Ilse Jurriën : March 25th 2008 - 08:35 CET
Geotate interview : Digital photography continues to develop. If not on the field of image quality or intelligent software then on extra features. Shooting itself is no longer satisfying. We want to do more and more with the captured pictures. Such as linking them to the place where they were taken. Geotagging is becoming the magical word. We witnessed a few of these applications and possibilities at the PMA show. One of the companies is Geotate, spin-out of NXP. They have invented a method of geotagging accurately without hardly any power consumption. We interviewed Hans Streng, CEO of Geotate about geotagging, GPS and naturally about their application. A pleasant talk about a near future subject. |
Question : Many people are not familiar with your company. Can you give us a brief explanation about Geotate?
Answer : Geotate is a start-up, a spin-out of NXP. Geotate is a company dealing with software GPS and services of software GPS. NXP is mainly dealing with the hardware while our core business is making the software. NXP acknowledges the possibilities for Geotate but cannot be into everything as semi conductor manufacturer so they spin out. NXP keeps a minor share. The major part of the shares is owned by an investors group. The group focuses on every possibility coming up in the geo domain. The geoweb, developing very fast at the moment, requires for all the assets to come with markers of where they are and where they have been. A geocode gets tagged to the features of every object. There are many ways of doing so and we chose to do it with GPS. We have a different approach than the hardware solution and with it we are capable of geocoding objects, such as pictures which would not be able with the existing GPS. One of the most logical applications is the one inside cameras. And because it is so obvious we have decided to only aim at this application in 2008. It is so very likely.
Question : Geotagging is becoming more popular. Why is it becoming so popular? What's the fun about geotagging?
Answer : It is a totally different dimension than just watching pictures. We live in a mobile community. People are interested in trying to find out what's happening globally and where it is happening. You can put everything on the globe with Google Earth, a virtual globe, but nonetheless a real globe. It enables you to search for place-bound, but not time-bound, items. For pictures it is very simple. If you want to know which pictures you took in New York, at this moment there is no possibility to search for New York, (as long as you haven't added anything in IPTC, BdM) and you just have to recall when it was. Time-bound you know it very well, but place-bound you cannot search for it yet. However; you can search for place-bound pictures of your buddies, an application Flickr is offering. Now if you want to know where your whole community went or where they took pictures, you can search for the place and see where they are at that moment or where they have been. Tagging pictures is coming on strongly and not only pictures, also tagging other segments is coming up. John Hankey, Google's director, informed us about people in Silicon Valley tagging all the traffic lights. John added that you can discuss the usefulness of it, but then everything could be questioned. If you like it you just do it. And it also offers many possibilities on the field of advertising for example. We can tag anything, even videos.
Question : Geotagging is fun, as you just said yourself. But how does geotagging work for the professional photographer?
Answer : That is an interesting question. You can think in two directions. First it is the fun as you correctly stated. And secondly it has to do with everything professional, b2b, security etc. For example; a photographer wants to know the assets available on the location he is going to work. So if you are an advertiser for the Olympic Games and you have to invent a campaign, you can search on Google Earth for assets, movies and pictures that are useful for the campaign. As a professional photographer you can offer shooting material this way. We have spoken to companies specialized in licences and if someone wants to purchase a picture he can purchase the rights and the professional place-bound picture through these companies. A photographer can go to Beijing even at this moment and take a number of nice pictures to put online in a tagged portfolio. If you are searching for a special place you will find these pictures as thumbnails. And if you want to purchase them, you can order them through an agency.
Question : And the archives of the photographer… are they still useful?
Answer : Yes of course, you can also just search the archives. That is also the fun factor people have at home. They can say: 'show me all the pictures of…' and they can select from the selected section. You can even put your whole archive on the globe, just as Flickr wishes it to be for the consumer.
Question : I use a separate GPS receiver for geotagging and I link the pictures and the GPS data through a program. It would be excellent if I saw in the picture which street I was in because at the moment I still have to look it up. Will this belong to the possibilities?
Answer : Yes, we've got one running here now. We have made a few pictures in Las Vegas. We store the raw data and the various reverse geocoding facilities are available for free on the web. Inserting your coordinates will give you the information as to where exactly the picture was taken. Mind you, you have to log into the APIs. The business target of these companies is to increase traffic to their website of which they benefit too.
Question : And these things can also be written to the IPTC?
Answer : Yes, that is no problem at all. lat/lon alone is of no use to you. But there are thousands of people who have said, 'IF' it is available, we can do these kinds of things with it. And we have enabled it. We have the disposal of many APIs so processing is easy. Everyone can do that really. But not everybody can offer lat/Ion.
Question : What are the limits for current geotag systems?
