Negative Notions: Proper digitization of stereoscopic negatives for parallel viewing

written for the stereosite by Ian Ference

Foreward

My inter­est in stere­og­ra­phy is pri­mar­i­ly informed by two things – my back­ground as a pho­tog­ra­ph­er, and his­toric stud­ies of the Great War. For this rea­son, I have amassed quite a few stereo­scop­ic neg­a­tives from ~1914–1921. The image I am using for the prac­ti­cal demon­stra­tion is des­ig­nat­ed “FasserNeg56”, a part of the Alexan­der Otto Fass­er Col­lec­tion, the most impres­sive and impor­tant of the many ama­teur Great War stere­og­ra­phy col­lec­tions in the Boyd/Jordan/Ference Col­lec­tion. It depicts a Nieu­port 10 sesqui­plane being worked on by the crew, and has a cor­re­spond­ing pos­i­tive that you can view on my per­son­al blog. Need­less to say, it is of great his­toric impor­tance, as are the oth­er 136 neg­a­tives in this par­tic­u­lar set.

Your inter­est may be else­where. While the images used in this arti­cle are of the war, and from the Fass­er col­lec­tion, the tech­niques are uni­ver­sal. Whether your inter­est is in trav­el stere­oscopy, 3D erot­i­ca, or macro stere­oviews, the the­o­ry and prac­tice described below apply equal­ly well to all sin­gle-sub­strate stereo­scop­ic negatives.

---

Introduction

Stereo­scop­ic neg­a­tives are, by nature of their cre­ation, trick­i­er drag­ons to con­quer than are those made by tra­di­tion­al two-dimen­sion­al cam­eras. They are vicious chimeras, prod­ucts of dis­tinct pho­to­graph­ic and stere­o­graph­ic process­es, and dif­fi­cult to tame, espe­cial­ly when we talk of those in Euro­pean-style. What do I mean by “Euro­pean-style”? Sim­ply that both parts of the stereo pair appear on the same sub­strate – usu­al­ly glass. Amer­i­can and British stere­o­g­ra­phers often used large-for­mat cam­eras to pro­duce paper cards. These cam­eras would cre­ate two neg­a­tives, one for each tak­ing lens. These are not dif­fi­cult at all to print (or dig­i­tize). For the remain­der of this arti­cle, I will be refer­ring to Euro­pean-style sin­gle glass neg­a­tives. The the­o­ry, of course, also applies to cel­lu­loid negatives.

Most sam­ple scans of neg­a­tives made by sell­ers on auc­tion sites are abject­ly ter­ri­ble. The above image rep­re­sents a bet­ter-than-aver­age attempt to dig­i­tize a stereo­scop­ic neg­a­tive, made by my late friend Doug Jor­dan. How­ev­er, as you exam­ine it, you should be struck by three prob­lems. Most obvi­ous­ly, the expo­sure is ter­ri­ble – you don’t even need to attempt to view the image to see that! Next, upon attempt­ing to par­al­lel view the scene, you’ll notice that you can­not – it will either appear flat or cause eye-strain. The pan­els are reversed, and it appears as if it were a cross-view image! This is not cor­rect. Final­ly, upon care­ful study, you might think that the sol­diers stand­ing about the mor­tar are wear­ing their medals back­wards. They aren’t – each pan­el is actu­al­ly a mir­ror image, and the entire stere­oview is flipped hor­i­zon­tal­ly! From the same neg­a­tive, a decade lat­er, I cre­at­ed the fol­low­ing dig­i­tal ver­sion, with­out much pain or effort:

