Topic: Distorted 3D reconstruction! (possibly due to error: "Could not find ccd width"?)

Report Abuse Report Abuse
swami_worldtraveler (Over 1 year ago)
Here's the synth in question: "Credo - Songwriter's Stage..."

Just did first synth w/ new Canon G12. Noticed log reported "Could not find ccd width for 'Canon PowerShot G12'" and "Computed ccd width of 7.416800mm." It's unfortunate the G12 data is not in the database, but I understand not all models can be covered (hopefully, this one will be added, considering it's a fairly popular model). Of more immediate concern is the very low computed value, which is only ~0.3"! The G12 has a slightly over-sized CCD (for a point-n-shoot) of 1/1.7".

The synth is of a long aspect, rectangular building, yet the point cloud (see Overhead view) does not represent an orthogonal structure!:(


P.S. I just noticed, and just threw this up will it's fresh in my mind. I'll go check the forum and GetSatisfaction for similar reports (and possible solutions)...

Nathanael (Over 1 year ago)
Swami, long time no see!

The link to your synth got the full stop added to it by the forum, breaking it.
Here it is again:

As far as help on Get Satisfaction, the closest thing I can find is Amir's suggestion here:

Also related, but less helpfulin terms of solutions:

That's all I've got for you for now.
swami_worldtraveler (Over 1 year ago)
Hey Nat!

Nice to see you're still around:) I've done a little synthing, and a lot of showing (doing presentation this month for travel-centric photog group I started in CouchSurfhing), but haven't had a need for help. But, here I am now!:)

Tnx. Yes, I found those two links, as well. I don't see "FocalLengthIn35mmFilm" in my EXIF. Just Googled. It's a TIF tag. And it's different than the "Focal length" datum. Even if I had an equivalent, I'd have to read the FL, do the math, and write the result. Of course, this might be cam depend. Perhaps somebody has already done this(?)

Of course, I'm not even sure if the 'ccd width' is the problem. I will have to do some testing. There was a somewhat dense 3D 'crowd,' which might have affected the synther. I just don't know. I've bee looking at house examples, etc. Anyhow, the "good" news is that the distortion is not apparent when navigating the 3D photo space.

I let you know if I come up with anything...

Tnx, again.

swami_worldtraveler (Over 1 year ago)
P.S. Unfortunately, those posts were over 3 yrs ago. And nothing was resolved. But maybe it's good news that not much can be found on the problem (of distorted t. cloud, not ccd width issue). Of course, the problem could just be rarely noticed +/or under-reported. Currently, I file it under "mystery"...

P.P.S. One thing's for sure: the ccd width is considerably bigger than just 0.3".
Nathanael (Over 1 year ago)
Hopefully you'll hear from someone on the Photosynth team, but you might want to also repost your question at and