The Joys of Photogrammetry
As I move forward with prototypes for producing my line of HO Studley mallets I am aided by a number of different assets. For starters I have in hand almost 20 of the castings from Bill Martley, in a variety of alloys, to work with in taking the rough metal castings to a point of “finished” that I would feel comfortable in providing to interested customers.
Second, I have a boatload of photographic images of the mallet from about every angle possible (it might seem adequate to have several dozen images, but they are never enough). Reviewing them does make me reflect on the unbelievable resource residing in my compewder; between my images and Narayan’s images I have almost 7,500 pictures on my hard drive .[As a snarky aside, I note that on occasion someone on the interwebz requests (or worse, demands!) that Narayan make all the high quality unpublished images available on-line for free, somehow not registering the facts that 1) the images are the creative property of Narayan and/or Chris, and 2) the images are the result of an investment of (literally) tens of thousands of dollars. If you have ever expressed this sentiment, grow up. — DCW]
Third, a topic I will address in a later post, I have several molds made directly from the original mallet.
Finally, and perhaps the strongest impetus foundational to this project, I have the enthusiastic approval of Mister Stewart to proceed with my effort to make replica mallets because, and I quote, “People should be using this mallet.” That endorsement is a great encouragement to me and I will produce a tool and make it available only once I am proud of it bearing my imprimatur. At some point very soon I will embark on designing the logos to be stamped or engraved into the mallet itself.
During the many episodes of us examining and photographing the tool cabinet and all its contents, one of the first projects Narayan and Chris conducted was to take “study shots” of every tool so that I could use these pictures as pneumonics for my own work in constructing the book. A second exercise was to formally photograph every tool with a scale included in the frame. Though a standard tool for documenting artifacts in the museum/conservation field from whence I emerged, it has never been more valuable to me than it is right now. Since my documentation for the mallet was not infinitely complete I still had a number of minute details I needed to establish in order to replicate the tool. To cross electron beams with Roubo, I wanted my replica mallet to be accomplished “With all the precision possible.”
Thus the images with the scale in the frame are my “go to” tools for making sure I get the details exactly right. I can fiddle with the file to make sure the printed image is precisely the size of the original artifact, but it is much easier to simply employ the tools of photogrammetry to the task. Although I believe software packages like Sketch-Up do this automatically, I am old school having learned classical photogrammetry in Architectural History classes back in the 70s. Once again the learning of the past solves the problem of the present.
Printing out one of the pictures representing the orientation I needed to make the handle (made from dalbergia, just like the original; I do not have much dalbergia so I might switch to swietenia mahoganii or another tropical hardwood at some point) with the scale included, I simply cut out the scale and used it to measure all the critical dimensions. As you can see the photographic representation of the scale differs greatly from the true scale, that difference is irrelevant because the image of the mallet itself is exactly the same scale as the photographed scale.
This exercise of using a “wrong sized” scale may be disconcerting to some, but it was the daily order of business when I was a patternmaker. The patternmaker’s tool kit includes a variety of rulers called “shrink scales.” These modified rules are used to lay out and construct a pattern to be used in making the molds for metal castings, and since each metal alloy shrinks a little bit when it solidifies the precise size of the pattern must reflect that reality.
At this point, especially as I create the handle, photogrammetry is more than a mere tool. It is the irreplaceable element.
How comparable is the scale of the ruler to the object in doing this kind of thing? Is parallax an issue? Is the error negligible?
Easy for frontal view; but not that easy for lengths oblique on the picture.
My very first thought when you said you had 7500 images was “I hope he has his files backed up and somewhere safe.” (This from a guy who has run head first into Murphy’s Law on more than one occasion with my computer.)
Also, has Mister Stewart ever said, “People should be using this vise?”