Naked Oak

[Frederick_P]

I have been looking for an easy way to convert photographs to patterns. I need some patterns of classic nudes for client keepsake boxes, sconces and the like. Michael from Carvebuddy agreed to make some from some of my reference photos. Unfortunately recent family issues and the demands on his time have put this project at the bottom of the list. I saw a program called 3DSEE a while back but the site no longer is active and so I have no idea what is happening with this project. It seemed perfect for me as it used a series of images shot in sequence and stepped slightly from side to side to feed the software with the 3D information and produce a greyscale height map. This would allow me to shoot live subjects in the studio to produce the information I would need to create a true 3D relief for the CW machine. So if anyone knows what has happened to 3DSEE or knows of another program that would allow me to do something similar, I would be interested to know.

Michael was, however, able to produce one pattern. I decided to try it out by making this box to hold my office papers. The sides of the box was made from oak baseboards and the only carving is the figure on the lid. Being a photographer I have high expectations for detail which seems impossible to achieve from carving in wood. Certainly the original photo is better. I think it also would have worked better from a hardwood such as maple as the wood grain in the oak is not flattering to the figure, but this project was just a test of the pattern and so I chose to use whatever scraps I had lying around the shop. The edge of the lid is purple heart.

Michael also asked me to shoot some cowgirl pinup images as he has received many requests for this. If you have other suggestions I am sure he would like to hear them.

[WRW]

Great work as always Frederick, I liked the title you picked for the post rather appropriate.
Have a great day.

[hess]

Fred it looks cool This is the subject of many of my Lithos have not attempted in wood yet. I use a program call MeshCam Art. You can use it free for 30 days and he will let you have more time if you ask.
I have to say that all I use is the smoothing part and it does what I need but there is much more to the program I know wood will not be as clear as a litho most of the time but it has great detail

[Frederick_P]

Hess, this program accepts stepped images as input? How is the 3D info produced?

[mtylerfl]

Hi Frederick!

Your projects are outstanding - I love the boxes you are making. Sorry I'm so behind! Getting caught up gradually, though. Anxious to 'get cracking' on more specialty patterns along these lines, but we also need to wait until the CarveWright store programmers can devise a system of a 3-way payment split in order to start offering these to the public. More complicated than originally thought, is my guess.

Below are a couple photos...one is a montage of the Aspire software screenshot (the software I used to create the model), your original photo, and the carve...the other is the finished carve alone. I agree with you that a photo wood-carve relief does tend to look a little better with a 'plainer' wood grain. I carved this on Select Pine.

Creating great quality relief patterns from photos can be challenging - I am not currently aware of any reasonable shortcuts and/or software that makes the process particularly 'easy' or particularly 'fast' (unfortunately).

I don't know if folks on the forum realize that you are listed as one of the top 100 photographers in the world! Your professional work certainly speaks for itself. And, I don't know if the folks here realize how notable Oleanna is in the modeling world. What an amazing and unique team you both make! I'm looking forward to providing more relief models based upon your photos in the future.

[Frederick_P]

Michael, it's good to hear from you again! The western style pinups are located on my web site under the portfolio section "western theme". Have a look and see if anyone has any interest in these as carving patterns. I am still hoping to get more detail out of the carvings. I can appreciate what a challenge this must be which is why I was hoping anyone could suggest any other software that would help. If there is any way the photos could be enhanced or shot differently to assist in your efforts, please let me know. I understand that there are 3D scanners out there that will scan a full body image but the cost of these units are more than Sarah Palin's yearly ammunition budget.

I understand how busy you are so no worries. Just keep us all up to date on how they are going.

Regards,
Frederick

[henry1]

Hess, this program accepts stepped images as input? How is the 3D info produced?
Can you share this program would love to make one as a gift great one ,,keep it up

[eelamb]

Frederick, try this program called shadermap pro, sells for $19.95. Examples are below, Images brought into software, and map generated, giving you about 5 outputs of maps types. One below you will know where it came from, but remember it was a smal image.
One horse image is the orginal image used to make the pattern.

[mtylerfl]

Hi Eddie,

Just so I'm clear on this...the center image was NOT created using the shader map pro, correct? The center photo looks like it was created in a much higher quality software program to me (ArtCAM or Aspire?).

