The Purpose of installing SDM StereoData?

[univurshul]

Hi,

Just curious, what is the purpose of SDM StereoData software? What does it do? I've been briefly looking around online, and there are no detailed explanations of what it does for the cameras. I'm thinking it has something to do with getting the cameras to fire at the same time, maybe?...which leads me to my next question: is there a way to wire the switch to get the left camera to fire just before the right? --This would print a earlier file creation date, and I can import all my pages into Adobe Acrobat Pro in sequential order by sorting the files from their earliest file creation date.

...Uh, yeah, so not sure why Stereo Data isn't supported on Macs. Is this software even necessary to scan books with the Canon Powershot A590 IS?

Thanks.

[univurshul]

Correction: I want the right camera to fire before the left. I was thinking left page.

[daniel_reetz]

it has something to do with getting the cameras to fire at the same time, maybe?...

StereoDataMaker is custom firmware that runs on top of the existing camera software. When loaded on the cameras, causes each camera to fire when they receive a 5V signal. With our switch setups, each camera gets a 5V signal at the same time, so they fire at the same time.

which leads me to my next question: is there a way to wire the switch to get the left camera to fire just before the right? --This would print a earlier file creation date, and I can import all my pages into Adobe Acrobat Pro in sequential order by sorting the files from their earliest file creation date.

This is a neat idea. Yes, but you'll need to use some kind of circuit to send a pulse to one cam, and then the other. Perhaps you could use an Arduino, or on Linux (maybe Mac, too) you can use GPhoto2.

...Uh, yeah, so not sure why Stereo Data isn't supported on Macs. This is a hassle and pretty lame, and is this software even necessary to scan books with the Canon Powershot A590 IS?
Thanks.

You can install it from a Mac. It's not lame, though it might be a hassle. You can also use just-plain CHDK if SDM doesn't work as easily as you'd hoped.

[univurshul]

Daniel, thanks. Sorry for the 'lame' tone. Perhaps I didn't look into it close enough.///Yeah, I'm trying to look at page-building in the simplest form; less software, real simple and fast processing. One program to do it all...just curious (because I have yet to get my scanner operational), if the cameras fire at the same time, then the creation date for left and right will be the identical each trigger, correct?---so, what if you change the "internal clock" on the left camera to be 1 second behind that of the right camera's time clock? Would this effectively create the conditions I'm looking for instead of wiring a mechanism to trigger a pulse? i.e., is file creation date on the SD card sensitive to what time you establish on your camera?

[daniel_reetz]

You can easily set camera time to whatever you want -- I think it will accomplish the same thing you're after. I think that's a really clever idea and one that could be useful for all of us with two-camera rigs...

[univurshul]

Daniel,

It's looking like there are no "seconds" when setting the time on the Canon Powershot A590 IS, but I was thinking about attempting a time snyc to lag 1-2 seconds via the :59, :00 minutes clock. I'll let you know how this works once I have the firmwares updated and wired.

This brings another questions about SDM StereoData and how power can be supplied to the Canon A590 IS: do the AC adapters that power the A590 IS conflict with the 5V power pulse sent to the USB port with StereoData? I ask this because I'm trying to eliminate using batteries all together on this rig, by powering my cameras via their respective power supply ports, 5V.

Thanks. Best Regards.

[univurshul]

...Just scanned my first series of books and had some really good results. I ended up using CHDK. There's a "delay sync" option on CHDK for remote, but I haven't really toyed with it too much as of yet

What I did find out was that delaying the camera's internal clock MIGHT create a later-creation date for the left camera, but it was fairly chaotically organized once viewed in Mac OSX X 10.6. --So setting the internal clock on the cameras to be offset by 1-2 seconds doesn't organize my page ordering as hoped it would--not yet anyway, it may be a computer clocking problem or a camera+CHDK problem, or something else too, hard to say for sure...

But here's what did work: I imported my captured jpegs into separate "events" in iPhoto 8.1 (left camera & right camera), where I then "batch processed" the groups of right camera's files by altering their dates to be set 1 minute apart starting at a given date, and the left camera's files to be the same as above +00:00:30 (+30 seconds). Once the two "events" were merged into one, and viewed by their "date", the page ordering was perfect. I then renamed the "titles" of the photos to be in sequence by assigning a number in place of the title (all of the above process took 90 seconds to establish, because it wasn't modifying the files, but tags to the files (908 page book, FYI)). My last step in iPhoto: I exported all the files to rename to their titles, i.e., page numbers on a file in the desktop (5 min). Lastly, I imported these new files into Adobe Acrobat, where I performed a final process of OCR, margin trimming, etc. (~2 hours; margin trimming is tricky on this book, but Acrobat allows you to trim even and odd pages, so that is super cool). I have Readiris, but for most OCR in English, Acrobat Pro does a really great job, I think.

Done Deal. Putting my first ebook on the cloud. Best Regards.

[daniel_reetz]

Hmm, so in other words we could just make a date-shifting utility and handle it that way? Cheers on your captures!

[univurshul]

I think the date-shifting utility is a great idea. I've been toying with Aperature 3 and iPhoto; one is a fancier version of the other. At the moment, it's a fast way to organize the raw jpegs and re-date them to stagger and order correctly...

Being new in the book scanning environment, I'm breaking-down the software process into segmented apps on the Mac. Here's what I'm thinking as a Mac OS X 10.6 user for comprehensive digitizing of a book and magazine library:

1--Import raw jpegs into iPhoto or Aperture (Dating, Title Organization and Image Retouching). Export-->
2--Import into Readiris (OCR & Page descewing) Save as high quality PDF
3--Open PDF in Adobe Acrobat (Margin Trimming and Final PDF Mb size/resizing.)
4---Open finished PDF into Calibre.app (ePub Conversion for iTunes; iPad, iPhone apps like Stanza or iBooks)

Some of these apps do the same thing, and I'm trying to find out what does what best, and run an Automater.app script to process a book in one go. I may get a chance to try Adobe Photoshop Elements in a week or so; the real key is intelligent batch processing. Fixing pages individually is out of the question.

If there is better and more efficient ways to do the above, I'm all ears. (But I run Macs only).

Thanks.

[StevePoling]

Do I understand this correctly?

Application takes as input two directories designated "left" and "right" it then changes all the times of image files in both directories so that each image file has a distinct time that is interleaved between two other file's times.

For instance, one way is to find the time of the earliest file in the left directory. Call it t-sub-zero, or t0. Then find the image file with the 2nd-earliest time, then change it to t0+2 seconds later. Then repeat this process for all subsequent images, setting the n-th such file to t0+2*n. Then repeat the process on the "right" directory. Setting its earliest image file to 1 second after "t-zero" or t0+1. Then iterate as before, setting the n-th file to t0+1+2n.

With both directories in this state, we combine all image files in a single directory and its page-sequence should correspond to its time-sequence. And (except for endpoints) all files should all be 1 second older and 1 second younger than another file that originally lived in the other directory.

Or better, the process should copy the image file to the destination directory, then change its time as described above. This avoids the possibility that the new t0+2n or t0+1+2n time might interfere with the "oldest file" time.