Yea, I can't imagine where, outside of aerial photogrammetry/LiDAR or climbing, where elevation is relevant, much less precise and accurate elevation. But different people do different things!

http://en.wikipedia.org/wiki/Photogrammetry

Thanks for taking the time to try and explain it to me Drab. It seems to me that for the vast majority of users, a simple tag showing longditude and latitude is all they're looking for? The type of photography where altitude is also usefull to log is quite specialised I would imagine? Being someone who has never had the need to log any location info, your and the other members comments on this thread have opened my eyes to the problems associated with accurately recording that third axis. I'll never hold theTomTom in my car in high esteem again! ;-)

SkintBrit said:
Regarding this GPS conversation, I think I should have payed more attention in school, as you are making my head spin. Haven't got ANY idea what you're talking about, but sounds interesting :-)

In a nutshell WAAS is a free service covering North America which gives you a correction for the errors inherent in the GPS system. It gives you accuracy within 10 feet normally. More and more cheap consumer receivers support it. While some third-party sites say the Nikon GPS-1 supports the WAAS satellites, Nikon's website doesn't and their accuracy claims are more inline with an autonomous position. Since Nikon likely isn't designing their own GPS chips this feature will likely come to them for free if/when they release a GPS-2 - most of the newer chipsets on the market today support it natively.

http://en.wikipedia.org/wiki/Wide_Area_Augmentation_System

Differential GPS is the style of solution correction which is at the heart of WAAS. You can use WAAS, one of a number of subscription services offering coverage outside north America, or higher quality within North America. You can also run your own correction station. Cost is from $100/yr to $20,000 in equipment. This is truly a specialist solution. I'll wear a dress and call myself Nancy if any of the camera manufactures ever offer a (non cellular and non WAAS) differential service.

http://en.wikipedia.org/wiki/Differential_GPS

Kinematic GPS is another style of solution correction. It must have a base station within 10km of the rover (ok, some exceptions apply), and can give you sub-cm accuracy. While it is completely overkill if the desire is sub meter vertical it is the only way to be sure - as kinematic (unlike differential) involves a robust initialization which is constantly verified.

http://en.wikipedia.org/wiki/Real_Time_Kinematic (not the only way to do kinematic, but the only way discussed in depth on wikipedia.).

AT LEAST $10,000 in equipment cost.

This too is a very specialized setup and will likely remain so for a long time to come. Farmers are just now starting to use it but their needs for precision location is far from us photographers.
EDIT:
Or you can post-process GPS data recorded with a dual-frequency receiver ($1000 minimum) and get great solutions. There are two main ways to do this. One is static GPS which (as the name implies) is only effective for a non-moving receiver but can give you insanely tight results, the others are post process versions of the differential and kinematic styles mentioned above, with the quality being as mentioned above as well.

SkintBrit said:
Regarding this GPS conversation, I think I should have payed more attention in school, as you are making my head spin. Haven't got ANY idea what you're talking about, but sounds interesting :-) (Picture nuclear physicist teaching a kindergarten class, and you'll get where I'm at) :-)

You're not missing anything. What they keep going in circles about is trying to get the elevation of where they are shooting - And I have zero clue why or why it even remotely matters. People just arguing for a dream of having a $10,000 GPS system in their camera.

NikoDoby said:
Yes you did. Several recent threads and post seem to have disappeared for some reason.

Yes that happened to a post I made on the pocket wizard thread yesterday. I felt sure it had been sent successfully, but then it disappeared. I'm pleased to hear there was actually a problem! Regarding this GPS conversation, I think I should have payed more attention in school, as you are making my head spin. Haven't got ANY idea what you're talking about, but sounds interesting :-) (Picture nuclear physicist teaching a kindergarten class, and you'll get where I'm at) :-)

Ade Barkah said:
Hmm, I thought I had replied to this thread, I guess it got deleted.

Yes you did. Several recent threads and post seem to have disappeared for some reason. Keep the disscussion on-topic (Nikon GPS) or I will close this thread again.