[eelamb]

MT
The center image is a photo taken of the actual carousel horse. It was used to make the first image. As with any rendering of photo's light spots and dark spots need to be taken care of before doing the rendering, after all it is a height map output. I did nothing to the original images just brought them into the program and rendered the output that you see there, then imported the image into CW. Again no editing at all, what you see is what the CW softare read it in as. With a little work in PE the first image would be much better.

For details as in the center photo you need to model it in a 3d program, Silo that I use will give these details too.

[Frederick_P]

Eddie,
Thanks for the info. However, please remember that my images are copyrighted and, regardless of the quality of the results, please do not use them to make .mpc or any other files without written consent. Thanks for the info, I will look into this program.

[mtylerfl]

...

For details as in the center photo you need to model it in a 3d program...

Thank you, Eddie. That's exactly what I thought.

[hess]

Hess, this program accepts stepped images as input? How is the 3D info produced?

Hey Fred
Not sure about the 3D as all my stuff is for lithos

[hess]

Can you share this program would love to make one as a gift great one ,,keep it up

here is where the trial is http://www.grzsoftware.com/art.php

[eelamb]

Frederick, sorry I removed the mpc, had not thought of that since I was just showing you. I had already removed the original from my pc after making the pattern for that very same reason.

[Frederick_P]

Thanks, Eddie!

[eelamb]

With the comments on 3d and detail, as frederick pointed out when carving you loose detail, this is due to the STL importer, where some detail will be lost and the tapered bit, where more will be lost. Therefore making GREAT models with lots of details may not give you a great carve. Even the type of wood used to carve the pattern on will give variations in details. Red oak for one is not a good choice, it has open pores, white oak is not as porious. Maple, cherry, and walnut are just a few to mention that carve great. Even soft woods like pine and cedar do a good job. I have been impressed with cedar as to the quality of the carve.

Mind you I am not complaining, just stating some of what I have seen, the software is great as well as the STL importer, could they be improved sure, but look at other CNC and their software and cost, you can not beat the CW and STL importer software.

Frederick, I had also removed the pattern and mpc from my pc, they were for demonstration only, even went back to make sure I did not miss any.

[henry1]

here is where the trial is http://www.grzsoftware.com/art.php

just dont no how to use it yet,, and I meant can you share the pattern thank you in advance

[mtylerfl]

Hi Eddie,

As far as detail...it's really not so much to do with the STL Importer - the detail tolerance that can be entered for an STL exported from Aspire for example, can be as small as 0.000000000000001" (although I've never attempted it and don't plan to - I don't know if that actually works, even though it allows me to type in all those zeros - I am fairly certain it would "kill" my computer!). As a rule, I export my STL's at a detail tolerance of 0.0001" (a ten-thousandth of an inch, I think), but have done exports at 0.00001" (a hundred-thousandth of an inch, I think). Anyway, you can readily see that the STL end of things doesn't really account for lack of detail, unless it is a low-res STL in the first place.

The primary practical limitations are as you stated...the size of the bits we use and the material we work with. Super-fine detail won't show up, either because the bit tip is just too big to render it and/or the super-fine detail is "erased" due to material machining/chipping. As a pattern designer, I always try to keep the practical limitations in mind while creating a relief model. It's a balancing act and always a mental challenge for me, deciding what I can and cannot "get away with" in a relief model destined to be machined with a tapered 1/16" or a straight 1/8" bit. However, more often than not, I am surprised at the detail that actually CAN be rendered by our "miracle machines"!

[Frederick_P]

Michael,
What are the detail limitations of the CW software? What is the maximum resolution that can be turned into a pattern? I always want as much detail as possible because one never knows how big the pattern may be output to. The limitations of the bit and wood are more apparent when the carve is small, but, as in the case of this box lid, if the pattern is output at a very large size, the "hidden" detail that may not be visible on smaller carves would be revealed. But if the CW software limits the resolution of the greyscale height map, then this would be the technical limit that a pattern could be rendered.

Just wondering.

