Are Digital Images Really Free? By John Kossik
John Kossik has been on the site before, with interesting insights into shooting film and camera rebuilds. It is safe to say John has an inquiring mind. And now he addresses an issue that I have been thinking about for a long time. What sort of impact the digital revolution has been having on the environment. John presents us with a clearly researched and well composed piece on the relative effects of film and digital photography on the environment.

Abstract- A preliminary evaluation was conducted of the consumer yearly cost and the yearly carbon footprint of the use of a modern full frame digital SLR and a modern film SLR. This evaluation came up with a yearly consumer cost of the digital SLR to be $ 533/yr while the similar yearly cost of the film SLR ranged between $ 855/yr to $369/yr. In the evaluation of yearly carbon footprint of these two image-capturing methods it was found that the yearly CO2 emissions from a digital SLR was 5.5 kg while that of the film SLR was 2.7 kg. Both these values are probably skewed in favor of the digital SLR as the recyclability and biodegradable potential of photographic film were not fully accounted for in these methods.

INTRODUCTION:

It is obvious for the professional photographer that digital imaging has distinct advantages over traditional film photography. But this fact is predominantly due to the vast number of images professionals take and their need to quickly transmit and/or print these images. Their very livelihood in most cases depends on it. Professional photographers are of course not what allows imaging, whether digital or analog, to exist.
Imaging technology was created and is only really sustainable if the general public embraces it. The needs of the casual photographer greatly differ than those of the professional. The casual photographer takes far less images and is not under the same time constraints. Due to this the difference the monetary and environmental difference between analog and digital for the majority of the public is far more difficult to assume.
Since the general public is not burdened by an immediate cost when taking a digital image the first assumption is that the image is “free.” This is compounded by the tangible immediate cost seen by the customer of “old-style” analog images in the form of purchasing film and getting it developed. Though the costs of film and its processing are small compared to the purchase cost of a digital camera, the cost of a digital camera is a one-time event while with film the customer is continually reminded that there is a cost, be it incremental, to produce images.
This paper will try to normalize these costs by concentrating on the “life-cycle” or yearly costs of digital and analog imaging. In a like manner the environmental costs of each of these imaging methods will be compared by attempting to assign a yearly Carbon Footprint to both digital and analog capture.

YEARLY CONSUMER COST OF DIGITAL V.S. FILM:

To compare these costs we will be looking at the consumer costs of the actual cameras used to capture these images and well as the cost in the analog case of the film and its development. To get a comparative cost of the camera hardware itself it is assumed that a new digital and film camera is bought. It is also assumed that the cost of these cameras are the same. To arrive at this cost the list prices of a D610, D750, F6, and FM10 were found from Nikon’s website.[i] These prices were averaged to come up with a hardware cost for both digital and analog of $1,660.

To determine the yearly cost of a full frame digital imaging camera first the typical useful lifetime of this device must set. The average operational lifetime of a digital camera has been estimated at 3 years.[ii] This lifetime is not usually determined by the physical failure of the device, as the typical mode of failure of a modern digital SLR will be shutter failure and modern shutters are rated for between 100,000 to 300,000 actuations[iii] which everyone other than professional photographers will never reach. The lifetime of a digital SLR is usually determined by the evolution of technology and it becoming technically obsolete in 3 years. Using this lifespan the calculation of the yearly cost of a digital full frame SLR becomes simply:

$1,660 ÷ 3 Years = $533/ yr for a digital full frame SLR for a digital full frame SLR

To determine the yearly cost of a film SLR the cost of the film and its developing must be included. In doing so an actual number film images per year has to be arrived at. In addition the yearly capital cost of the film SLR needs to be determined. Due to the advancements in technology usually being limited to the film itself, film cameras typically have a much longer lifespan than a digital cameras.
It also must be accounted for that a person using a film camera will take much less shots than one using a digital camera. Some quantitative data on this needs to be determined before we can move forward. For this I will fall back on my own personal experience with my own cameras, specifically the digital SLR I currently use, a Nikon D300.
I probably take more images than the typical camera user, especially when my kids were still in high school and I took pictures for most of the sports teams in which they participated. Combine this with trips all over the US and to the UK and France and this camera has seen a lot of use. Yes, I have been using it far longer than the typical lifespan of 3 years. As of this November the shutter count on my D300 was about 65,000. I have owned and used this camera extensively since I purchased it in June of 2008. Thus I have had it for about 90 months or 7.5 years. Doing the math my yearly shutter count is thus:

65,000 shutter actuations ÷ 7.5 years = 8,667 actuations/year

Of course if this were a film camera I would have a far smaller number of actuations per year as the “Spray and Pray” method used especially in sports and wildlife photography would be far too costly. But how to come up with a reasonable value for the equivalent number of film camera actuations? Well here again I have to fall back on my own personal experience.
I was recently down in Florida for a wedding in which I took a substantial amount of digital images. Along with this were digital images taken down there at various beaches and on a kayaking trip through the Everglades. Due to the number and variety of locations in which I used my digital camera at this time I feel it was fairly typical. Upon arriving home and evaluating the images I took, as well as post-processing in Photoshop, I culled the images that were useful to upload and thus share with those concerned. Of the 903 digital images taken during this trip I ending up post-processing and uploading 184 of them. This gives me a “usable” image value of:

184 ÷ 903 x 100% ≈ 20% 

If we assume that the number “usable” images from my digital camera equals the number of film images I would have taken instead, then we have the following number of film images typically taken in a year:

(8,667 actuations/yr)× (20%) =1,733 filmimages/yr

Now we have to determine the lifespan of a film SLR. Using again shutter actuations as the limiting factor here, but this time using 50,000 actuations as the limiting value as this was typical in the days when film was dominant,[iv] we get:

50,000 actuations ÷ 1, 733 actuations / yr = 28.9 yr 

With this lifespan and the same initial cost as a digital SLR we get the fixed cost to be:

$1,660 ÷ 28.9 yr = $57.4 / yr 

Now of course we have to add on the cost of film and its development. Taking 48 different types of negative and B&W film we get an average cost of a 36 image roll to be $6.56.[v] My friendly neighborhood photo store here north of Seattle will develop, scan, and give you a CD of a 36 image roll of C-41 processed film for about $10.[vi] This gives us a cost per image as:

$6.56 + $10 ÷ 36 images = $0.46 /image 

Using the number of film images per year determined above we can get the cost of film and its developing per year as:

(1,733 images/yr)×($0.46/image)=$797.18/yr

Adding the yearly cost of the film camera itself we get the overall yearly cost of the film SLR as:

$57.4 / yr + $797.19 / yr = $854.6 / yr for a film SLR

By these calculations the yearly cost of the digital full frame SLR is $533.3 and that of a film SLR is $854.6, or the cost of a digital SLR is 62% of a film SLR.

Now, one could argue that the above analysis is skewed in favor of digital as the cost for film includes developing. But then you say, if you do not develop the film then you have no images to look at. The competing argument in digital would say that even though the images are on the flash media in the camera you still need a computer to view and show them as well as some type of internet access to transmit them. The fact that digital imaging is not viewable without these secondary and many times ignored elements could be viewed as significant.
If you consider the latent image on the film in your camera the same as the digital information on your flash media (both of which need further processing to create a usable image to view) then the yearly cost of film photography in this analysis should only include the cost of the film itself and not the development. If this approach is taken the cost for film is reduced to $0.18/image and the cost of film photography reduced to $369.34/yr (the author will allow the reader to do the math on this). If this is the case then instead of digital imaging costing less than film it now is 144% the cost of film imaging.

Either way you look at it the consumer costs of digital and film imaging are similar, and thus digital photography is far from “free.”

YEARLY CARBON FOOTPRINT OF DIGITAL V.S. FILM PHOTOGRAPHY:

Carbon footprinting has been a growing concern lately and is seen as one of the metrics in which companies and individuals measure their effects on the environment. This is useful as the consumer cost of an item in most cases does not reflect the “true” cost of this product.
Costs, especially environmentally and socially are not included in the purchase price and thus carbon footprinting is one way in which to try to quantify these. Carbon footprinting is far from an exact science but it is a useful tool in getting a handle on order-of-magnitude comparisons between products and their modes of production. This is what we are going to attempt here in comparing digital and film imaging.
To do this we are going to concentrate only on the image-capturing elements of these two technologies, namely the digital sensor and one frame of photographic film. This approach implies that the hardware carbon footprints of both digital and film are the same, minus the digital sensor itself. This is a good for a first approximation but will probably end up estimating the carbon footprint of digital imaging to be smaller than it actually is due to the fact that digital cameras are used for a much shorter time frame than film cameras.

First we will look at the carbon footprint of one frame of photographic film. A single frame of 35mm photographic film measures 35mm x 37mm. It is primarily made up of three components, Cellulose Triacetate base, Gelatin, and Silver embedded in the Gelatin.[vii] Of this film the Cellulose Triacetate base has a thickness of about 125 microns[viii] and a weight of 32.3lb/1000sq.ft.[ix] The Gelatin layer is about 38 microns[x] with a density of 1.3 gram/cm3.[xi] As for the amount of Silver it is approximated that for 35mm film 5000 images can be produced from one troy ounce.[xii] To determine the carbon footprint of one film image we must first find the mass of each of the components and then use estimates of how much CO2 is produced in creating these various amounts of material. From the size of a 35mm film section we have the following surface area of one image:

(35mm)×(37mm)=1296mm2 =12.96cm2

First start with Cellulose Triacetate:

(12.96cm2) × (32.3lb/1000ft2) = 0.2042 grams/image

Now the Gelatin:

(12.96cm2) ×(38microns)× 1.3 grams/cm3 =0.064 grams/image

And lastly for Silver:

troyoz ÷ 5000 images = 0.0062 grams/image 

Now we have to assign a CO2 footprint to each of these materials. These footprints are traditionally given in kg of CO2 per kg of the material in question.

