I can barely pay my bills! so I figured out I could be better off making panoramic imaging and the Gigapan EPIC Pro is awesome. Basically, this is what Google Art project does to produce the Gigapixel images of artworks. Gigapan EPIC Pro it’s still affordable, a bit expensive – maybe – if we consider it is basically a motorized panoramic head. Also there is a big difference in the price of the Epic Pro and the other two lighter models, BUT you really need the EPIC Pro. I use it both with a digital and an InGaAs camera. Among stitching software I like PTGUI. You can get high-resolution images in few minutes and without the hassle of forget to shot a picture in your panorama!
Gigapan Epic Pro is an must have tool for high resolution art documentation
Fantastic work Antonio! Great video.
Funny, I had the same idea and just bought a Gigapan Epic too, for the same purpose.
By the way, small spelling mistake in first U V title ( you wrote Ultravilet)
I’m using Gigapan, and it’s really so handy. I’m looking to have them develop an adapter so I can work with the InGaAs camera with it
I just updated a spelling corrected version of the video. Thanks for notice it!
Hello,
The distance from the sensor to the ends or in the center of the picture is different.
How did you solve this problem, because I do not think that the device can successfully focus each photo?
You close the lens focal length?
I hope you understand my message.
Edouard from France.
Hi Edouard, thanks for your comment. Indeed, look at this, it should help you. Depth of Field Calculator http://www.dofmaster.com/dofjs.html
Thank you, but that is the fow diaph is a solution that degrades the quality of the photo. The only solution is to use a telephoto over 400mm, but the “Gigapan” hardly bear this weight. It also asks to have a lot of recoil.
I thought maybe you had a better technique.
In any case, thank you for sharing.
Edouard
Salve Atonino, Thanks for sharing all this. I have used stitching for taking highresolution photos of artwork for some time now. I have one problem which you talk about passingly above this entry. I use a rather longer lens then you do (CANON 70-200mm) to get more photos of the art work and thus getting a larger resolution. This gives me a shorter field of depth, and means that I have smaller leave way in the focus range going over the painted suface. Using the depth of field range doesn’t cut it for me (it is around 2 cm with the amount of light and the focus and shutterspeed I use). Certain areas of the painting photographed will have to be in less focused than others when using the “sweet spot” focus on my lens (usually at somewhere around 160mm lenght and at the 8f which for this lens is the sharpest). Using the automation of the Gigapan Epic pro is very tempting, but sometimes, when the picture surface is very smooth (with little change in colour or techture), the autofocus of the camera will not work. Do you have any ideas as what to do, in order to remidy this? I’m currently using a manual Rollei camera mount for paning photography, which works fine, but it is slower to use than the Gigapan Epic obviously.
You don’t talk about the use of polarized filters in front of the lens as well as the lights…..is there any explanation for this?
All the best,
Mikael Melbye (painter)
Ciao Mikael,
grazie for your comment! I’m actually starting to get many of these comments from other professionals working in high resolution imaging of art. Which is great. And therefore I’m working on a extended article on this subject, trying to give more insight into this topic than the bare post. My first suggestions for you would be: 1. Use a prime lens, they are faster than zoom lens, I use an old Nikon Nikkor 200 mm F4 AI. 2. In order to have an idea what the theoretical depth of field far and close limits are I found this website great: http://www.dofmaster.com/dofjs.html You should check your painting’s center and far borders stay into the near and far focus limits. 3. I don’t recommend auto-focus. I always do manual focus with LIVE VIEW on my laptop monitor and then, before doing the shooting, I check that the photos of the center and of the borders are on focus (I use a PC tethered camera). 4. I’m preparing a post regarding polarized light for art photography, so keep be posted..But, briefly, the problem with polarized light is that you need a lot of lights! Polarizing light means to allow just a fraction of the original light to reach your camera. In high resolution photography (we use slow lens), this can be prohibitive! ciao!
Ciao ancora, I have some comments that might challenge you, but don’t be cross with me….I would like to hear your reply very much….we can all get wiser if we work together…ahahahah
so: I use the canon 70-200 mm lens for two reasons…. first it is one of the sharpest lenses on the marked (including the primary lenses). It does have a sister version at an opening os 2:8, but the lens is much costlier, and simply not as good. Secondly I prefer to use the zoom lens because many of the paintings I photograph vary enormously in size, so in order to be able to get them all within the same bracket of resolution, the zoom lens is more practical. If not, I would have to carry too many lenses around which is costly and heavy. I used dofmaster, but even when the photo is within the parameter of the depth of field range, there is alway a difference from the middle to the outer perimeter, and even at relatively long distances like for instance 5 to 6 meters. And this gives an inconsistent result as far as I’m concerned. The tethered approach is good, but you can’t use that with the Gigapan system (at least if you want to use the automation of it…..it is too quick to see if the image is actually sharp). And then I might as well use my Rollei manual mount.
