Photosynth gets Geekbriefed

Cali and Neal from visited us in October, and they've just released their video "brief" of Photosynth based on that visit.  Here it is:

 And here's the synth that was made on-camera during that shoot. My color balance was a little on the sickly yellow side for this one. Sorry Cali!


Comments (1)

  1. Nate Lawrence says:

    Well that was pleasant. I need to look up their previous episode on Photosynth, apparently.

    It’s also good to hear someone besides Blaise emphasise the difference in results between the determined (and reasonably informed) photosynther vs. typical mass photography.

    I have to admit that my earliest synths all used portrait shots in my panoramas. I believe I had 50 percent overlap more often than not but I wasn’t exactly overjoyed with the results of that method.

    I’ve wondered before whether the 1.5 or 2.0 megapixel versions of our photos that are used to harvest image features are always defined by a maximum pixel size in the largest dimension (width or height). This is to ask, "EXIF data notwithstanding, would square photos that were centred on the same subject and the same number of pixels tall as landscape photos actually generate more image features?".

    I will say one thing. I feel like somehow the description of Photosynth needs to be distilled to be much more clear while still remaining compelling and concise. This is certainly difficult to do, but I can’t help feeling that it would result in better reactions from people.

    Reading the comments over at raises my blood pressure a little over people’s dismissal. Time and again you see the word ‘underwhelmed’. Well, what is the reason for that?

    Somehow in describing what Photosynth is, expectations are raised above what Photosynth is delivering today. I’m certain we all feel that the future holds nearly unlimited promise, but I wonder whether it isn’t better to show people before you tell. I actually think that the order of presentation in Blaise’s TED Talk had a lot to do with its success. People saw it in action for themselves before any description could conjure up a vision in their heads that transcends Photosynth’s current position.

    In any case, this is a great little way to wrap up the year on the heels of the second version of the Silverlight 3 Viewer, Geo-Alignment, Bing Maps Beta, Overhead View, etc. It’s been a great year of progress, guys. Thanks so much for all your hard work.

    That being said, I can hardly wait to see:

    > Linked Synths! (this is going to buy you so much credibility as it’s really the big piece that’s been MIA since the TED Talk and the Live Labs promo video days)

    > Dense Reconstruction!

    > aligning each separate pointcloud in a fractured synth (and even orphaned photos) using the geo-alignment tool

    > MSR ICE integration!

    > hardware acceleration for proper 3D and stable pointclouds

    > view the map tiles as a ground plane under the pointcloud as in the geo-alignment tool (also allowing us to play the slideshow in overhead view).

    > the ability to suggest geo-alignment for other people’s synths (there are so many people who are just never going to do it on their own!)

    > the ability to send other Photosynth users a direct message, rather than having unrelated conversations in comment threads on their synths.

    > notifications to replies in the forums

    > Bing Maps Silverlight leave beta (hopefully ditching the black void around the satellite tiles when you dive into a synth)

    > the ability to log into and comment on the official Photosynth blog with my official Photosynth account. 😉

    and the list goes on and on.

    I have prioritised the above for the most part as far as what I would love to see, for whatever it’s worth. Obviously some of it extends past the Photosynth team (such as Silverlight’s capabilities.)

    I’m definitely looking forward to seeing what your team has been working on in the coming year. Thanks, again.

Skip to main content