[Pratyeka]

Just to add my 2 cents on that last comment from Michael, the amount of details also is dependent upon the size of the final carving. You just can't squeeze as much fine details into a 6 inch carving than into a 36 inch one. For that reason a carver will choose which details to put into a carving, and will select those details that will be aesthetically closest to the artist's desires.. which brings me to my point: Taking a photograph and turning it into a "good" carving requires extensive editing, as the amount of details into any photograph is way above what is needed for a "good" carving. The art here is to drop those details that are detrimental to the "feel" of the carving.
Also, consider that some carvings are created using "classical rules of proportions", which will subtly impress beauty into your mind without you consciously knowing why you find the carving so beautiful. Let's face it, most of us are not modeled with classic proportions. The talent is to create a pose that will present the subject from an angle that will result in those magic proportions.

So, in short, if you want to turn a photograph into a carving and find the result disappointing, blame the details of the photograph that are out of place on the carving. With practice and technique, you can improve the result, but to make a great carving, you need to compose the photograph with the carving in mind, then remove from the photograph all those unnecessary details.

Frederick, I took a long look at your website, WOW! And I used to be an avid reader of Heavy Metal when I was younger, I'll have to dig for the one with your feature in it.

[dbfletcher]

Michael,
What are the detail limitations of the CW software? What is the maximum resolution that can be turned into a pattern? I always want as much detail as possible because one never knows how big the pattern may be output to. The limitations of the bit and wood are more apparent when the carve is small, but, as in the case of this box lid, if the pattern is output at a very large size, the "hidden" detail that may not be visible on smaller carves would be revealed. But if the CW software limits the resolution of the greyscale height map, then this would be the technical limit that a pattern could be rendered.

Just wondering.


I have always thought the "limit" was the carving depth divided by the decrete steps along the z. In our case the depth limit is .8" and still 256 levels of gray... so without taking in to consideration of the bit thickness, we have a resolution along the z at best is = 0.003125. I know there are 16 bit ptns now, but i believe i read that it still get translated to 256 discrete steps.

[Frederick_P]

Doug,
But what is the limit of the software? In other words, can the software accept, say, a 3000x6000 pixel image and create a pattern that has this much detail in it?

[dbfletcher]

The Z axis is the limiting factor. And the current depth of carve along the z is .8" (althought it used to be 1.0"). It is hard to equate your 3000x6000 pixel 2D image to a 3D work space. Pratyeka can probably provide a much clearer answer.

[mtylerfl]

Michael,
What are the detail limitations of the CW software? What is the maximum resolution that can be turned into a pattern? I always want as much detail as possible because one never knows how big the pattern may be output to. The limitations of the bit and wood are more apparent when the carve is small, but, as in the case of this box lid, if the pattern is output at a very large size, the "hidden" detail that may not be visible on smaller carves would be revealed. But if the CW software limits the resolution of the greyscale height map, then this would be the technical limit that a pattern could be rendered.

Just wondering.

Hi Frederick,

When it's all said and done, the machine is responding (carving) based upon 256 shades of grey, so that is the ultimate limitation. That's why we will sometimes notice slight "stepping" on an otherwise "smooth" curved or tapered surface. There is some type of interpolation going on with the "16-bit" support for greyscale during the PTN conversion process which has improved this behaviour. I don't claim to have a complete understanding of what is going on under the hood in that respect, but it seems to be better than it was before.

Before the STL Importer was available, I always saved my relief models as greyscale heightmaps out of Aspire and ArtCAM, then imported those into Designer where they were ultimately converted to the PTN format. However, I now save my relief models as STL files, import those via the CarveWright STL Importer, then save them to the PTN format. Usually (but not always), the STL import method yields a better quality pattern than the greyscale import method. For awhile, I was doing both, then doing a side-by-side comparison in Designer to see which one looked better...sometimes it was too close to call, but usually the STL-import-conversion was better, so that's my preferred method now.

As far as size of carving, I am of the school of thought now that I should design my reliefs as shallow as possible, trying to stay fairly close to the "standard" 1/4" depth. There are exceptions to this, but that's what I strive for. The reason is that any pattern will "flatten" the details when enlarged at a certain point. The only cure for this is to increase the pattern depth and height settings in Designer to restore the detail. If a pattern was already maxed out at say, .7" and a user is using 3/4" material, there is little or no room to increase the depth. Thus, if the user enlarged the pattern and the details flatten out, he is stuck. By designing in relief at the lowest possible depth that still looks good, we can leave the user some "wiggle room" in case he/she needs to increase the depth to restore detail after an enlargement. And yes, an enlarged pattern can show more detail that would otherwise be hidden at a smaller size, BUT the user has to know to increase the depth and height to get those to show up.

