In-Camera Geoencoding and the Nikon D4: Case Study In Product-Development Costs, Ignorance, and Naïveté
NOTE: Images with an icon next to them have been artificially shrunk to better fit your screen; click the icon to restore them, in place, to their regular size.

Now that Nikon has announced its next flagship pro SLR, the Nikon D4, with much flowery prose but few hard details, discussion and debate and speculation and flames and praise have filled camera circles.

This post is long, but here's the two-sentence summary for the “tl;dr” crowd:

Nikon didn't include built-in geoencoding in their new camera.
I wish they did, but perhaps this will help you understand why they didn't...

As with most any technology release (hardware or software), folks tend to frame their personal whims and desires as “absolutely required!”, while features that they don't personally care about become unneeded fluff. Photographers who don't care about video, for example, lament the assumed cost of all the new video features.

This is human nature, and in this respect I am human (why on earth doesn't Lightroom have geoencoding built in!?), but my moral outrage at the omission of personally-desired features is tempered by many years of experience with product design, and a long-learned understanding that in practice things are much more complex than they might seem from the outside.

A feature not included in the D4 is built-in geoencoding, an omission that some people find unfathomable, as evidenced by a thread at DPReview started with this message:

D4 no internal GPS? Seriously?

I mean... No words...

Those who don't care for it can just turn it off.

Those who do... Gosh. I really hate saying this, but what is nikon thinking?

This rant is clearly emotional, by someone who doesn't realize that it's emotional, nor that others may not share their particular emotions on this issue, so it's easy to dismiss. However, it started a long thread — one of many, I'm sure — where people debated the usefulness of having a GPS receiver in camera, and the business decision of having omitted it.

Of course, the responses included equally emotional rants on the other side of the aisle, such as this gem from Curtis Smith:

4 real?

Are you that stupid you need the camera to tell you where you are? Its a freaking camera! Not a swiss army knife!!! Who gives a rats a$$ about GPS? I so sick of you old farts complaining about this ,that and you don't even have the camera. Yes a was complaint a while back for nikon to hurry up and give us the D4 and they have so guys give it a rest its not going to have all the tricks you wanted as its just a camera and by the looks of it a Dam good camera!

Personally, I've been geoencoding my photos for the better part of a decade (the ones I've posted, mostly in Kyoto, are mapped here) and would love to have it built into the camera.... but only if done well. I understand enough to know that it's not necessarily easy to do well, so I followed up the initial message with one of my own about some of the likely costs that came to mind — costs such as the physical space the GPS unit would add to an already bulky camera body, and how it would lower the value/cost benefit ratio (that is, the marktability) for those not interested in geoencoding their photos.

Each specific “cost” has its own level of impact, but the manufacture must weigh them as a whole against their benefits, and at this point in Nikon's product roadmap (at this point in the economy, at this point in their market research, perhaps in light of Mother Nature's harshness in 2011), built-in geoencoding didn't make the cut. Maybe it will when they reevaluated everything for the next camera. Or maybe not.

Anyway, as is common on faceless online discussions, some folks mix in a healthy dose of ignorance and antagonism with their emotion, such as this bit of arm-chair logic from Robin Casady replying to a “there are costs” post similar to my own:

Re:Seriously? But you don't really want to face the facts...
Cost (1): Adding it to the camera (system) entails engineering costs.
          :
          :

The software aspect must already be done because you can plug-in a Nikon GP-1 and geotag each image.

So, fact #1 = bogus.
          :
          :

Here, Robin is referring to a small add-on product that Nikon sells that includes a GPS antenna; when you plugin it in to the camera and it has any kind of satellite fix, photos are geoencoded automatically. This has been available for years, and if I were to use it with my current camera, it would replace what I actually do (which is carry a separate GPS receiver while out with the camera, then sync its tracklog to the photos in Lightroom, with my geoencoding-support plugin).

I don't choose to use Nikon's add-on because I feel it's an ugly wart on the camera, I worry about its speed and accuracy, and it lacks the ability to display its status beyond more detail than a blinking light. And even if I did use it, I'd still need a way to geoencode after the fact, for locations where I didn't have it with me, to geoencode the photos I get from others on the same outing with me, for locations it couldn't get a good signal, and for times where a depleting battery prompted me to turn it off.