Hmm, I thought I had replied to this thread, I guess it got deleted.

In any case it's not generally possible to get sub-meter accuracy with a calibrated barometer (or a consumer GPS augmented with one). Even aircraft altimeters built and maintained per FAR standards have a tolerance of +/- 20ft up to 1000ft, and +/- 80ft at 10,000ft. See the tables in FAR part 43 Appendix E.

Not to mention that in real life atmosphere does not follow ISA standard pressure, temperature & lapse rates. Hence we have radio altimeters.

This thread went plain funny there for a while. Barometric pressures? And here I thought that was just for weathermen.

WAAS is the huge addition for accuracy. I saw my delivery trucks go from 200 yards or more accuracy (were not always able to connect to enough sats) to 1-2 meters easily - constantly. We even lost a handheld device on a highway that connected to the truck for the gps and it pinged each truck that drove by and sent the GPS data. We found it less than 2 feet away in the ditch from where the GPS data said it was.

I haven't seen any (doesn't mean they aren't out there) GPS devices for cameras that utilize WAAS yet. I was never lead to believe that it was part of legacy devices. I have an old logger that I matched with files but rarely used it.

If it was built in, that would be neat but not necessarily a big selling point for me.

The kicker to me is that attaching more devices to my camera is just a pain. I would rather see bluetooth added to the camera with the ability to connect to add-ons where I could clip the device to a bag or something. Same goes for RF wireless triggers and anything else. I am not a fan of connecting items with wires or plugging them in on my camera. Especially when everyone wants them to attach to my hot-shoe.

dormant said:
I need it when I'm shooting from a helicopter, so elevation can vary. We have used mega-expensive continuous-GPS kit to get high-accuracy locations, but I'm interested in how I can get moderate accuracy (~1m) without spending the big bucks, and without post-processing of the GPS data.

You can not get 1M vert w/o differential GPS (WAAS counts as differential - as Testing123 can tell you (he works in the WAAS office down the hall from me)) You'll be lucky to get 3M vertical even with differential. Kinematic GPS is the only sure way to reliably get sub meter vertical - or a calibrated barometer. ;)

EDIT: I've been told 3M vert is a bad day. You should be able to get 1.5M vert with one sigma confidence if there isn't a solar storm. Many handhelds should be using WAAS corrections now. Remember, a precision approach at an airport is 4M vert.

@dormant - have You tried connecting garmin gps with a dedicated cable?

Testing123 said:
3 - If you aren't in an airplane consider simply pinning your positions to the ground. For almost every lat/long there is only one ground elevation. ;)

I need it when I'm shooting from a helicopter, so elevation can vary. We have used mega-expensive continuous-GPS kit to get high-accuracy locations, but I'm interested in how I can get moderate accuracy (~1m) without spending the big bucks, and without post-processing of the GPS data.

@Testing123:

Drab concluded that a barometric altimeter can be kept "decently calibrated" for "weeks at a time" [without external input]. This is a misleading conclusion which does not follow his/her earlier points. The problem is whenever the barometer is not at rest, there is no way to isolate pressure movement due to changes in altitude vs. changes in weather.

As a counterpoint, I cited a fatal incident where a professional crew perished using an altimeter setting from just two days prior.

Besides, 'dormant' was worried about ~50 meter GPS accuracy. It seems to me that even drab's calculated 631ft average error for an uncalibrated barometer is much worse, and only supports my earlier point.

Also, after my last post, I went out to shoot, and forgot to attached the GP-1 again. :)

Cute Drab, but unfortunately that kind of analysis is rather misleading (and potentially dangerous).

Suffice it to say that at a given location, atmospheric pressure fluctuates constantly. In Canada (where I live) a transport helicopter crashed because they used an altimeter setting from "just" two days before. Their altitude was off by about 830 feet due to this error, and unfortunately the entire crew of four paid for this mistake with their lives. And no, there was no hurricane nearby.

