more difficult. The angle of illumination for transmission is pretty easy to arrive at -- straight through. And i haven't even read the ANSI standard! - donl {sun,pyramid,adobe,decwrl,ucbvax,allegra,research}!sgi!donl donl@sgi.com
beyer@houxu.UUCP (J.BEYER) (02/12/88)
I checked last nite: it is a T14 scale and cost $6. a few years ago.
briand@tekig4.TEK.COM (Brian Diehm) (02/14/88)
>>It would not be as accurate as a real densitomiter, but it would seem like >>an inexpensive way to figure out 0.1 above film base + fog, or whatever, >>with accuracy good enough for a hobbyist. And it would not require finding >>a place to put another machine. > > In the section on film testing procedures in _The Negative_, Ansel >Adams states that a spot meter may be used as a densitometer. >According to Adams, every 1/3 stop mark on the meter scale is >equivalent to .1 density units. To find .1 above fb+f, set your >negative on a light table and take a reading of an unexposed developed >and fixed negative, and that is your fb+f density. All you need to >find now is the exposure which reads .1 more. All readings are >relative to the fb+f frame so no other calibration is neccessary. >I have tried this with my Pentax Spotmeter V and it works fine. This does indeed work, and work well. However, I've found a few warnings from my experience. 1) Use a broad light source - the light table suggested above will work fine, as will a large globe light PROVIDED THE NEGATIVE AND METER ARE BOTH VERY CLOSE TO THE LIGHT. If you don't fill the frame of the spotmeter with the light source, you will get specular density rather than diffuse density. Because silver particles actually block photons in bunches with the grain, the specular and diffuse densities of B&W materials vary widely; with color materials the silver is washed out and the remaining dyes scatter the light rather than block it. This is also why B&W negatives will shield your eyes while viewing a partial eclipse but color slides won't. And why diffuse enlarging light (cold light) makes a difference in B&W printing but matters less when printing color. Anyway, diffuse density is what most agrees with densitometers, in my experience. 2) The fluorscent source of many light tables isn't corrected in any way for color balance; if yours isn't then you may have some spectral response mismatch between your meter and your source. Theoretically this doesn't matter if you are comparing densities. However, you might want to be cautious. I have used tungsten with good matching to real densitometers. The relationship between 1/3 stop and 0.1 density is very close to absolute; you needn't take Adams' word for it. Basically, it is because the base 10 log of 1/3 is very nearly 0.1 - to about 4 places. If you're getting into this, it is a useful exercise to work out this relationship for yourself; you will get a good understanding of the sensitometry and mathematical relationships. Finally, one-hour labs all have process densitometers for color QC. I have gotten gray-scale density readings from these labs simply by taking in my 4x5 negatives during low-demand hours; invariably they are willing to give me a few readings for a smile. -- -Brian Diehm (SDA - Standard Disclaimers Apply) Tektronix, Inc. briand@tekig4.TEK.COM or {decvax,cae780,uw-beaver}!tektronix!tekig4!briand
jtn@potomac.ads.com (John T. Nelson) (02/15/88)
The title of this discussion sounds like the title to a Tom Swift book... "Tom Swift and his Homebuilt Transmission Densitometer." -- John T. Nelson UUCP: sun!sundc!potomac!jtn Advanced Decision Systems Internet: jtn@potomac.ads.com 1500 Wilson Blvd #512; Arlington, VA 22209-2401 (703) 243-1611 "Hi... My name is Hobbes. I'm the product of a malicious 5-year old's twisted and destructive imagination. Would YOU like to be my friend?"
beyer@houxs.UUCP (J.BEYER) (02/15/88)
The trouble with makeing a transmissioni densitiometer at home involves the Callier effect. If collimated light is passed through a diffusing medium, such as exposed and developed film, it is scattered. If a densitometer is made by passing collimated light through the film and the sensor is 'far' from the film, only the unscattered light is measured. This is measuring specular density. If collimated light is passed through the film and all the transmitted light is gathered by the detector (by using an integrating sphere or suitable opal glass arrangement) the setup measures diffuse density. [it is equivalent to pass diffuse light through the film and use a detector 'far' from the film.] If diffuse light is incident on the film and a detector measures all the light transmitted through the film, doubly-diffuse density is being measured. These readings will all be different, especially in the higher density areas. Now in most books and articles, diffuse density is what is referred to. If your densitometer measures something else, you can get confused. If you report something else, you may confuse others. What density should you use? If making contact prints, you probably want diffuse density. If enlarging with a diffusion enlarger, you probably want diffuse density. If enlarging with a condenser enlarger, you could probably use a specular density measurement, but that raises the issue of how specular. The ANSI standard does define 2 kinds of specular density but these seem to be related to movie and slide projection. It varies depending on the aperture of the optical system.