Again, it's a bit of a balancing act on the part of the pattern designer, and it is good practice to consider a few "user variables" when creating a professional pattern. Often, when doing custom work, I will ask the client "What size do you plan to carve this at?" If I know from the get-go the preferred size, it makes my life a little easier and I can design a pattern accordingly. For the "masses" I use sort of an "industry standard", that is, I design most patterns at a fairly high resolution and try to keep them between 1/4" to no more than 1/2" in depth. For example, a 6" x 6" square work area will need about 3 million pixels to give me enough resolution to be able to sculpt smoothly and get good detail. The pattern must cover as much of the 6x6 space as possible in order to benefit from the resolution setting. If I have a 6x6 area and create a pattern that is only 3x3 in size on that workspace, then I have used only a small fraction of the total resolution available to me. The result would be a low-res model, even though my workspace was set at high res. I hope I'm not losing you on this, but I'll bet you get the idea alright. There are many other factors to consider too, but it would take a few more pages of explanation!

Bottom line...design a pattern at as high of resolution as practically possible, so that when it is converted, the interpolation during PTN conversion has more data to draw from and gets you a better pattern. I think. ;)

[eelamb]

Before the STL Importer was available, I always saved my relief models as greyscale heightmaps out of Aspire and ArtCAM, then imported those into Designer where they were ultimately converted to the PTN format. However, I now save my relief models as STL files, import those via the CarveWright STL Importer, then save them to the PTN format. Usually (but not always), the STL import method yields a better quality pattern than the greyscale import method. For awhile, I was doing both, then doing a side-by-side comparison in Designer to see which one looked better...sometimes it was too close to call, but usually the STL-import-conversion was better, so that's my preferred method now.


Very well said MT, this is what I was referring to with the STL importer. I too have learned to size the pattern in the .25 to .5 range. If you have to resize to get this area for the cut, then you loose details, and the end results are a flatter pattern. I agree with you in your post, along with pratyeka, and Doug too. There is more than just the pattern to account for when doing a pattern. I am not a professional and do not plan on being one, it would be nice if there was a book written on the CW, usage, techniques, STl importer, repair, designer, and so on. To help new buyers to understand this great machine. As it stands I see a lot of frustration from new users attempting to create the great patterns others are making. The CW and software alone are a large learning curve for any new user.

Maybe some of you guys should get together and create a book! (LHR's documentation is good, I am talking about a book to take over from there)

[Frederick_P]

OK, so at the risk of appearing dense, if there is no limitation in the software as to what the pattern's resolution is (and I am referring to importing a greyscale height map, not using the STL importer), then if the machine can cut to a depth of 0.8", we divide that by 256 (shades of grey) and come up with a depth resolution of 0.003125". The other limiting factor would be the minimum incremental steps of the stepper motors moving the traction belts and the carving head. If we assume that our image contains square pixels, then the resolution would be limited to the smallest step the machine can make. We would then multiply that number by the maximum size of board that the machine can take and the result would be the optimum resolution of the greyscale file used to make the pattern.

Right?????

[mtylerfl]

...

Maybe some of you guys should get together and create a book! (LHR's documentation is good, I am talking about a book to take over from there)

It's a little embarrassing...we've been talking about creating a third-party book for a couple years now...but, still hasn't gotten much beyond the talking stage (Jeff Birt and I did a rough table of contents and that's about it). Another project on my/our "back burner" is creating a professional, high-quality video tutorial series...already have the chops, equipment and software...looking for 48-hour days to actually produce them! We'll get there...

[mtylerfl]

OK, so at the risk of appearing dense, if there is no limitation in the software as to what the pattern's resolution is (and I am referring to importing a greyscale height map, not using the STL importer), then if the machine can cut to a depth of 0.8", we divide that by 256 (shades of grey) and come up with a depth resolution of 0.003125". The other limiting factor would be the minimum incremental steps of the stepper motors moving the traction belts and the carving head. If we assume that our image contains square pixels, then the resolution would be limited to the smallest step the machine can make. We would then multiply that number by the maximum size of board that the machine can take and the result would be the optimum resolution of the greyscale file used to make the pattern.