Back on topic... I'm constantly on the road and last spring I was introduced to the GP-1 by photographer Louis Au (who works for the Masterfile stock agency). I really like it so far, but I wish it was integrated into the camera body since I sometimes "forget" to attach it when on location.

Ade Barkah said:
Not really. Unless you already know your altitude (e.g., from a topo map), barometric altimeters are next to useless without an accurate "altimeter setting" to translate atmospheric pressure by reference to sea level pressure (QNH) or some known elevation (QFE). This setting changes all the time -- literally with the weather -- and isn't generally available outside of airports or weather reporting stations.

Piqued my interest, so I ran some numbers: (Sorry for the imperial units)

1000' of elevation change = ~1.1 inches of Mercury.

Cat 5 Hurricane = ~-4 inches of Mercury. F5 tornado = ~-5 inches,

Delta between typical high pressure and low pressure events = ~1.5 inches = 1363 feet of potential error swing if you aren't under a cyclone = ~681 feet of error on average for an uncalibrated barometer.
;)

Though the USA has only like 100,000,000 benchmarks for calibration. Easy to do. My barometer has two settings "I'm moving so assume all pressure changes are elevation" and "I'm still assume all pressure changes are weather". Helps keep it decently calibrated for weeks at a time.

dormant said:
I've also seen a GPS unit that uses a barometric altimeter for altitude, which is much more accurate than the GPS altitude, which can be up to 50m in error.

Not really. Unless you already know your altitude (e.g., from a topo map), barometric altimeters are next to useless without an accurate "altimeter setting" to translate atmospheric pressure by reference to sea level pressure (QNH) or some known elevation (QFE). This setting changes all the time -- literally with the weather -- and isn't generally available outside of airports or weather reporting stations.

In most GPS units having a barometric altimeter, the primary altitude information still comes from the satellites; The barometric altimeter is only used to help "smooth out" the GPS readings, like a filter. But if the GPS altitude is off by say +50 meters then the barometric altimeter wont help anything.

Anaxagoras said:
Yes, I would agree. Cost, weight, and bulk are all important.

The cheapest GPS I know of is 25 (GBP - about 40 USD) which isn't much in the context of a dSLR. Weight is 20 grams (about 0.7 ounces!).

Bulk is more of an issue with a standalone unit being about half the size of a camera battery. But presumably the clever engineers at Nikon could integrate the actual chipset (no casing needed) without adding too much bulk.

By the way, and no offence whatsoever intended, if you know exactly where you took an image, and by implication will still remember in ten, fifteen, twenty years time, shouldn't you be leading a more exciting life, my friend?

You're right, EXIF doesn't support compass bearings.

School is my life right now. :D

Like I said, "Ehhhhhh". :)

You're neither right nor wrong. There are no compass fields, but none seem to mind. ;)

Testing123 said:

And while this might seem to be a dirty hack, it is the method used by every camera manufacturer I've run across so far.

Mea culpa.

I did hesitate before I wrote that, because I had heard that there was a way of working around the problem, although I had seen no evidence of how to do so. I was also led to believe that it would not be a realistic option for most photographers - in terms of quick, easy, workflow.

Perhaps I should have said "EXIF doesn't support compass bearings CONVENIENTLY".

My apologies, everyone, if I have mislead you.

Anaxagoras said:
You're right, EXIF doesn't support compass bearings.

Ehhhhhh.

A - EXIF supports custom fields.
B - EXIF supports multiple assigned fields suitable for this, as there is no reason not to place a compass bearing in one of the GPS bearing fields. If you want to distinguish between derived (GPS) bearings and measured (compass) bearings use one the bearing method ASCII tags.

See:
GPSImgDirectionRef
GPSImgDirection

http://www.exif.org/specifications.html

And while this might seem to be a dirty hack, it is the method used by every camera manufacturer I've run across so far.

De facto / de jure. :)

NSXType-R said:
I kind of like the idea of GPS, but I also don't want it if it would add significant cost to a body, or bulk.

Besides, I never really go anywhere on vacation, so for the most part I know exactly where I took it.

Yes, I would agree. Cost, weight, and bulk are all important.