About the polariser filters……it is true that it does take 1,5 stop off your aperture when you have filters in front of your lamps as well as your lens. But I looked in to how much potential damage flash lights (without pilot lights on) would have on the paint layers of even very fragile paintings. I discussed it with several conservation experts here in Denmark, and they all agreed that the very short bursts of the flash light (equal to normal sunlight) at 1/1000 of a second would be less harmful then a quarter of an hour with tungsten lights on. Also the tungsten lamps alter their temperature over time, which means that you have to change the bulbs simultaneously and quite often to maintain a steady working environment light wise. I use two flash lights which are known to be 5600 kelvin at pretty much their entire lifetime, so I don’t have to worry about this at all. Also the tungsten lamps do tend to get quite warm, and this is never good for old paintings, who do not like any change of temperature at all I’m sure you agree. To get back to the effect of the polarisers…….this takes away any trace of false glare. You see the pure pigment on the painting. The added colour intensity that the double filters make, will be taken care of by the colorchecker passport if you make the profile with the filters in place. If you want to show the impasto work on a painting, you simply move one of the two lamps further out to one side until the effect of shadow in the brushstrokes is obtained.
ps: dove sei in Italia. Io abito a Cortona la meta del anno…forse si possiamo incontrarci ?
Ti saluto, Mikael
Hi Mikael,
Your comments are more than welcomed! If I had already all the answers I’ll not be blogging! I want this blog to be a tool to “share and learn.” This is the spirit! We all want to be better professionals. Now I’m back in Sicily, last year was in New York, so probably I got into this blogging since, even if living in Sicily is awesome for many aspects, I really miss access to professionals in the field, such as you can have in a big city as NY. But blogging has shown to be exstremely effective. got in contact with many nice people.
Regarding your technical comments. Sure I understand your points. I think, in general, one should develop its equipment and methods for the specific professional tasks. for example, I use the halogen lamps since I do infrared photography, so I need a strong source of infrared, together with visible.
Fammi sapere se passi dalla Sicilia! ciao. Antonino
Voluntieri e altretanto se venga a Cortona. Sto li tutta l’estate. Volevo dirti che tutte queste tencniche non sono affatto conosciuti a Danimarca. Sono l’unico qui che le fanno. Quindi mi sembra come fosse in Sicilia…..nessun a debattare….:o) Ma, in somma, hai raggione, solo con dialogo si puo migliorarsi. Io non ho molto uso per le luce infra rosso quindi non e un dovere per me a avere la luce constante. Spero che ci incontriamo….tanti saluti della ghiaccia di Danimarca,
Mikael
Grazie Mikael. Hai Linkedin?
Si certo!
Hi, due to the fact this tripod head works by changing the perspective I suppose the stich software makes a big correction of the distort result, isn’t? Will love to clarify this basic doubt because I cannot find an exaustive video or text explaining how it actually happends.
I usually work in a manual mode, not changing the tripod head angle but their height (of course I don’t produce gigapixel images like this) and then the stich software (or even a manual stich) will have a lot less of correction, due to the fact all images where taken in a ‘frontal’ mode.
Will really appreciate if you can confirm it.
Kind regards,
Hi,
thank you for your comment. You can find very good material on the subject of panoramic photography and in particular correction of image distortion, perspective and rectilinear projection on PTGUI.com. Taking picture frontally, moving your camera up and down, sure doesn’t introduce distortion, so I guess, you can stick with this method since it is solid and sound and it works for your needs.
Thanks for your answer, I just wanted to confirm that distortion where produced and corrected later in ptgui (I love it). This automated procedure stills of course very interesting since it, under well supervision, will procude correct results saving a lot of very valuable work time.
Great blog!
that’s the spirit! saving valuable working time!
Antonino, ciao. Intanto grazie per aver accettato il mio invito su Linkedin: mi fa piacere essere in contatto e confrontarmi, per imparare, con professionisti del settore.
Leggevo con interesse la discussione nata da questo tuo post (mi scuso per non intervenire anch’io in inglese, ma vado più veloce così… 😀 ) e mi ha incuriosito quanto detto su PTGui e le distorsioni.
Premesso che è un sw che uso con piacere ormai da anni, ho sempre considerato un “difetto”, nel montaggio di una griglia di foto fatte su una superficie piana, il fatto che egli tendesse sempre e comunque a correggere le deformazioni prospettiche, anche dove non ve ne sono.
Mi spiego meglio: nella riproduzione di certi dipinti o disegni – soprattutto se non incorniciati – è impossibile posizionare e fissare l’opera in verticale.