Anyway, at first blush Robin's logic may seem reasonable... Nikon already has the hardware as a separate product, and the cameras have had for years the firmware needed to geoencode the photos, so the simple step to include the hardware inside the camera takes almost no work, so citing “engineering costs” seems unreasonable, Robin concludes.

But that's an exceptionally naïve, ignorant point of view, even when limiting the scope of “engineering costs” to the “software aspect” that Robin brings up. (Update: In light up the updates noted below, “exceptionally naïve, ignorant point of view” is a perhaps harsher characterization than Robin's statement merits. See more below...)

For example...

  • If you move it from an external dongle to an internal component, you've lost all UI (“user interface”), so now you've got to come up with a way to interact with the user. You have to come up with a way for the user to turn the GPS feature on and off, and you have to come up with a way to indicate to the user whether the unit is on or off.

  • Let's look at just that latter point for a moment, the oh-so-simple “indicate to the user whether the unit is on or off”. That might naturally go into the viewfinder display, but since that very-limited real estate has been staked out by other UI needs for years, you're forced to change or remove things that until now were considered critical. Maybe just make everything else slightly smaller to squeeze in one more indicator, but geez, that was done the last several iterations and now things are starting to get way too small....

    So that's not straightforward, and the same angst applies to the LCD displays.

    Update: In a touche!-worthy reply, Robin points out that there's already a GPS-related indicator in both the viewfinder and LCD-body displays, which flashes to indicate that it's on but has no position fix, and is solid if it has any kind of fix. This clearly makes this bullet point of mine completely moot and I have learned something, but I don't think removing one or two points removes the idea that there's a cost/benefit balance that most of us don't fully understand.

  • Then you're faced with the fact that “on” and “off” are not really sufficient... you need to be able to communicate to the user whether the unit is actually getting a signal, and if so, how accurate. If I recall correctly, the hardware dongle uses a bi-color LED to indicate the difference between “on with good signal” or “on but no good signal”, and this is really very minimal because “good signal” means different things to different people (100m accuracy vs. 3m accuracy, for example). Also, “Not good signal” may really mean “warming up” or “can't see satellites”. So if you're going to bring it in house, you'll naturally want to tackle that issue to provide the user with a more-usable level of information, but maybe all of that doesn't need to go into the viewfinder, so now you've bifurcated the information (some in the viewfinder, and some somewhere else... likely in the menu system), so finding the balance there is another issue.

  • Then you've got to consider the situation of when the battery is getting low... it makes no sense to run the battery into the ground just so the GPS receiver can keep its fix while the camera is idle.... what good is it to have had a good fix for the last hour if you now don't have enough battery to take a shot? So you've got to come up with some measures to handle this situation (perhaps GPS auto shutoff when battery is below X% full), and this means that you have to have a whole new UI to communicate to the user when this has happened, what the current status is, and perhaps to allow the user to make adjustments to the heuristics for their specific needs.

    Update: In the same reply noted above, Robin cites that there is already an option just as I postulated, that turns off the GPS receiver when the exposure meter turns off, some user-selected time (default: six seconds) after the shutter button is released. I offered the idea merely as an example of something that might be done, but I'm quite dismayed to find it's actually there in that form because one would generally want the GPS receiver to be active a lot longer than the exposure meter; the exposure meter comes back on instantly when you press the shutter button half way, but Nikon's GPS receiver takes a minimum of five seconds to get a fix when already hot, but is rated at 45 seconds when warm (unused for more than 15 minutes, I think). I'd like to the GPS unit remain on much longer than the exposure meter, and would like the ability to have it wake up every 14 minutes or so to keep itself “hot”...

  • But we've forgotten the more common situation of power management... should the unit be on all the time, or should it auto-shutoff in the same way the image review shuts off, coming back up when the shutter button is half depressed? (I'm half depressed just thinking about this.) Both make a lot of sense, each applied to its own situation, but no one solution makes sense for everyone, so again you need to add UI to allow the user to indicate their wishes, and more UI to communicate to the user this additional status of the settings chosen and the current state within the context of those settings.

Notice how the “simple software aspect” immediately turned into a hardware aspect: both the viewfinder display and the LCD display require physical hardware manufacturing changes to update. There are other hardware aspects one could turn to (such as the LED seen in the external GPS unit), or one could keep it all software by burying everything into the menu system, but these are substantially less desirable.