beyer@houxs.UUCP (J.BEYER) (02/16/88)
While I don't normally measure density with an enlarger and enlarging photometer, I think it is a pretty good way of controlling the process, and even takes into consideration things like enlarger flare, Callier effect, etc. However, the readings should not be equated to those measured on a true densitometer. (They may even be *more* meaningful in some applications, but they are not the same.) To get an idea of what is going on, and to calibrate 3 new boxes of paper I just bought, I ran some tests last weekend. Those results that I thought might interest others are posted here. Relative speeds of the various papers are not, since the box-to-box variations are the important ones that must be measured individually. I normally use a Macbeth TD901 transmission densitometer, a Beseler 45MCRX enlarger with Aristo cold light head, or Beseler condenser head, and sometimes use a Beseler PM2a color analyzer. As a test, stimulated by recent postings here on netnews, I measured a Kodak T-14 control scale four ways. With the densitometer, over the color analyzer, and projected through a Schneider Componon S 180mm f/5.6 lens. When projecting, the scale was 'pretty well' surrounded with an opaque (litho-film) mask. When using the densitometer, I placed the emulsion side up (toward the diffuser and detector). In all other cases, the emulsion side was down (toward the detector -- where the paper would be). The "white" filter was used. Step D D D D Number TD901 Scale on Scale Scale 3mm aperture Analyzer Projected Projected ColdLite ColdLite Condenser ------------------------------------------------------------------------- 1 0.056 0.06 0.07 0.1 2 0.184 0.2 0.2 0.27 3 0.313 0.34 0.35 0.46 4 0.463 0.5 0.52 0.69 5 0.612 0.66 0.69 0.89 6 0.764 0.82 0.85 1.08 7 0.916 0.98 1.0 1.29 8 1.08 * 1.16 1.48 9 1.22 * 1.32 1.68 10 1.39 * 1.51 1.90 11 1.55 * 1.7 >2.0 12 1.71 * 1.88 13 1.86 * >2.0 14 2.02 >2.0 * indicates too difficult to read (can't see where the steps are). Assuming the TD901 to be the most accurate (I believe this assumption), this seems to show that the PM-2A is pretty accurate, tracking the TD901 quite closely and linearly. Projecting with the cold-light head is also quite close: I believe the slight upward curvature is due to light falloff at the edge of the image where the higher densities on the control scale were. A separate measurement seems to confirm this. The Callier effect is quite pronounced with the Condenser head, but it seems to have more even light distribution than the cold light head. Makers of home-made densitometers can expect similar differences if they are not careful about illumination and collection angles of their instruments. Such makers are urged to plot the graphs, either from my data, or (preferably) their own to see how much this difference is. For speedy printing, I don't bother with the color analyzer (in B+W, that is), I just set the enlarger for an 8x10" print, put the 4x5" negative in, and print 10 seconds at f/11. Sometimes this comes out right. Other times it is close enough so the next print is correct. This is only true for "average" scenes. I do use the analyzer when changing sizes, but one does have to calibrate each box of paper.
donl@foist (02/17/88)
In article <191@ritcv.UUCP>, kar@ritcv.UUCP (Kenneth A. Reek) writes: > In article <2659@dcatla.UUCP> ittfb@dcatla.UUCP (Thomas F. Blakely) writes: > >A handy tool around the Black & White darkroom these days is a > >transmission densitometer. While relatively inexpensive ones are > >available (ca. $200) it might be an interesting project to build > >one. Has anyone had any experience along these lines? > > I have built one of these using a CDS cell from Radio Shack. As was > pointed out, the spectral sensitivity of the cell (and the lamp, too, for > that matter) is of little interest to B&W workers. The problem is to get > the answer in the right kind of numbers. I like this because it's cheap and it works! A handy programmable calculator or equivalent might be an easier way to get from the wrong numbers to the right numbers. If the response of the cell is linear, you should be able to come up with an equation to turn it into proper density values. I set up mine using the first value as zero; it gets subtracted from all the rest as film base + fog. - donl {sun,pyramid,adobe,decwrl,ucbvax,allegra,research}!sgi!donl donl@sgi.com
mikkel@cg-atla.UUCP (Carl Mikkelsen) (02/17/88)
In article <2659@dcatla.UUCP> ittfb@dcatla.UUCP (Thomas F. Blakely) writes: >A handy tool around the Black & White darkroom these days is a >transmission densitometer. While relatively inexpensive ones are >available (ca. $200) it might be an interesting project to build >one. Has anyone had any experience along these lines? I have built an analogue enlarging meter/densitometer that I use routinely for both B&W and colour enlarging, although theoretically it should not work as well for colour. I have a CDS cell on a probe that measures the amount of light transmitted through the film and delivered to the paper. There is a "pot" to null out the intensity of any particular portion of the negative (such as base, point of full blackness, etc), so that by subsequently probing the highlight areas, can read "stops of negative contrast" directly. Based on negative contrast, I pick a polycontrast filter (yes, I am not a purist). With the proper filter, I set the pot to a calibrated position, and adjust the f-stop to null the meter reading. At this point, I have a pretty good (often "near perfect") selection of both contrast grade and exposure. From this point on, I adjust for special effects desired, dodge and burn, etc. Internally, the meter contains two (2) crude log-amps, a differential amplifier, and a metering circuit. The whole thing is powered by two 9v alkaline cells (that have so far lasted for over two years), that also drive two small lights to illuminate the meter. BTW, although probably obvious to you all, I did not anticipate a need to turn off the safelights when making meter readings. I modified my enlarger timer to kill the safelights by adding a DPST NO relay energized by the enlarger circuit switching the safelight circuit. All parts came from Radio %$#@ or my junk box. I have also a Beseler Analite 400, and find it to be total trash compared to this meter. May your good shooting come out fine in the soup! +--------------------------+--------------------------------+ | Carl Mikkelsen | ..!ism780c\ | | | ..!cbosgd!ima>!cg-atla!mikkel | | Compugraphic Corporation | ..!ulowell/ | | 200 Ballardvale St. | ..!decvax/ | | Wilmington, Ma. 01887 | (617) 658-5600 x 5220 (voice) | | | (617) 658-0200 x 5220 (TT-auto)| +--------------------------+--------------------------------+