Answer : The usual limits we all have. GPS is for outdoors not indoors. We have overcome the limits of hardware GPS inside cameras. But it is a fact that GPS works outdoors and not indoors. You can invent some tricks to estimate where about you are and we do as a matter of fact, but in essence it doesn't work inside. We are working on it to see if we can overcome this with Wi-Fi or something like that. It might be possible but we want to stick to our principles and offer a perfect lat/lon. If it is not accurate we don’t want to offer it. Other than that there are no limits as far as we are concerned. Well, you have to 'look up'. That is a hardware requirement. But we are working hard to find a solution for that too. Also with a compass so you know which way you were watching.
Question : What is the disadvantage of your system since every system has disadvantages?
Answer : Yes, our system too has disadvantages. The big minus is the fact that it is not real-time. So you don’t know where you are at a specific moment. We cannot navigate with a camera. I wouldn't know why you'd want that at all, but there are people out there thinking it is a nice feature. All we do is check afterwards where something or someone was. It’s not like tracking people. We could incorporate it, but that would be doing the same thing as all the other hardware manufacturers do. We'd rather do things differently. I don't see any more disadvantages other than the usual ones every GPS suffers from.
Question : Can you explain in a few words as to how Geotate's system works?
Answer : We have a small radio receiver that takes a snapshot of the satellite signals it has traced and stores it in a memory. Storing takes around 200ms, very fast, we open an antenna quickly, pick up the signal and put it in a memory. We don't do any calculations with it; we don't do anything with it yet. On a later stage, varying from one minute to ten years or whenever, the snapshot can be put on the PC and the PC will obtain the additionally required information from our server. Next a calculation will be made as to where the snapshot was taken. So 200ms raw satellite data suffice to calculate where the snapshot was taken, independent of the place and the circumstances. That is in short what we do and it is completely secured with patents. It is very simple and hardly consumes any energy. The energy we need to make a snapshot is not even 0.1% and even better it is almost reaching 0.01% of what a camera uses to take a picture. It works immediately; you don’t have to wait until you receive a signal. Everything happens in those 200ms.
Question : How do you make it happen so fast? An average GPS receiver makes you wait thirty seconds or even longer before you can work with it.
Answer : It is the way the GPS works. Satellites are sending a signal continuously; a standard signal. If you receive that you should be able to calculate where you find yourself on the globe. But you have to know the exact position of the satellites at that moment. And that is the additional information satellites continuously send out with intervals of approximately 30 seconds to one minute. Also correction data comes with it. And this amount of data makes it take so long with an average GPS receiver. Your GPS can only begin calculating when all this data is received. We pick up the extra data and store it on our server so we only take a snapshot from what the satellite is sending at that moment. The moment we calculate where the snapshot was taken, we will obtain the extra data that applies to that moment from the server and we combine it with the snapshot. We do exactly the same as the average GPS receiver in its 30 seconds, only afterwards.
Question : Well, that requires storing a huge amount of data for years and years. I suppose it takes up a lot of room.
Answer : You have to store it that is correct. And if you want to store the additional worldwide data for ten years it costs 1TB. Nowadays that is nothing. It is not a lot at all, but satellites do have to send it. And they do it in a low byte rate. That makes you having to wait so long for it. There are methods of obtaining the extra data through another source. Via a phone for example you can obtain the data from a server that has already received it. It is called; assisted GPS. You have to use the 'send and receive' function which consumes energy. And a camera or MP3 player cannot offer that. We make GPS functionality possible for this class.
Question : When I walk with my Garmin 60Cx, with excellent antenna, in a city, I seem to be jumping up and down in a very high speed while I am just standing still. How do you solve that problem?
Answer : It is something standard in a GPS analysis. These reflection errors are found in every GPS. Everybody is encountering these so-called urban canyons, the reflection from buildings etcetera. Normally GPS analysis is carried out in two phases. The first phase you just calculate straightforward where you are. And what comes out of that is full of these crazy jumps. Besides that the second phase is linked to the analysis; the so-called positioning engine and that is software related only. The software will check how reasonable it is that you jumped from one side to the other. It also takes your speed into consideration. When you are driving using navigation software and you carry on a different road for example, it takes a while for your screen to show the different road. And that is simply due to the positioning engine busy analysing if it is at all possible. This calculation manner is a unique feature from every brand or supplier. We do the calculations on a separate computer and have therefore many more sources at our disposal to analyze loads. We can estimate a more correct calculation of where it really was, than anyone else.
Question : What is the hardest part of integrating GPS in a camera?