Update: So, with my first viewfinder/LCD point having been completely wrong, the previous paragraph is pretty clearly unreasonable. You do still need to handle on/off for the GPS unit, but in my experience as a geoencoder I think it'd be sufficient to have it in the menu system.

All features have costs. The ones above are just what came to my imagination at first blush, but because I have no experience with embedding GPS receivers inside cameras, I'm sure there's a whole litany of additional problems I've neglected to consider.

Update: And it's because of this ignorance that we all have (at least those of us who are not camera-design engineers) that I characterized Robin's curt conclusion of “Bogus!” as I did. In light of my own ignorance on the viewfinder/LCD issue when I wrote it, my use of “exceptionally” seems overdramatic. Sorry about that, Robin.

Then there are the followup costs, such as the need to document all this in the user's manual, and translate that to all the languages that Nikon supports. (Of course, they'll have had to translate all the related UI in the camera menus.) Then there's the recurring need to provide support: more features mean more support calls, etc.

None of this is unsurmountable, and it's all certainly trivial compared to all the video features that did get added, so it wouldn't have surprised me in the least to have seen Nikon include this in the D4. But to do that they would have had to shift resources from something else, and if they had done that, we'd see these same online flamefests about whatever that “something else” was that got the short stick.

We'd also see additional flamefests about whatever GPS support did get built in and how lame it is...

What, GPS only? No GLONASS? What is Nikon thinking?!

This applies to pretty much every new hardware and software product you'll come across (I see it every time Adobe releases a new version of Lightroom, or Apple releases a new phone), so the next time something near and dear to your heart is released, realize that like everyone else you're talking to, you don't know it all, and take care to frame your wishes and opinions as just that, wishes and opinions.

(And as a matter of courtesy, don't point out my own advice to me when I don't follow it. :-))

By the way, several years ago I read an excellent writeup from a software developer about the extensive ripple of high-impact ramifications that he came across while trying to implement a “trivial” feature request from his users. I think it was about some kind of email-notification aspect of an online forum of some sort. My web-search kung-fu is weak on this one.... does anyone have a link for this article?


All 11 comments so far, oldest first...

You’re right, of course. But as your little comment in grey near the end quite clearly demonstrates, this is advice very few people are going to follow. We all like to make our opinions and disappointments known sometimes.

Regarding GPS in the D4… Even if I could afford to get a D4, and even if it had GPS, I’m not at all sure I would use it. You mentioned most of the reasons already: battery problems, likely accuracy problems, and of course the need to correct things later anyway. When it comes down to it, having a tracklog is just much more useful.

I’m interested to hear that you GPS encode photos you get from people. I often find myself using my tracklog later to do just that when sharing photos with friends from the same outing.

To summarise: your geoencoding tools make any in-camera GPS pretty much obsolete, as long as you are willing to go to the effort of making tracklogs. And frankly, since you can do so these days even with your iPhone, I think it’s going to be a while before our workflows change away from this paradigm.

Long may your plug-ins prosper! And since it’s easiest to say in Japanese, Lightroom 4もよろしくね!

We all like to make our opinions and disappointments know” — Yes, which is fine if we frame them as opinions and disappointments, not pronouncements from an all-knowing omniscient judge. I rarely geoencode others’ photos in bulk, but if I’ve been in close proximity with someone on an outing, I’ll use my tracklog as a basis to encode theirs. (I would never use a tracklog from an iPhone except as a last resort because it mixes in some location methods that can be wildly wrong, to the tune of, literally, hundreds and hundreds of miles. I’ve been told my location is Tokyo when I know I’m in Kyoto, for example.) —Jeffrey

— comment by Thorf on January 8th, 2012 at 6:10pm JST (5 years, 11 months ago) comment permalink

Some more rants about a GPS receiver.

Nikon sells an external box for it, so they know how many peoples are really interested in this feature, not just the people who likes to complain. Perhaps there is only a few of them.

Where do you put the antenna for a GPS receiver in an environmentally sealed magnesium alloy body ?

How many DSLR models does have a embedded GPS receiver ?