In addi­tion to prop­er expo­sure in “devel­op­ing” the dig­i­tal neg­a­tive – an ana­logue to the dark­room con­cepts of expos­ing, using con­trast fil­ters, choos­ing a devel­op­er, etc – knowl­edge of how lens­es work, and appli­ca­tion of that knowl­edge to the more com­plex stere­o­graph­ic set­ting, allows me to quick­ly and eas­i­ly flip the entire “dig­i­tal print” hor­i­zon­tal­ly and then exchange the pan­els to cre­ate a nice rep­re­sen­ta­tion of what I per­ceive as the orig­i­nal photographer’s intent. Keep in mind that this involves some artis­tic inter­pre­ta­tion on my part – just like real dark­room work! But once ones grasps the the­o­ry intu­itive­ly, putting it into prac­tice makes dig­i­tiz­ing stereo neg­a­tives more excit­ing than dig­i­tiz­ing stereo pos­i­tives. Instead of being a pain in the arse, it’s an oppor­tu­ni­ty to exer­cise cre­ativ­i­ty and deci­sion mak­ing in inter­pret­ing the neg­a­tive into a final work. To the end of mak­ing this acces­si­ble to every­body, I will first dis­cuss the the­o­ry behind cre­at­ing stereo­scop­ic neg­a­tives, before address­ing the prac­tice of prop­er­ly dig­i­tiz­ing them.

Theory

Your eyes are lens­es. What hits the back of your eye is an upside-down rep­re­sen­ta­tion of what you are see­ing; your visu­al cor­tex steps in and sorts it for you. You may not know this, but cam­era lens­es do the same thing. Because many peo­ple pre­fer dig­i­tal pho­tog­ra­phy these days, you may not have strips of neg­a­tives lying around to proof that effect. So I will demon­strate with a sil­ly lit­tle car­toon drawn by my won­der­ful wife Stacey.

After devel­op­ing a neg­a­tive, you’ll wind up with a piece of glass (or cel­lu­loid) which looks like this:

You can then make a con­tact print from the neg­a­tive, with the emul­sion side of the neg­a­tive fac­ing the emul­sion side of the print­ing sub­strate (usu­al­ly paper or glass). This has the effect of again invert­ing the darks to lights, as well as the effect of revers­ing the image – which faith­ful­ly repro­duces the scene:

Now all you must do is take the print in front of you and rotate it 180º and voila! – a faith­ful rep­re­sen­ta­tion of the scene. Our pen­guin, igloo, and snow­man are all in their right place. Fol­low­ing from this, you might think that the same process should work on a stereo­scop­ic neg­a­tive. If you go into your dark­room and try it, you will get a dis­ap­point­ing result sim­i­lar to the first scan shown in this arti­cle, though not identical:

Two of the flaws of the ini­tial image are still appar­ent: the expo­sure is ter­ri­ble and the pan­els are reversed.So what of the third flaw – why is it not the case that on the con­tact print, the entire stere­oview is flipped hor­i­zon­tal­ly? The answer is sim­ple: Con­tact print­ing is not anal­o­gous to mak­ing trans­paren­cy scans on mod­ern scan­ners. In mak­ing a con­tact print, or enlarg­ing, the image is nat­u­ral­ly hor­i­zon­tal­ly reversed. This makes up for the inver­sion of the image. Scan­ning is pred­i­cat­ed on mak­ing an iso­mor­phic rep­re­sen­ta­tion of the phys­i­cal object. This is why emul­sion-side-down dia­pos­i­tive scans look just like the orig­i­nal trans­paren­cy – and why scans of neg­a­tives are always reversed.

How­ev­er, we still need to address the trans­po­si­tion of the pan­els – why the heck does nei­ther a con­tact print nor a dig­i­tal scan pro­duce a prop­er stereo­scop­ic image, free-view­able in par­al­lel or in more detail through a scope? The answer to this goes back to the notion of the fact that the Euro­pean-style stereo­scop­ic neg­a­tives under con­sid­er­a­tion con­sist of two images on a sin­gle sub­strate. Let’s con­sid­er our sil­ly pen­guin car­toon again – but this time, let’s con­sid­er pho­tograph­ing it with a glass-plate stereo­scop­ic cam­era. Let’s pre­sume that the pen­guin is at the zero plane, the igloo is clos­er to us (neg­a­tive par­al­lax), and the snow­man is far­ther away from us (pos­i­tive par­al­lax). What we want is to obtain a stere­oview that looks like this:

But let’s see what hap­pens when we load up a glass plate and click the shutter:

Both images are tak­en simul­ta­ne­ous­ly, and the light cre­ates a stereo­scop­ic neg­a­tive on our sin­gle glass plate:

