Victorian Porch Brackets

[skintigh]

32019

The 1865 originals were cut from a single piece of wood around 35"x48"x1.75" but I had trouble finding that stock at Home Depot.

I stripped one side of a mostly intact original bracket and after multiple failures over a couple weeks I managed to scan it in 35 overlapping images on a flatbed scanner, then photo-stitched the scans together and stretched it until the dimensions were right because they were off a few inches. (My scanner was adding artifacts on 2 sides that confused the photostitcher and made me miserable until I trimmed 1 pixel off each side. On other projects I stripped, sanded, and painted white the side I wanted to scan and that worked great.)

Each is cut in several pieces out of 3/4" MDO plywood (the local Shopbot can only do 2'x4', and I made 8 halves per sheet this way instead of 4), I glued them up, did tons of sanding and used tubs of wood putty repeatedly as I was new to CNC and my first cuts were UGLY. Then water sealed, primed, and painted a few coats. These are the smallest of the 12, others have longer tails as every span on the porch is a different length (I made a batch of interchangeable tails).

In later revisions I made some of the gaps between the details large enough to fit things, things like a file or a paint brush or a finger with wood putty on it. Turns out that's important. I have 7 done, if I ever finish the porch I'll post photo.

[Chuck Keysor]

Hello Seth. That is an interesting project. Thanks for sharing.

Do you by any chance happen to have a picture, or a partial anyway, of the original bracket? I am interested in Victorian houses, and would like to see a picture of the house this goes on, IF that is handy.

I once had to make some replacement parts, but I didn't have anything to work with except for the old photo. Well, I had the house, that had had all of the trim ripped off before siding was installed in the 1960s. I had to climb on the house, and look for nail holes and shadow lines (after the siding had been removed) to get a fix on the actual dimensions.

Thanks, Chuck (PS: The parts in the new picture are covered in primer, and they looked much better when painted in contrasting colors.)
http://www.talkshopbot.com/forum/attachment.php?attachmentid=32020&stc=1http://www.talkshopbot.com/forum/attachment.php?attachmentid=32021&stc=1

[phil_o]

Beautiful!! What software did you work with?

Phil

[Chuck Keysor]

Thanks Phil. I used Aspire for the design work. The first image shows better how I had to work on my models. http://www.talkshopbot.com/forum/attachment.php?attachmentid=32024&stc=1But because of problems I had with the resolution of my 3D model, and problems with my Shopbot, I had to do a HUGE amount of hand work. So I resonated with Seth's comments about all of the hand work he had to do. http://www.talkshopbot.com/forum/attachment.php?attachmentid=32025&stc=1http://www.talkshopbot.com/forum/attachment.php?attachmentid=32026&stc=1

Seth, instead of using a scanner to capture an image of the original bracket, why didn't you simply take a digital photo? As long as you get far enough back to avoid camera distortion of the image, you could fix the image in Photoshop, and be spared what seems like a lot of work to stitch together all of those scans.

Chuck

[skintigh]

Phil were you asking me? I can find the name, I think it was Microsoft software for photo-stitching then VCarve Pro.

Chuck, I tried taking photos from far back but didn't trust them -- not crisp photos, there was like 1/4"+ thick paint on the details, pieces broken off or wedged in the wrong place and then caulked/painted, most were rotted and split, etc. And scanning seemed like it would be "easy..." So I had to take them down anyway to strip the paint. I also tried taking lots of photos from different angles with another project and using software to make a 3D image but didn't get great results and gave up, since I was sure scanning was the easy way to go. Which it might have been, had I realized my scanner added lines at the edge and had a slight lip that made scans un-level which killed alignment.

In the end my first brackets are almost identical to the one I copied, and very close to the others which all differ a little. Later I kinda averaged them and made some gaps slightly bigger, but unless you lay one on top of the other you can't really notice. However mine are 1.5 inches thick instead of 1.75, I didn't think 50% more work was worth it for 1/4" more material.