For Cellulose Triacetate we have an estimate of 3.5 kgCO2/kg.[xiii]
For Gelatin we have an estimate of 3 kgCO2/kg.[xiv]
For Silver we have an estimate of 100 kgCO2/kg.[xv]

Thus we get a carbon footprint of one frame of photographic film to be:

(0.2042 grams×3.5kgCO2 /kg)+(0.064 grams×3kgCO2 /kg)+(0.0062 grams×100kgCO2 /kg)=0.0015kg

Converting this to CO2 per year using the number of film images taken per year denoted earlier we have:

(0.0015kg)×(1,733 images/yr) = 2.65kg/yr Carbon Footprint for Film

Now we need to determine the same carbon footprint for a digital full frame sensor. Digital sensors are produced in the same manner and in many cases the same facilities that are computer chips. Due to the extremely clean environments needed to produce these devices and the numerous aggressive chemicals needed to do the multiple etching steps involved, these small devices have a very large carbon footprint. Combine this with the relatively high failure rate of sensors due to contamination during production, some ranging up to 35%,[xvi] and the environmental costs of these digital devices are very high despite their small size. We start by determining the energy used to make a 300mm diameter wafer that these digital sensors will be cut out of. The energy to produce one silicon wafer is estimated at 833.16 kWhr.[xvii]   The conversion of energy used to CO2 emissions of course depends on the nature of how the energy, in this case usually electricity, is generated, for this estimate a value of 1.35 lb CO2/kWhr was used.[xviii] Thus for one wafer we have:

(833.16kWhr) × (1.35lbCO2 / kWhr) = 1124.8lb / wafer = 510.2kg / wafer
To include the electrical circuitry and cutting of the digital sensor out of the wafer each sensor measures about 40mm x 28mm.[xix] Cutting these out of a 300mm diameter wafer allows about 48 sensors/wafer. With a 65% success rate (called die yield technically)[xx] we get:

(48sensors / wafer) × (0.65) = 31.2usablesensors / wafer

Dividing this into the CO2 emitted per wafer and we get the carbon footprint of a digital sensor:

510.2kg/wafer ÷ 31.2sensors / wafer = 16.35kgCO2 /sensor 

Per our discussion earlier we know that a typical digital camera lasts 3 years so:

16.35kgCO2 / sensor ÷ 3yr / sensor = 5.45kg / yr Carbon Footprint for Digital

This estimate for digital is probably low as it does not account for all the other parts of the digital camera that will be thrown away in three years, but it gives us a first estimate. Using this estimate we see that the carbon footprint of a digital image sensor is at least 206% of that of a film image.

Now the above carbon footprint does not include the development processing of the latent image on the negative for film. One could argue that this should be included thus raising the carbon footprint of film making it closer to that of digital. Perhaps in a future evaluation this carbon footprint of film developing can be added. I would imagine though that the development carbon footprint per image would be very similar to that of the film itself, essentially doubling its overall value. In this case the carbon footprints of film and digital would be very similar as it was for their consumer cost.

FUTURE WORK:

The estimate for the number of film images a casual photographer will take in a year needs further analysis as it has a large impact on the cost and footprint of film. I personally think that the value used here is high making film more costly and having a larger carbon footprint than it should, but I will leave it to others to address this point.

In addition, the environmental impact of digital (other than carbon emissions) was not really fully accounted for in this article due to its inherent un-recyclability. The silver in photo film is almost 100% recoverable for color film and some 60% recoverable in B&W film.[xxi] Matter of fact, even the perforations made in 35mm film are collected to recover the silver therein.[xxii] Also the chemicals used in both C-41 and B&W processing (and E-6 but not to the same extent) are relatively simple and far more benign than those used in semiconductor manufacture.
Home developing aside, these chemicals can be recovered, reused, or repurposed fairly simply by modern chemical processing standards. If the markets were large enough commercial development labs would have places to send their spent chemicals so that recovering of the materials beyond that of silver could be optimized.