Wait a minute – it says “right pan­el” on the left, and “left pan­el” on the right? Wor­ry not that your san­i­ty is slip­ping away, dear read­er.
 What we’re look­ing at here is the emul­sion side of the neg­a­tive. What the left lens saw is on the right; what the right hand lens saw is on the left. Visu­al­ize your­self stand­ing in front of a stereo­scop­ic cam­era and this should intu­itive­ly make sense. If you’re think­ing that this might have some­thing to do with the image trans­po­si­tion dis­cussed above, you’re on the right track. But to demon­strate, we’ll go ahead and make a con­tact print from this neg­a­tive, rotat­ed 180º as we would with a flat negative:

Uh-oh! This doesn’t look like the above stere­oview, and for a very salient rea­son: in rotat­ing it so that the fig­ures appear upright, we have placed the expo­sure made by the left-hand lens on the right, and the image made by the right-hand lens on the left. This is the rea­son that many stere­oviews (pos­i­tives) print­ed by ama­teurs from sin­gle glass plates appear to be slight­ly mis­aligned – they are! One side would have to be print­ed off the oppo­site side of the neg­a­tive, and then the oth­er, each while the pos­i­tive plate was half-masked. Now that we under­stand the process­es by which stereo­scop­ic neg­a­tives are cre­at­ed, we can hope­ful­ly bet­ter appre­ci­ate that although it takes a bit more work, we can make amaz­ing dig­i­tal recre­ations from the neg­a­tives in the “dig­i­tal darkroom”.

Practice

What fol­lows is an exam­i­na­tion of the process I use in cre­at­ing the best pos­si­ble dig­i­tal pos­i­tives from stereo­scop­ic neg­a­tives. It must be not­ed that, as with most things in dig­i­ti­za­tion (as well as the tra­di­tion­al dark­room), your mileage may vary as regards avail­able equip­ment and tools (soft­ware) – and that one’s own indi­vid­ual aes­thet­ic can come into play heav­i­ly when pro­cess­ing from a neg­a­tive. As stat­ed pre­vi­ous­ly, a neg­a­tive has far more data than a pos­i­tive – and there­fore, is far more open to inter­pre­ta­tion when print­ing or dig­i­tiz­ing it. The process of scan­ning a pos­i­tive is a study in cre­at­ing as close to a repro­duc­tion of the orig­i­nal work as is pos­si­ble. The process of cre­at­ing a dig­i­tal pos­i­tive from a scan of a neg­a­tive is an inter­pre­tive process. My own aes­thet­ic is to cre­ate true B&W works from B&W neg­a­tives, unless I have a cor­re­spond­ing pos­i­tive that shows the photographer’s intent to tone or tint the image. But it all starts with the scan.

Scanning

For most pos­i­tive scans, it is suf­fi­cient to scan direct­ly to JPEG or TIFF for­mat – minor alter­ations can of course be made, but gen­er­al­ly you’re just look­ing for an accu­rate repro­duc­tion of what is already estab­lished. How­ev­er, neg­a­tives are far more sil­ver-dense than are pos­i­tives. In this dense sil­ver, there is a lot more data. That’s why it is essen­tial to use soft­ware that can scan to the Adobe DNG (dig­i­tal neg­a­tive) for­mat, because this is the only cur­rent for­mat that ade­quate­ly rep­re­sents the amount of data con­tained in the den­si­ty of a neg­a­tive. I pre­fer Sil­ver­Fast, an inex­pen­sive and pow­er­ful soft­ware bun­dle that allows direct use of the scan­ner bed in “trans­paren­cy – glass” mode:

Sil­ver­Fast also offers the option of mul­ti­pass scan­ning – that is to say, tak­ing two pass­es over the neg­a­tive to cap­ture max­i­mal detail – which is anoth­er option not offered by the lousy soft­ware bun­dle that’s usu­al­ly deliv­ered with a scan­ner. In any case, the most essen­tial aspect of any scan­ning soft­ware is its abil­i­ty to cre­ate Adobe DNG files. Once you have picked a soft­ware prod­uct, it’s time for the fun to begin. Here’s my work­flow, in order:

1. Clean the neg­a­tive and scan­ner bed. This may seem like a no-brain­er, but most glass slides pur­chased from online auc­tions, estate sales, and so on are at least some­what dirty. Pure water (not tap!) can be used on the clear side of the glass, by appli­ca­tion with a microfiber cloth. On the emul­sion side, a dry cloth is prefer­able, but a lit­tle water vapor from breath­ing can come in handy with tough dirt / smoke dam­age. Always err on the side of cau­tion. Also make sure your scan­ner bed is clean and free of dust; a blow­er bulb can be your best friend.
2. Place the neg­a­tive emul­sion-side down and make a pre­view scan. If you’re not using a scanner/software com­bo that allows for place­ment direct­ly on the scan­ner bed, you may have to use a hold­er to hold your neg­a­tive. These are often prob­lem­at­ic, as the image area of a neg­a­tive can cov­er an entire plate and thus be blocked by the hold­er; when pos­si­ble, use the direct scan method. In any case, use pre­view scans to select the image area and make sure that, upon eye­balling it, the slide appears to have 0.1º or less vari­ance from being com­plete­ly straight. Every major action per­formed after the image is in your edit­ing soft­ware – such as unnec­es­sary rota­tions to cor­rect for slop­py scan­ning – only degrades image quality.
3. Dou­ble-check your set­tings. Acci­den­tal­ly scan­ning to the wrong fold­er or with the wrong image name only caus­es headaches. But much more impor­tant­ly, make sure you’re uti­liz­ing the entire Adobe RGB col­or space (more on why you should scan B&W neg­a­tives in col­or lat­er), and scan­ning at a high enough res­o­lu­tion. I rec­om­mend 4800 ppi if pos­si­ble, mul­ti­pass to DNG in 48 bit color.
4. Check your envi­ron­ment. If some­one in the flat above you is thump­ing their bass, you’re run­ning appli­ances which cre­ate vibra­tions (air con­di­tion­ers, loud phono­graphs, old hard dri­ves on the same table), and so on, you’re liable to get a less­er-qual­i­ty result.
5. Hit the scan but­ton. You’ll prob­a­bly have to wait 15+ min­utes for the result. Don’t for­get to check each scan before remov­ing the slide from the scan­ner – it might be the case that you need to res­can for any num­ber of rea­sons. Best to make sure you have a good base scan before clean­ing and plac­ing your next slide!

And that’s it – you should now have a DNG file that looks alto­geth­er unlike the orig­i­nal neg­a­tive. But don’t wor­ry about that; in our next step, you can make it look like the orig­i­nal neg­a­tive if you want – but you’ll prob­a­bly want to improve on it. One of the few upsides to mod­ern tech­nol­o­gy is that it’s much eas­i­er to improve on a dodgy neg­a­tive by mak­ing a dig­i­tal neg­a­tive and prop­er­ly pro­cess­ing it. At the very least, it wastes less pho­to­graph­ic paper!

“Developing”

Although it is com­mon par­lance in the dig­i­tal world to refer to pro­cess­ing a dig­i­tal neg­a­tive as “devel­op­ing” the file, this is rather sil­ly – you are already work­ing on a neg­a­tive that was devel­oped before you were born. Much bet­ter to refer to it as pro­cess­ing in my opin­ion. But I digress. This is the step where you do most of the fun­da­men­tal work of mak­ing the image look nice – that is to say, cor­rect­ing the bad expo­sure that we iden­ti­fied as one of the three major flaws of most neg­a­tive scans. Most stereo­scop­ic neg­a­tives are scanned direct to JPEG or TIFF. Our inter­im step allows you to do things that you can do in the dark­room – but instead of dic­tat­ing that you want greater edge acu­tance by using Agfa Rod­i­nal instead of Ilford ID11, you slide a lit­tle slid­er bar to the right.

In order to process a DNG file, you’re going to need soft­ware. I use Adobe Pho­to­shop CS6; oth­ers pre­fer Adobe Light­room after I con­vinced him into scan­ning to DNG. There are gen­er­al­ly some­what infe­ri­or free prod­ucts avail­able as well. The afore­men­tioned Adobe prod­ucts are eas­i­ly obtained in cer­tain fash­ions, but doing so is out­side the scope of this arti­cle. In any case, I will be using Pho­to­shop for pur­pos­es of explain­ing my work­flow; feel free to try the prod­uct of your choice and exper­i­ment – that’s half the fun any­how, is it not?