Here's a photo of the original 6. I might be able to reuse 2+/- of them, but it takes sooooo many hours to remove all the paint.

32027

Why are my photos baby size?

[Chuck Keysor]

Thanks Seth for sharing the photo of the original porch. That is a great looking porch.

As to the process, from what I see, I would have used a photograph, then correct tilt/visual errors and get the proportions exact in Photoshop. I would have imported the photo into Aspire, and then manually placed vectors/nodes in Aspire on top of the photo. Then I would have done any visual touch-ups on the vector image to fill in any missing parts. But that is just what I would have done. You have to do, of course, what you can do with the tools that you have.

Now, since you mentioned that you may be able to reuse some of the originals..... I have seen where original parts are still on a building, new ones are made, and all the old ones are dumped. The problem is for future historians. Any old house investigators would instantly see that the brackets are not original, based on materials used, their total freedom from old paint, and the thin dimensions not filling the old spaces/paint lines on the porch. Then they wonder, how well was this reproduced? Maybe this is nothing like what was originally here...…. So as a historic example, the value of that house will be diminished, because without photographic evidence, no one will reliably know what was originally there. If you can re-use the best original, you will have left an important artifact for future architectural historians. Any old house expert will quickly spot the original. And at the worst, if the original best bracket goes into the attic, with notes written on to the part... Though such things can easily get thrown away by some clod 30 years from now cleaning out the attic.

Sorry, I have digressed, but such points are interesting aspects of working on old houses, and realizing that you are helping to preserve things for the future.

Also, as to your small pictures, I don't know.

Thanks again, Chuck

[phil_o]

You should post this on Vectric's forum. :)

Phil

[scottp55]

Seth..As far as the pic(the last one anyway)...If you click it, it will appear larger in a window, click that pic again, and it will show up full screen and with a zoom icon.
And by the way..Welcome:)
scott

[skintigh]

Still not sure who Phil is talking to, probably not me though.

Chuck, manually placing vectors sounds like a nightmare unless I'm missing something. Every line is a unique curve. When I scanned them I'd get say 200 vectors in each little curve, I'd clean up the knots, then have the software make a best fit curve. Is there a way to do that manually with just a few vectors per curve? Or do you manually stretch and manipulate each curve vector until it matches? The little tuning I did in the Aspire software made me want to murder whoever wrote that software... It's like MS Paint, but with a much much worse and archaic interface.

I don't plan on trashing the old brackets, unless one is hopelessly rotted, I'll just store them in the basement or garage. And I may re-use 2 of them in the center of the porch, so people will see the proper thickness when walking under it. I'm re-using corbels where I can, but I needed more of everything (the porch was 1/2 Victorian, 1/2 Craftsman but with terrible detail work and mis-matched columns.)

I really wish the previous owner had stored the original windows when they replaced some. The vinyl replacement windows are already warped, cracked, don't stay open, don't lock closed, are un-fixable and un-serviceable and need re-replacement already. Meanwhile the 150-year-old windows just needed a little TLC and they were good as new for another 150+ years of use, and now they open, close and lock effortlessly.

[Chuck Keysor]

Hello Seth.

a) I am guessing that Phil is suggesting I post my old example to the Aspire forum. I think that I did, maybe 4 years ago. Though I'm not really sure. It is so old now, but it seemed relevant to Seth's work, so I dusted it off.

b) Drawing in Aspire, by roughly placing points/nodes, and then adjusting them to draw good curves is an important skill to work on. If you use auto-trace, you get gobs of knots, far too many nodes, AND, the auto-trace function will not correct for missing details!

These drawing tools aren't difficult because they archaic, they are difficult, because they are powerful. The fact that you find manual drawing to be difficult and/tedious just shows that you need to learn how to do it. I actually took a class in Adobe Illustrator, maybe 10 years ago, and at first I hated trying to draw with the "pen tool". But with practice, I got to be proficient.