Also, the resulting film negatives are extremely benign. Even the silver is not as “bad” as people claim it to be. After all it has recently gained a renewed use as a bactericide, as our grandparents knew very well. That’s why they used real Silverware, not utensils made from stainless steel.
As for biodegradability, well the Cellulose Triacetate and especially the gelatin in film are easily returned to nature, matter of fact keeping it from doing so is one of the major challenges of archiving traditional negatives.

Of course none of this is true for semiconductors. Perhaps a very small amount of a printed circuit board can be recycled or repurposed, but virtually none of this is true for the semiconductor chips and their related enclosures. Combine this with the fact that most semiconductor-based products made today are designed specifically to be un-repairable and un-recyclable as well as having a short lifespan (so people buy new ones, Apple is one of the largest offenders in this case) and we have the tragedy of piles of E-waste accumulating in poorer countries.

This is one of the items that should be explored in future analyses with the understanding that the recyclability of film and the un-recyclable nature of digital needs to be factored in somehow.

CONCLUSIONS:

The above is of course just a first approximation of the comparison between digital and film imaging as it pertains to yearly consumer cost and yearly carbon footprint. The quantities calculated here are subjective especially when it comes to the carbon footprints which are difficult to quantify accurately. I welcome others to look at these parameters quantitatively and hopefully come up with estimates better than those I have given here. That said, this at least lends some quantitative data to the continuing evaluation of these two image-capturing formats beyond the typical qualitative comparisons widely present in the past.

In the final analysis, due to both these evaluation parameters being of the same order of magnitude for both image-capturing formats, we can confidently see that Digital is not Free.

ACKNOWLEDGMENTS:

I would like to thank Robert L. Shanebrook (http://www.makingkodakfilm.com/) and E.J. Peiker (http://www.ejphoto.com/) for their valuable input to this article.

John Kossik  (jmk@63alfred.com) graduated in 1983 with a degree in Chemical Engineering from Michigan State University, he currently works for Beacon Engineers, Inc. (www.beaconengr.com) a small industrial engineering firm in Bothell, WA, USA and is co-founder of Steadfast Equipment (www.steadfastequipment.com). Some of his images can be found at http://63alfred.smugmug.com/. You can also see John’s other articles on JCH by clicking here.

Thanks it https://www.ifixit.com/ for the images of e-waste.

Notes:

[i] Nikon USA website, [http://www.nikonusa.com/en/index.page], (12/5/15)
[ii] Viability of recycling semiconductors in imaging devices, [Link here], (12/3/15)
[iii] Patowary, Kaushik , What is The Shutter Life Expectancy of a DSLR Camera?, December 26, 2012, [Link here] (12/7/15)
[iv] Thein, Ming, Some Thoughts on Digital Camera Lifespan, 7/6/2012, [Link here] (12/3/15)
[v] B&H Photo Website, [Link here] (12/4/15)
[vi] Kenmore Camera Photo Print Service, [Link here] (12/7/15)
[vii] Shanebrook, Robert, “Making KODAK Film: The Illustrated Story of State-of-the-Art Photographic Film Manufacturing,” Shanebrook, 2010, Rochester, NY, USA, p. 9.
[viii] Ibid.
[ix] Ibid, p. 18.
[x] Ibid, p. 9.
[xi] Densities of Miscellaneous Solids, The Engineering Toolbox, [Link here], (12/7/15).
[xii] Shanebrook, p. 35.
[xiii] Lower greenhouse gas emissions, Solvay, “CO2 footprint* for the production of 1 kilogram of plastics (in kg eq. CO2),” [Link here] (12/7/15).
[xiv] Desjardins, Raymond at el. , “Carbon Footprint of Beef Cattle,” Sustainability 2012, 4(12), 3279-3301, [Link here], (12/7/15), p. 3294.
[xv] Ashby, “Materials and the Environment: Eco-informed Material Choice,” Butterworth-Heinemann, 2013, p. 127, [Link here], (12/7/15).
[xvi] Personal email E.J. Peiker to John Kossik, 12/3/15.
[xvii] Dessouky, Yasser at el., “Computing the Carbon Footprint Supply Chain for the Semiconductor Industry: A Learning Tool. “ Industrial & Systems Engineering Charles W. Davidson College of Engineering, San Jose State University, Proceedings of the 41st International Conference on Computers & Industrial Engineering, p.4.
[xviii] Ibid.
[xix] Personal email E.J. Peiker to John Kossik, 12/3/15.
[xx] Ibid.
[xxi] Shanebrook, p. 35.
[xxii] Shanebrook, p. 76.

I would like to extend my thanks to John for taking up his personal time to produce this thought provoking and meticulously researched piece of work. there is certainly food for thought there and it definitely opens up many avenues of discussion.
We would love to hear your thoughts and your feedback in the comments below.

Thanks
JCH