The cost of software development decrease with the number of product sold. So adding software function can be very cheap. The cost of an hardware function don’t decrease.

That’s a good point about them knowing the sales numbers for the external unit, but at the same time, I think it’s fair to guess that the vast majority of people who do geoencode their Nikon-camera-produced photos do so without using Nikon’s external unit, so using that unit’s sales would severely underestimate the demand (which even when not underestimated may not be as great as the vocal passion of the minority that does it). As for the magnesium body, I’ve heard that argument before but it seems to me that the miniscule antenna proper could easily go under a bump in the rubber outside the magnesium shell. My Garmin GPS+GLOSNASS receiver is waterproof, so it’s certainly not impossible, but I’d guess that issue is trivial. —Jeffrey

— comment by Fabrice Bacchella on January 8th, 2012 at 8:31pm JST (5 years, 11 months ago) comment permalink

Re: iPhone tracklogs: don’t knock it until you’ve tried it, Jeffrey.

I use the MotionX GPS apps, and I’ve never had a problem yet. The accuracy is comparable to the Garmin GPS tracklogs I use most of the time, and more than good enough for geoencoding photos. It’s wonderful to have GPS in your pocket wherever you go without having to carry yet another gadget.

I have tried it… years ago, before “Find your Friends”, I ran a script on my jailbroken iPhone that updated a web site with (the iPhone’s best guess to) my current location so that Fumie could see where I was, but found that the error jumped around between 50m and 500km, often on a minute-by-minute basis. Maybe it’s related to the sketchy nature of the “enhanced” part of “enhanced location services” in Japan? I sure wish I could have made it rely on only the GPS receiver. Even with my current iPhone 4s, I’ve found that the “mail my location” feature of WhatsApp is sometimes 100km off. Maybe dedicated tracklog apps can ensure that only actual GPS data is used? In any case, my dedicated unit just sits in a pocket of my camera bag so it’s not a burden. Next time I go out, I’ll keep a parallel tracklog with the iPhone, and compare them. That’ll be fun… —Jeffrey

— comment by Thorf on January 8th, 2012 at 10:49pm JST (5 years, 11 months ago) comment permalink

You still read DPR forum ? *gasp*

Sure, sometimes. There’s a lot of crap, but it’s fun to browse and there are still plenty of sane, reasonable folks there. —Jeffrey

— comment by Chan on January 8th, 2012 at 10:52pm JST (5 years, 11 months ago) comment permalink

While these are all valid points, I do think that we’re forgetting the legal aspect.

For starters, somebody may have a patent on the GPS-in-camera feature and who knows what their licensing terms are (if any)? Going a bit further down the rabbit hole: IIRC, Russia has explicit rules that any device with GPS must also support GLONASS, so it isn’t quite as simple as just any cheap GPS receiver… Which leads to that all-important question: what are the places where having GPS support in the camera makes it illegal or forbidden by the location the photographer is trying to work?

So, while I don’t want to minimize Nikon’s greed as a factor (they are a consumer electronics company, after all), I can’t help think that there are very good reasons why neither Nikon nor Canon are willing to embed a GPS receiver into their DSLRs. That sort of blanket avoidance reeks of legal reasons rather than technical or marketing ones: otherwise they would never pass up the one-upmanship opportunity that having the feature brings (even if the implementation is awful and tacky — there isn’t even a crappy, barely working, incredibly cheap to implement, version of this anywhere in the range).

Good points. That also brings to mind what I’ve heard about video… some limits on how long a video clip can be are related not only to sensor overheating, I’ve heard, but also to import tariffs that tax “video cameras” at much higher rates than “still cameras”, with “video camera” being defined for these purposes as having the ability to record clips of at least X minutes…. so to avoid those import tariffs, the artificial limit is added. I don’t know whether it’s true, but it wouldn’t surprise me. —Jeffrey

— comment by Carl Irving on January 9th, 2012 at 12:56am JST (5 years, 11 months ago) comment permalink

Yes I did make the following commit about your rant and rave on the D4!
Ill say it again!
Are you that stupid you need the camera to tell you where you are? Its a freaking camera! Not a swiss army knife!!! Who gives a rats a$$ about GPS? I so sick of you old farts complaining about this ,that and you don’t even have the camera. Yes a was complaint a while back for nikon to hurry up and give us the D4 and they have so guys give it a rest its not going to have all the tricks you wanted as its just a camera and by the looks of it a Dam good camera!