However, Illustrator's pen function seems more complicated, and more powerful than Aspire's drawing tool. I say that as I seem to be able to draw well enough with Aspire when manually tracing. But when I tried to draw something in Illustrator a couple of years ago, it was very difficult, and I had simply forgotten too much. So that says that while Illustrator and Aspire's drawing tools are similar, Aspire is much easier, so you should draw some comfort from that as you learn how to use it.

Honest, for the work that you seem to be tackling connected to old houses, learning how to easily draw/trace with Aspire, would be very important in improving your work. (Learning Photoshop and Illustrator would also be very important, in my opinion.)

c) I agree with your observations on new windows. I am starting a small project on a neighbor's house tomorrow. The person who measured for the replacement windows on the front of the house measured to the old sill. Before the new plastic windows were installed, someone replaced the window sills. The replacement sills were undersized by about 1/2", and one part of the sill was horizontal,,, ie no slope, so rain would sit on the sill instead of running off! Then, when the plastic windows were installed, they rested the bottom of the jambs on the sill. But because each sill was undersized, there was a 1/2" gap between the top of the plastic window, and the original window frame..... So the fix, was to install a 1/2" strip at the top of the window to fill the gap! Now, there is a gap between the bottom of the new window, and the flat sill...…….. Someone went and globbed a bunch of caulk to fill the gap a couple of years ago, but it is coming loose...………. As you observed, everyone would have been better off to have just fixed up the original window...……… Oh well...……

Good luck, Chuck

[skintigh]

> These drawing tools aren't difficult because they archaic, they are difficult, because they are powerful.

I don't know, maybe that's true for generating the CnC files. Or perhaps my hopes were too high, but I'd say the image tracing tools were difficult and terrible because they lack any power/intelligence... Maybe it would have been faster to manually click tens of thousands of points for my large designs, then maually making those curves, but that just seems like something that should have been automated decades ago. I thought tools like Adobe Photoshop do a great job of tracing an object, I didn't expect to have a nightmare when scanning in Aspire. I also thought scanning would be far more accurate than a human, and I wanted "perfect" copies. But in the end I manually adjusted just about everything so that was a poor goal from the onset.

While I was able to get slightly better results and reduce knots in Aspire by mucking with the resolution of nodes and reducing color levels, I think the biggest fix was to do an end-run around the software. I painted one side white and re-scanned it, and left the rest dark. That way, even the stupidest software on Earth cough aspire cough couldn't screw it up too badly. The high contrast let me reduce the color levels way down, too, without all the glitches VCarve made the first time. Dents and flaws in the bracket would still get traced, but I think I also used image editing software to clean up some spots before letting Aspire see/butcher them. In other spots I'd erase a know then use the curve-fitting tool. I wish I took more notes, but towards the end I did the smaller corbels in much less time.

The Aspire interface kinda lacks common sense. It seems designed for a 640x480 screen or smaller like it's still 1990, with tiny, text-less buttons with hieroglyphics on them. I believe the technical name for that interface style is "Mystery Meat."

Then if you're adjusting nodes and you accidentally click just one single pixel off, the mode you are in changes against your will and the function you were using is gone and all of your work selecting nodes is undone. Whoever programmed that obviously had very little common sense.

Speaking of common sense, obviously people are going to want to print a design and compare it to what they scanned or compare it to physical objects to be sure the size is correct. So naturally Aspire left out that feature.

So you try to convert it to another format so you can print it from less terrible software. Aspire thought of that too, and intentionally crippled cross-compatible formats.

Just getting a design out of Aspire and into something the a laser cutter could use was an Odyssey of wasted time. I had to save in one Aspire version then open it in a different version then save it to some vector format then open that in different software and convert it into something I could us on the laser cutter. I don't remember exactly, it took an hour or 2 to figure out and I made a hand-written chart of which formats each software could use. All of which suggests that if Aspire ever goes out of business all your designs may be forever locked away in an unusable format.

HULK SMASH

[phil_o]

Obviously I wan't clear. I was suggesting that the work on the Victorian brackets should be shared on the Vectric forum since the work was done with Aspire.

Phil