Answer : That is a beautiful question because we had to ask this ourselves of course. The hardest part of getting the GPS inside the camera is the interference with the camera itself. The camera is some kind of a jamming transmitter. The image sensor receiving and processing the signals is almost like a microwave oven when sending out, especially at high resolutions. It all happens with high frequencies being sent out. The LCD behind it is also an enormous source of emission and also the processor. It means a huge amount of noise coming from the camera. The GPS signals are sent down from a satellite and weaken on their way down. The noise of the camera has the effect of being at a cocktail party. The antenna of the GPS is listening to someone whispering in its ear while many are screaming around it. What we do, because we only need 200ms, is making a deal with the camera manufacturer that is going to incorporate our GPS in his camera, that the camera stops its noise for 200ms while taking a picture. For example, the time right before the picture appears on the monitor. At that exact moment we will take the snapshot. And because it is such a short time span, the user will not notice a thing. A camera manufacturer has everything worked out as to how the camera should work and then we appear and tell them they have to do it differently. And they have to be willing to put time and effort into it. Those are the issues we encounter.
Question : And a metal housing is not a problem?
Answer : No, that is not a problem. And when it is shielded we can always make a hole in it or make it thinner. There is more than one solution for it. Even a piece of copper would do it. This is how we work with the manufacturer. Find the best spot on the camera. For example on an LCD monitor. You make the snapshot after one or two seconds because the user would be watching the LCD which automatically gets the antenna upright.
Question : When your system is incorporated in the camera, are the snapshots written in the EXIF or do I have to work with the data in a separate file?
Answer : No, what we do is incorporate our engine in the applications that the manufacturer delivers with the camera. The entire calculation to lat/lon is carried out at the moment you upload your pictures to the computer through that program.
Question : Do you have to be online constantly when you copy your pictures from the card?
Answer : You'd have to be online at least once. You don't have to go online every time. The application should be able to have access to the extra data from the satellites to be able to calculate. The application will ask through the timestamp in the picture for the extra data related to that time. You'd have to get online at some point.
Question : So for the calculations you have to rely on the software of the manufacturer? I for instance never use that software. How can I make it happen then? Is there a plug-in available for Adobe Camera Raw?
Answer : Right at this moment we are talking with some people handling professional applications as you describe. The calculation programs can also be available as stand-alone features. It must be a professional one. And that leaves you with Photoshop or ArcSoft-like companies that should be able to offer a full photo handling solution. And this could well be a built-in feature. That is what we are doing right now, you touched the sore! But you are right; you don’t want to depend on one or another feeble program. And for the large bulk it doesn't make too much of a difference.
Question : In my RAW file I would find only snapshots and if I were to open Photoshop if your engine had been integrated there, then the lat/lon and eventual location etcetera comes into the EXIF?
Answer : Yes, that's correct. The program will calculate at that moment, without you noticing it. For a professional it should be a dedicated program. We also have to monitor how you use it. That is why we thought of the camera. We just had to integrate it in there. People don't have to change anything for it; just the camera is somewhat different. And this is how we keep on inventing and developing. If you look at the technique only you won't accomplish anything. You have to start with looking at the user.
Question : Looking at the future of GPS... Will it remain GPS or will the European project Galileo become the standard?
Answer : There are even more systems, from the Russians for example. Everybody is developing them, especially to become independent from the Americans. GPS could be very more precise but the Americans have incorporated two modes. One for military use, extremely precise, and one for remaining use. The European system is compatible with GPS, a well thought-out move. So, most signals coming from Galileo are one-to-one similar with those of GPS. Our radio function to receive a snapshot is the same as for GPS, for Galileo or any other system with a very small tweak. If you have a hardware system it will work for GPS but not necessarily for Galileo. You have to build in something that recognizes whether it is from the one or from the other. And when the one is working the other is not and vice versa. It means spending money on something you won't use. If you put it in the software as we do, we take a snapshot and it is the same, after that we can calculate where it was. No matter if it is GPS or Galileo which is a lot more efficient.
Question : When can we expect Geotate products to be put on the market?
Answer : We are working on it now and you can see a few cameras here already. The fastest we can deliver is the upcoming summer. We are negotiating with a number of camera manufacturers at the moment to put our products on the market full speed. Most of them are aiming at autumn so before the Christmas season. And the Japanese are somewhat slower and want to go for next year's spring. Besides that we are offering accessories like the module that can be placed on a hot shoe. It is easier and quicker to put accessories on the market, so I think we will be on it during Q2 of this year. These accessories will be in the shops for a few ten dollars.
Question : Like Jobo?
Answer : We have worked with them but they have chosen something else, more hardware targeted. I cannot comment on it as you will understand.
Question : With such an accessory snapshots are made at the exact moment of taking a picture. But I do delete pictures in the process. So how do I know which picture I took where?
Answer : Through timestamp. That is the only possibility. Otherwise you'd need a marker in your picture and in your logger and that would have to be the time. If the camera is set completely incorrect, you do have a problem. That is a standard problem of an accessory which can be solved through the software. We are also working out with camera manufacturers if we can make a marker it not being the time. You have to make a link somehow.