Right?????

Hi Frederick,

Sounds good to me! I never really thought about the "formula" too much. One of the CarveWright Techs could probably offer some insight. (BTW...I believe the CW uses servo motors.)

[Smoken D]

That would be great Michael. I am one who understands watching much better than reading about it. Reading not my thing when it comes to understanding something.

[dbfletcher]

We would then multiply that number by the maximum size of board that the machine can take and the result would be the optimum resolution of the greyscale file used to make the pattern.

Right?????

The only comment I would make here is that along the X axis there really isnt a limit (well.. eventually even with supports the weight of the piece would be a limiting factor, but for this type of discussion I think we can say X is not constrained.) Along Y we are limited to 14.5" but there are work arounds for that... so it still comes back to the Z axis is the biggest single limiting factor with the current machine.

[Frederick_P]

Doug,
so along the Y axis, what is the finest increment that the machine will resolve? Is it the same for the X axis? Are the traction belts able to move the piece in as fine an increment? At the end of the day, it is the machine that determines the final resolution.

[dbfletcher]

Doug,
so along the Y axis, what is the finest increment that the machine will resolve? Is it the same for the X axis? Are the traction belts able to move the piece in as fine an increment? At the end of the day, it is the machine that determines the final resolution.

From the Carvewright FAQ: (I'm not sure how up to date it is)

http://www.carvewright.com/2010CWweb/machinefaq.htm

What is the resolution of the machine?
The core capability of the CarveWright is as a carving machine. The machine configuration and control software are quite different than a classic mill. As such, the terms resolution and accuracy are hard to correlate to what you expect from a classic milling machine. Our controllable resolution is .00025" on the Y and Z-axes and .0015" on the X-axis. The accuracy of any single raster carving line is on the order of .005" - .010" in the Y and Z and the step width is between .005" and .010" depending on the quality setting you choose. When you are talking accuracy of position of two elements across a 60" board the value is even higher.

What is the ouput quality of the cutting passes?
Like a printer, projects can be configured to run in a DRAFT, NORMAL, BEST or OPTIMAL quality settings depending on the situation. Draft is the fastest and should be used on any first time "test" carves. BEST works great on tighter grain woods leaving little to no feathering or clean up required. OPTIMAL should be used for softer grained woods that will be final pieces and takes about 3 times longer than the BEST mode. The quality mode dictates the speed of the cut and amount of material removed per pass. On the best quality settting the spacing between passes is .006". Lower quality setting means less time on the machine, but perhaps more elbow grease. Higher quality setting means more time on the machine, but less elbow grease.

[dbfletcher]

btw... I hope i dont come off as sounding like a Guru with this machine. Most of what I spout is simply recycled or regurgitated info I have read elsewhere on the forum. I have been around long enough now that a lot of info has managed to "stick" in my head though. So when I see a question that I remember I have read about previously, i'm happy to relay what I recall.

[chebytrk]

btw... I hope i dont come off as sounding like a Guru with this machine. Most of what I spout is simply recycled or regurgitated info I have read elsewhere on the forum. I have been around long enough now that a lot of info has managed to "stick" in my head though. So when I see a question that I remember I have read about previously, i'm happy to relay what I recall.

Come on Doug... don't be modest. We know you a member of the Gran PooBaahs and rank up there with Obie Wood Kanobie ! : )

[Frederick_P]

Doug, so then to conclude, if we assume that the smallest reliable resolution is 0.0015" and we would not carve anything wider than 14", then the maximum usable resolution of a greyscale image used to create a pattern would be 9333 pixels on the smallest side. Am I correct then is assuming that if I created a file with this resolution that the machine (forget about the 1/16" bit or the type of wood) would be able to resolve this image?

[dbfletcher]

Doug, so then to conclude, if we assume that the smallest reliable resolution is 0.0015" and we would not carve anything wider than 14", then the maximum usable resolution of a greyscale image used to create a pattern would be 9333 pixels on the smallest side. Am I correct then is assuming that if I created a file with this resolution that the machine (forget about the 1/16" bit or the type of wood) would be able to resolve this image?


Your math is the same as mine (14" /0.0015) and if you limit the image to 256 shades of gray I think you would get a very close representation in designer to your image.