Dovendo lasciare il soggetto in orizzontale, su un piano idoneo, il punto di ripresa obbligato è quello zenitale e questo esclude la possibilità di fotografarlo con l’ausilio di una testa motorizzata come la Gigapan (la quale necessita di essere posta in orizzontale su di un treppiede). A questo punto, a meno di non possedere costose attrezzature, non resta che muovere il corpo macchina (o il soggetto) lungo una griglia predefinita… e qui entra in ballo il problema dello stitching.
PTGui, di fronte ad un set di foto riprese in questo modo entra in crisi, pur con la nuova funzione “align to grid” che dovrebbe, in qualche modo, semplificare le cose…
In qualche modo ho rimediato “barando” sulla lunghezza focale: in fase di importazione non faccio leggere gli exif dai files ed indico io focale e fov, ma non sempre questo è sufficiente.
Autopano, della Kolor, con l’ultima versione sembra aver previsto ed affrontato il problema ma non ho ancora acquistato l’upgrade.
Tu come ti regoli, in casi come questi?
Usi sempre PTGui? e se sì, con quali parametri?
A presto e complimenti per il tuo sito!
Alfredo
Ciao Alfredo, thanks for your comment, you are welcome! That’s the spirit of this blog. I must say, I never used gigapan for documentation of an object laid horizontally. I do document easel paintings so even if they don’t stand on the easel 90° vertically, I tilt the camera of the same amount of the painting.
In your scenario, documenting horizontally laid art, if you move the camera on a grid, theoretically you could just blend the images, since you don’t have any deformation to take into account, and so you don’t actually need to use a panoramic stitching software. It’s like the software used for IR Reflectography, NIP, if I recall the name correctly. You need not a panoramic software but a mosaicing one.
Hello,
What’s the best way photographing large paintings 240cm x 60cm hight ?
Move the camera in front of the painting,step by step parallel to the painting with a full sensor camera 24X36mm and a tilt shift lens 50mm on tripod with sliding bar?
Hi Thierry! Well, I think, in general, there is not a better way to make things. it’s also up to what equipment you actually have and what you are used to do. You have essentially two choices here, and both are valid. Panoramic photography, which is what I ‘ll do since I’m familiar with the method and I have the equipment or the “sliding bar” method you suggest, provided you have this precision sliding bars.
The video was first rate, and obviously displays wonderful work. It’s certainly interesting how many like minds one can unearth if one just expends a little effort. I have several high-res Nikon 35mm DSLR bodies…the infamous D800 being among them. I also have a Hasselblad 503CW with a Phase One back. So what else but that of late I’ve been looking into the possibility of acquiring a scanning back for my 4×5 Plaubel?
As anyone who has looked into the potential of this option, it’s not cheap, and using it can be extremely time consuming. So what should happen but that a light bulb went off in my brain. Why not just use a Gigapan? Admittedly it’s not very fast. but certainly faster than a scanning back. Of course it does a wonderful job with extremely high resolution, and it’s dirt cheap when compared to a scanning back. Right?
Well, surprise, surprise. Here we have several people who’ve beaten me to the punch forthwith. And all I can say is: wonderful. I love the idea of using a Gigapan to document fine artworks…or any other highly detailed smaller image, other than the usual grand landscape. For the longest time, (at least since the inception of digital imaging) much of this kind of museum/appraisal photography has by and large been done with medium format digital setups that cost the earth.
Using a Gigapan to generate a more conventionally rectilinear image…as opposed to a stock and standard panorama…turns this particularly egregious equation right upside down. As you’ve noted there is software available that deals with the issue of parallax displacement and other forms of distortion, etc.
Everyone seems to have their own theory as to lenses and so on. For me, the idea of documenting art works…or the like…seems to point in the direction of a medium high res DSLR body…maybe between 16 and 20 mpx max. and possibly a 45mm tilt/shift lens, using no tilt or shift to speak of, but having the capability available if needed. (long telephoto lenses are only necessary for landscapes…whether “grand” or not so…primarily to take full advantage of the zoom aspect in the finished product when displayed on the web) At any rate 45mm (or so) would lend itself to obviating depth of field issues, as well as finessing the distortion found in genuine wide angle glass. It would as well render finished images containing so many pixels that it would take a very high magnification loupe to detect any irregularities in detail. The very idea of employing a D800 simply boggles the mind. There would be so many pixels to manage, that the next expense hurdle would be in the area of the computer and it’s ability to store such monstrous files. Practically speaking, to a certain extent the storage problem is already rearing it’s head. It’s always true that when unanticipated consequences come up, photography is never cheap. But there are solutions like the Gigapan that can make it significantly cheaper.
Good heavens, to think that all it took was a trip to Mars, to develop this nearly other-worldly technology for use by the common, and even the not-so-common photographer. Wonders quite truly never cease. Bravo!
It’s clear, to me at least, that in the future the the Gigapan principle is going to find all kinds of uses that none of us can even dream of right now. It’s obvious that both the hardware, and particularly the stitching software are both going to become much more sophisticated and effective as time passes.