Can’t seem to get something to synth right? Curious about what lens is best? Ask your fellow Photosynthers here.
What's the optimum size for an individual for, say, a 30 image synth? My camera takes up to 10mp pix but that would be impractical to upload and I would have thought unecessarily large.
I take it that your Internet Service Provider limits how much data you transfer per month? That is the only reason why I could imagine it being impractical for you to upload the full 10 megapixel versions of the photos.
To be completely honest, the zooming and image streaming technology which Photosynth employs (Seadragon) is specifically made to quickly and easily view images of any size, so the bigger the better, really. When each image first loads on anyone's screen, the Photosynth viewer will only download as much of the image as is necessary to display at that smaller size. If someone wants to see more of a specific face or some other detail, only once they zoom into that part of the photo will more of it be downloaded to them as they look more closely at it, so never fear about the photos being too large.
It won't be too many more years and we'll be looking back at even our 10 megapixel photos and thinking that they are low resolution, so if your internet plan will easily cover the upload, I'd say just go ahead and upload the full quality images.
In terms of what image quality, today's synther program (which calculates and uploads your synths) needs in order to successfully match your photos together, all of the actual image matching is done with approximately 2 megapixel versions of the images, so any resolution greater than that won't get you more photos matched together, however uploading the full quality *will* give people the most imagery to zoom into and will also stand the test of time better.
It's also worth saying that the point clouds that we see in our synths today are only the first step in what sort of models can be built with our photos and the higher quality our photos are, the higher quality future reconstructions will be.
In case I've botched the explanation, here are a few other times that the question has been asked:
2008 August 24 (Photosynth was 4 days old)
2008 August 26 (Photosynth was 6 days old)
2009 March 30
And here are some videos that talk about Seadragon:
2005 February 12: Pre-Microsoft SeaDragon demonstration:
2007 March 10: TED 2007:
2006 July 28: Channel 9's 'Going Deep' show with Charles Torre
2007 January 07: Channel 10's Jesse Lewin learns about Seadragon
2007 October 03: PBS's Ziya Tong interviews Blaise for Wired Science
2008 January 25: TechFest IIT Mumbai 2008
More available at:
http://docs.com/XFO or http://photosynth.ning.com/video/video/listTagged?tag=Seadragon
Thanks for the tips and links. I hadn't realised that the image was created on my computer and then uploaded, which is obvious now I think about it.
Looking at my efforts so far, I've been underestimating the amount of overlaps and number of images. I'll have another go.
Use an image resizing application to resize single image to the best size ratio, like 1:2 or activate the batch image resizing function to handle multiple images' size at a time, which comes from an image processing library, including image flipping, scaling and rotating.