This is what FasserNeg56 looks like when opened in CS6:

And when opened in Pho­to­shop, and invert­ed (Command‑I):

Not too impres­sive, but rather char­ac­ter­is­tic of what you get if you don’t put in the work in edit­ing. Did I men­tion yet that you’ll be learn­ing to edit in reverse? You will. But that’s not as daunt­ing as it sounds, if it sounds daunt­ing at all – those who are still nos­tal­gic about their time in the dark­room are prob­a­bly look­ing for­ward to it, for though the smell of fix­er is not present, the adven­ture of inject­ing one­self into an image cer­tain­ly is. Those slid­ers you saw above, and the ones on the oth­er pan­els – these are your “dig­i­tal dark­room” tools. Expo­sure is anal­o­gous to the amount of light you use when enlarg­ing or mak­ing a con­tact print. Con­trast is the over­all con­trast of the image, where­as Clar­i­ty is sim­i­lar to edge acu­tance – the con­trast at bor­der regions between light and dark.

In any case, the best means of becom­ing expe­ri­enced in the use of these new­fan­gled slid­ers is by play­ing around with them – much like exper­i­ment­ing in a tra­di­tion­al dark­room! It took me dozens of tries to get what I want­ed out of the first stereo­scop­ic neg­a­tive I scanned; by my 20th I was get­ting it in one or two attempts. You learn to pre­vi­su­al­ize what will come out the oth­er side using this tri­al-and-error method, much as when you learned pho­tog­ra­phy, you learned to pre­vi­su­al­ize how a giv­en scene would look with what­ev­er lens and film you were shoot­ing. My final slid­ers for this image look like this:

With the end prod­uct (hav­ing been invert­ed) look­ing like this:

Is this the cor­rect expo­sure, con­trast, bal­ance, et cetera? There is no answer to this ques­tion; even though a pos­i­tive exists for this image, we don’t know that it was how the pho­tog­ra­ph­er (known to be Alexan­der Otto Fass­er) want­ed the image print­ed. As an Amer­i­can in Neuil­ly, it it far more prob­a­ble that he had the plates devel­oped and print­ed by a Parisian pro lab than that he took time away from his sur­gi­cal duties to do dark­room work him­self. Thus, this par­tic­u­lar print is my inter­pre­ta­tion of Fasser’s neg­a­tive, and can’t be oth­er­wise. And this is one of the lit­tle-known (these days) joys of work­ing with the neg­a­tives of anoth­er – one may, or indeed must, inject them­selves into the cre­ative process. Once you’ve refined your own cre­ative process, you’ll be ready for the final steps to pre­pare a pleas­ing par­al­lel-view­able pos­i­tive for dig­i­tal dis­play or printing.

Finishing Touches

So now we have a prop­er expo­sure at high res­o­lu­tion; we have over­come one of the three major obsta­cles dis­cussed in the intro­duc­tion. With Pho­to­shop, GIMP, or anoth­er soft­ware project, we can com­plete the final three steps. The first step is sim­ple – flip the slide horizontally:

Only one prob­lem to go! But first we must account for the slide’s oxi­da­tion. The sec­ond step is some­what more com­plex, and may vary depend­ing on the soft­ware prod­uct you are using. Remem­ber that we scanned to 48-bit col­or space, instead of sim­ply mak­ing a greyscale scan. This has left the oxi­dized sil­ver on the invert­ed image appear­ing as a blue/cyan fog over parts of the image – usu­al­ly the periph­ery. In Pho­to­shop, using the “Black & White” func­tion in “Adjust­ments” brings up the fol­low­ing menu:

And the cyan and blue slid­ers can be used to achieve a tonal match with the rest of the image, where­as a straight greyscale scan would not have allowed this! This can also be achieved in any num­ber of oth­er ways, includ­ing in the DNG pro­cess­ing stage. I chose this method because it is sim­plest to demon­strate; all that mat­ters is that what­ev­er means you use, you remove as much of the vis­i­ble oxi­da­tion as is pos­si­ble. We are only one step away from cre­at­ing a final image that we can be proud of. We must now do the third step – swap­ping pan­els and crop­ping. This is sim­ple to do in any image edit­ing soft­ware; I will share my Pho­to­shop work­flow, but you may fol­low what­ev­er makes log­i­cal sense to you – just don’t do any­thing to com­press the image! I start out by select­ing the entire image area, copy­ing it, and open­ing a new doc­u­ment (which uses the pix­el dimen­sions of the orig­i­nal, and will use whichev­er col­or I choose for the inter­sti­tial area between pan­els). I then copy what appears to be the right-hand side of the scan, and paste it into the left side of the new document:

Note that I care­ful­ly moved the new­ly trans­posed pan­el until all of the cen­tral over­lap area was cov­ered – this does mean a tiny amount of image loss, but that is more than made up for in cre­at­ing a pleas­ant view­ing expe­ri­ence. I then copy-paste the left-hand pan­el from the scan into the right side of the new doc­u­ment, and sim­ply crop off the over­lap area. Then I flat­ten the image, and am left with:

Of course, there is more that one can do to improve on this image. But retouch­ing stere­oviews is a top­ic for anoth­er arti­cle; suf­fice it to say that prac­tic­ing these tech­niques and work­flow allows one to process numer­ous stereo­scop­ic neg­a­tives in an hour, rather than spend­ing hours on each. There are a num­ber of oth­er things I could include here; for exam­ple, dur­ing the trans­po­si­tion step, it is very easy to paste the sec­ond pan­el on top of the first, des­e­lect the red chan­nel on that lay­er, and instant­ly have a pleas­ing anaglyph. If this makes intu­itive sense to you, then you might wish to add it to your work­flow, but for brevity’s sake I dis­clud­ed it from this article.

Conclusion

It is a shame that the mid-cen­tu­ry piv­ot away from stere­og­ra­phy (View-Mas­ter and Real­ist aside) has left a knowl­edge gap that has ren­dered most dig­i­tized stereo­scop­ic neg­a­tives ill-exposed, trans­posed, and back­wards. Most archives do not have a spe­cial­ist capa­ble of cre­at­ing prop­er par­al­lel rep­re­sen­ta­tions, and thus many are list­ed in find­ing aids but obnox­ious­ly unavail­able for view­ing. Those archives that do dig­i­tize their neg­a­tives some­times get the expo­sure right, but rarely prop­er­ly flip and trans­pose, lead­ing to the appear­ance of flat­ness and back­wards imagery. Read­ers who come from pro­fes­sion­al archival back­grounds should be able to quick­ly improve their skillsets, and thus improve the qual­i­ty of dig­i­tal rep­re­sen­ta­tions on their archives’ webpages.

Many stereo­scop­ic neg­a­tives linger on auc­tion sites for months or years; it is like­ly that the over­whelm­ing major­i­ty of these neg­a­tives have no cor­re­spond­ing pos­i­tive still extant, and thus are basi­cal­ly “lost arti­facts”. But things once lost can again be found, and hope­ful­ly this arti­cle will inspire some read­ers to seek out stereo­scop­ic neg­a­tives with­in their field of inter­est, dig­i­tize them, and make them avail­able to the pub­lic. Not only is this a pub­lic good, in that it helps to pre­serve his­to­ry, but it is a much more fun and cre­ative process than mere­ly dig­i­tiz­ing pos­i­tives. Just refrain from bid­ding against me when it comes to neg­a­tives por­tray­ing the Great War!

---

Afterward

For those of you who were intrigued by the the­o­ries behind the cre­ation of stereo­scop­ic neg­a­tives and how that impacts their dig­i­tal pro­cess­ing, there are fur­ther venues of explo­ration avail­able. In terms of the­o­ry of ana­log pho­tog­ra­phy, light­ing, lens func­tion, lens aber­ra­tions and coat­ings, and dark­room tech­nique, I high­ly rec­om­mend Aaron Sussman’s The Ama­teur Photographer’s Hand­book. The title might include ‘ama­teur’, but the sheer vol­ume of infor­ma­tion read­i­ly approach­es the entire­ty of some of my com­rades’ BA pro­grams. Mine is an 8th edi­tion, pub­lished in 1973, and hand­ed down to me by my father. I’m sure oth­er edi­tions are equal­ly won­der­ful. And for those with any ques­tions or who wish to dis­cuss any of this sub­ject mat­ter fur­ther, please reach out by email: ian@stereosite.com.