Hello.
I find myself spending way too much time taking photos.
I became really quick, but anyways, it takes ages because RC needs so many images.
I saw some drawings in posts here that shows how one should take photos, but it seems that this doesn’t really work for me:
-
I either spend too much time on too many photos (giving me good results, but I can’t spend so much time on taking photos).
-
If I take the minimum amount of photos shown in one of the drawings (sorry, I don’t find it anymore), the results aren’t sufficient. For example, areas in my render may be blurry because I didn’t get in close enough.
Does anybody know what I mean?
To sum it up:
If I take really much time and take many photos, it’s very fine.
If I take the minimum of photos, it’s not so great, I have blurry areas in the render.
If I would actually understand how RC, I could use it better, I think.
For example, I would like to ask if I need to provide enough images to give some “volume” information about an object, and RC will estimate the volume of the object and thus create the object from the volume?
If yes, in which way does it use color information? I have made the experience that for control points, color is very important.
If my assumption about the volume is correct, would it be sufficient to take some photos from far away so that RC has enough “volume information” to guess the shape, and then I make close up photos without any angles for the textures?
Thanks!
Hello Alice,
Well, to be honest with you the best way you gonna achieve the “right amount” is by having more and more experience and going through technical stuff in photography field.
I dont want to be rude here, but RC is not (in my point of view and if you want measurable and reliable data) a friendly software you gonna learn in one or 2 days, because there is so much knowdledge you need before. I’m using RC since end of 2015, and won’t say i’m mastering it.
Also its hard to say what is “sufficient”, highly relative on people, and what they need to do with the scan.
Did you had a look on the help section of RC, “How to take photo” part “Level of detail and precision”? great tool for distance placement.
How are you taking your picture? (iso, aperture, focal length?) so i can maybe improve your setup. depending shoot contxt, objects it can be very different, hard to tell without a clue ^^
What Aure said, including the bit about learning photogrammetry basics and 2015 !!! 
The help section you mention is only targeted at pixels, which is better than nothing I admit! :-) But blurriness can also be caused by imperfect geometry.
alice, what is “too much time” or “too blurry” for you? Those are all variables that no one but you can judge for yourself. What is way too long for you, somebody else might be thrilled about how little time it takes. Can you give us some numbers? For example if you spend 2 days for the little wall that you uploaded in one of your first posts, I would agree with you that it’s too much. Do you want to finish shooting a house in 20 minutes? I would say that is unrealistic, even for the oudside (although for g**gle earth it might work ok). I think the images you mention are just schematics to clarify the principle, not 1:1 maps for placing your tripod.
Also people work differently. Some use tripods and spend some time for each shot to be optimal, others just shoot from the hip and don’t care so much about slight blurriness here and there but are therefore much quicker in the field. Then in the office, the first one might be quicker - or not. It depends on so many things.
As with all other photogrammetry software there is a simple rule when it comes to the results.
The quality of your input images affects the quality of the result you get.
Basically, if you rush things and put garbage in you will get garbage out.
For my typical turntable workflow I spend anywhere between 20 minutes and an hour to take all the pictures I need.
The amount of pictures I take strongly depends on the complexity of the object. For simple objects (like a garden gnome) fewer pictures will suffice, for complex objects with intricate details and difficult surfaces only visible to the camera within a small angle I take a lot more.
Anywhere from about 30-40 to over 200.
But I always take the time to do things proper. Because if I don’t, I get to do it all over again.
If you gave examples of your work you could get more help about it , some tips and solutions.
I come from a photography background, so I find this whole photogrammetry thing very interesting. After learning about it recently, I decided to build a computer so I could explore this medium… but I digress.
What are you shooting, in what kind of light, and with what kind of camera?
I’ve been shooting with a 5dmk4 and a stabilized 24mm lens with no issues. But I am also shooting in the shade and creating a “full picture” of my subjects with my photographs. It would help to know more about your issues.
Can you expand on the penult sentence?
‘in the shade’ - can you restate what that means e.g. transferred to room-interior work, and exactly why important?
‘full picture’ - does that mean just covering everything, or some stand-back shots showing the whole picture - or what? and why?
In the shade - as in, overcast day, no direct sunlight but as much diffuse light as you can get to eliminate shadows. Applies to indoors too where you can have absolute control over the light.
full picture - as in, make your object of interest cover as much of the pictures you take as you can. Try for 80% or more.
If you can only cover part of the object because of that, you need adequate overlap when you photograph the other parts so RC can piece the object together.
That is how I understand and apply it.
“as much diffuse light as you can get to eliminate shadows. Applies to indoors too where you can have absolute control over the light” - as I thought - but then I read this interesting alternative view
Aleksandr Gorky in https://support.capturingreality.com/hc/en-us/community/posts/360002794972-Flash-for-building-interiors-for-RC says:
“If you use mobile light you will get texture mixed from different pictures and shadowed parts on one picture will mixed with illuminated part from another picture. Finally you will get uniformly illuminated texture.”
“You must capture in RAW and then correct shadows in Lightroom\Capture One”
Any opinions?