Thanks for sharing your thoughts, Curtis, but I should point out that your comment that I quoted was not a reply to me, but to the original “what is Nikon thinking?” poster. To date, I’ve not ranted nor raved one way or the other about the D4 (though I did rant about Nikon’s smarmy marketing the other day). And by the way, in case you didn’t realize it, the manner in which you chose to express yourself does not reflect well on the writer’s education or social graces. The brunt of what you intended to convey, that you don’t value a geoencoding feature on a device meant to capture photos, is certainly reasonable, though, and it’s apparent that Nikon agrees with you. —Jeffrey

— comment by Curtis Smith on January 9th, 2012 at 1:52am JST (5 years, 11 months ago) comment permalink

Which GPS are you carrying around to geoencode your photos?

I use a GPS+GLONASS unit, a Garmin eTrex 20, the successor to the previous two Garmin units I’ve used. Garmin is a mix of inspired design and utterly moronic design, glued together with an almost total lack of documentation. I get the strong feeling that they don’t actually use their own products, because the shortcomings become apparent very quickly when you actually try to use them. It’s very frustrating, but I know of nothing better for my needs. —Jeffrey

— comment by T.J. Powell on January 9th, 2012 at 2:44am JST (5 years, 11 months ago) comment permalink

“I use a GPS+GLONASS unit, a Garmin eTrex 20, the successor to the previous two Garmin units I’ve used. ”

Not desiring to take this post in a completely different direction… but having gotten the eTrex Legend HCx on your recommendation only 6 months ago (yep, knew it was “old” but I forgot my datalogger on the other side of the world and was desperate), what am I now missing with the eTrex 20? I’m curious how much more accurate or fast to latch onto a signal this +GLONASS unit is over a GPS only unit. I’d love to see a review similar to the one you linked to in the comment above.

I don’t use a bazillion features… mostly just keep tracklogs, but as far as I can tell, the Legend HCx is superior in most every way except the +GLONASS bit. The Garmin engineers really outdid themselves to make things klugy. (For example, tracklog files are now named in the form “Day DD-MMM-YY HH.MM.SS.gpx” rather than the commonsensical “YYYYMMDD.gpx” that sorts proper in a file listing.) It’s head-smacking “didn’t you even try to use this yourself?” incarnate. But it’s got better satellite reception. I’ll take it, my old one, and an iPhone out on a walk sometime once I can shake this cold, and post a comparison of the tracklogs. —Jeffrey

— comment by JasonP on January 9th, 2012 at 9:53pm JST (5 years, 11 months ago) comment permalink

I think people like Curtis are missing the point. I don’t geocode because I need the camera to tell me where I am, I do it because I enjoy documenting where I am as well as what I seeing. Since I was young I always enjoyed looking at the map while on road trips and figuring out where we were and what we were close to. That’s just me. Everyone all different.

Least we get too caught up in our own section of the internet where camera’s are the most important thing on the planet I feel compelled to remind everyone that the general population couldn’t give a rat’s ass about shutter speeds, apertures and ISOs…

— comment by Jesse G on January 11th, 2012 at 9:43pm JST (5 years, 11 months ago) comment permalink

Perhaps http://inessential.com/2009/07/30/anatomy_of_a_feature is the writeup of which you were thinking?

Yes, exactly, thank you! —Jeffrey

— comment by Ned on April 19th, 2013 at 6:21am JST (4 years, 8 months ago) comment permalink

I’d settle for bluetooth, I already have an external bluetooth adapter (wart) that pairs with my bluetooth GPS but it is only good for GPS. If Nikon were to add bluetooth to the camera, it could be used for much more. Could you imagine bluetooth speedlights? Bluetooth remote releases – you could even use your cell phone as one.

— comment by Bryan G. Jensen on July 16th, 2013 at 1:43am JST (4 years, 5 months ago) comment permalink
Leave a comment...


All comments are invisible to others until Jeffrey approves them.

Please mention what part of the world you're writing from, if you don't mind. It's always interesting to see where people are visiting from.


You can use basic HTML; be sure to close tags properly.

Subscribe without commenting