The cheapest GPS I know of is 25 (GBP - about 40 USD) which isn't much in the context of a dSLR. Weight is 20 grams (about 0.7 ounces!).

Bulk is more of an issue with a standalone unit being about half the size of a camera battery. But presumably the clever engineers at Nikon could integrate the actual chipset (no casing needed) without adding too much bulk.

By the way, and no offence whatsoever intended, if you know exactly where you took an image, and by implication will still remember in ten, fifteen, twenty years time, shouldn't you be leading a more exciting life, my friend?

dormant said:
...compass, but I'm not sure if the direction goes into the EXIF data or not.

You're right, EXIF doesn't support compass bearings.

dormant said: I've also seen a GPS unit that uses a barometric altimeter for altitude, which is much more accurate than the GPS altitude, which can be up to 50m in error.

This is true, but there are a few workarounds you can try depending on your workflow.
1 - if you are recording your NEMA stream at a high enough frequency (and post-tagging images) look into using GPSBabel to toss positions with a high VDOP. This will clean up your altitude noise quite a bit.
2 - if your software doesn't already do so look into using a GPS log converter which supports the use of geoid models. GPS measures height above the ellipsoid, which is not the same as elevation. Geoid models are the translation layer between ellipsoidal height and elevation. Depending on where you are this can be your largest source of elevation error.
3 - If you aren't in an airplane consider simply pinning your positions to the ground. For almost every lat/long there is only one ground elevation. ;)

Hi all,

It seems really cool, and from a news background position and general work position, really, really cool, and useful. Cost and weight are both on the table of course.

My best,

Mike

GPS is essential to some of my photography. But I also need a record of the direction the camera was pointing in when I took the photograph, which needs an in-camera compass. The new AW100 has a compass, but I'm not sure if the direction goes into the EXIF data or not.

The GP-1 is not the most useful GPS unit on the market. I use a now-unobtainable GPS unit which had a power-saving mode not available on the GP-1. I've also seen a GPS unit that uses a barometric altimeter for altitude, which is much more accurate than the GPS altitude, which can be up to 50m in error.

I kind of like the idea of GPS, but I also don't want it if it would add significant cost to a body, or bulk.

Besides, I never really go anywhere on vacation, so for the most part I know exactly where I took it.

NikoDoby said:
I don't really get the need for GPS on cameras. I understand what it's for but for me I'm in no hurry to geotag every picture's location.

....why can't I have built-in GPS in my D300s/D400! Now that would be cool! :)

You surprise me, Niko.

At my age, most of my photography has been pre-digital. And one of the major frustrations with film images is not knowing WHEN they were taken. Now I know precisely when every digital photo was taken. This, to me, is a huge leap forward.

A similar frustration is not knowing WHERE they were taken. Geotagging is now so easy, no more difficult than time-stamping, that I do it routinely. I don't NEED to know WHERE every photo was taken, but then again I don't NEED to know WHEN every photo was taken! The information is there if I do want it.

But I agree with you wholeheartedly that GPS should be built-in. I cannot believe that we are now well into the 21st century and it still isn't even an option on modern cameras. For example, Nikon offer a D7000 body; Nikon offer a D7000-with-kit-lens; why can't Nikon offer a D7000-GPS with built-in GPS?

(My apologies to any photography Luddites who are apoplectic at such a staggeringly radical idea!)

For the record, I now have two GPS solutions:

1) A "Foolography Unleashed" bluetooth unit linked to a "Holux M120-E" GPS logger
This is by far and away the neatest, tidiest, most discrete solution you will find. But remembering to keep the Holux unit charged (and switch it on!) is a bit of a hassle. Oh, and it's just a bit on the pricey side, too.

2) A di-GPS unit. Simple, basic, 100% reliable. I clip mine onto the camera strap where it is not too cumbersome. It has On/Off/Auto settings to control camera battery drain. Overall battery life? I've no idea. My battery (D90) seems to last forever (with or without GPS); it is so good, I often wonder why I always carry a spare.