165 replies to this topic

[atariksi]

Stephen, on Mon Dec 7, 2009 9:54 PM, said:

atariksi, on Mon Dec 7, 2009 9:47 PM, said:

You mean a turbo mode switch so we can toggle between 1.78979Mhz and 1.84374496Ghz (1024*original) so we can do 512 cycles of instructions every color clock?

And here I was hoping to just get ANTIC to double its data rate so we could get a nice 80 column display! I guess I should aim higher

Stephen Anderson

Nah, you missed the discussion about NTSC limitations. Even if you do 1024*40 = 40960*200 mode, you won't get clear 640*200 on NTSC monitor.
Perhaps, with a good multisync monitor and RGB connection...

[candle]

it seems like You need a drum scanner Curt, prepress studios, or dtp studios working for newspapers might have what You need
another option is to use fixture (table like construction, except it wouldn't have the top) and to use plain digital photocamera
fixture will give you constant angle and repetable coordinates for taking the pictures and can be build almost from anything - its verry similiar process to how old paintings are digitalized

[RevEng]

or he could go the cheap route and just take multiple scans from a regular scanner and auto-stitch the results.

Edited by RevEng, Tue Dec 8, 2009 12:01 PM.

[ivop]

RevEng, on Tue Dec 8, 2009 12:00 PM, said:

or he could go the cheap route and just take multiple scans from a regular scanner and auto-stitch the results.

Yes, and the stitching could be done by somebody else if it's too much work for Curt

[candle]

and rotate every piece of the schematics by few degrees to the right or left
but.. at least we would finally have them

[Curt Vendel]

I'm going to try this route first, I need to pick up a scanner where I can remove the top and then I'll use some white foamboard and weight it down ontop of each section and scan and stitch them together.


Curt

RevEng, on Tue Dec 8, 2009 12:00 PM, said:

or he could go the cheap route and just take multiple scans from a regular scanner and auto-stitch the results.

[Rybags]

People practically throw out the old parallel port scanners, and I'm fairly sure the low-end Canon ones don't have any electrics going through the lid.

[Mathy]

Hello guys

That's true. I own a Canon Lide 500F and if I fold the lid all the way, it's like I removed the lid. All I have to do is make sure the paper touches the glass, otherwise the scan is out of focus.

greetings

Mathy

[jaybird3rd]

Heh ... I have an old USB HP ScanJet 2100C with a plastic lid that broke off at the hinges. It still works fine, and I actually think I like it better without the lid attached.

[ClausB]

Curt Vendel, on Mon Dec 7, 2009 1:41 PM, said:

I have the schem's for the GTIA, Antic and others, they are on 6' X 6' print outs, I need to access to large format scanner to digitize them.

Why not hang them on the wall and photograph them with a good digital camera?

[RevEng]

There's not enough resolution in today's best cameras to manage that feat.

If he takes a picture of a 6'x6' print with a 12MP camera (4000x3000 resolution), it means the resulting image will be 42dpi. That's assuming the lens isn't the limiting factor, which is a huge assumption.

To get up 200dpi he'd need a 276MP camera. To get up to 300dpi he'd need a 622MP camera.

So either way, using a camera or scanner, he's going to be stitching. Scanning is less error prone route.

Edited by RevEng, Sat Dec 12, 2009 8:46 AM.

[Rybags]

I'd go the scan.

To help with rotating, use something like a fine red pencil and measure out and do dots every 6 inches or so to form a grid pattern.

Then all you need to do is rotate each image to align your marks - should make stitching them together somewhat easier.

IIRC, Corel Draw has a half decent stitch option and there's utilities around dedicated to joining panorama pics together. Photoshop... as usual the jack of all trades and not very good at most of them.

[BuddyBuddies]

Curt is DA MAN! GREAT news!

[Nathan Strum]

Curt Vendel, on Mon Dec 7, 2009 1:41 PM, said:

I have the schem's for the GTIA, Antic and others, they are on 6' X 6' print outs, I need to access to large format scanner to digitize them.

Curt Vendel, on Tue Dec 8, 2009 2:33 PM, said:

I'm going to try this route first, I need to pick up a scanner where I can remove the top and then I'll use some white foamboard and weight it down ontop of each section and scan and stitch them together.

Get one of these:
Epson GT-20000

11"x17", 600x1200 dpi (which is plenty), and the lid is removable (there's a small switch that pops up when the lid is off, but you can tape it down to get it out of the way). Excellent scan quality, good speed, and the frame of the scanner is flush with the glass - so you don't risk creasing whatever you're scanning when you weight it down. We have nine of its predecessors (the GT-15000) where I work, and it's the best large format scanner we've found for the money (and we've been scanning animation with them - so they hold up to a lot of use). We use VueScan to drive them, which is a full-featured, often-updated cross-platform scanning application. I'd be glad to scan them for you, however I know shipping something irreplaceable like that is probably out-of-the-question. But if you're ever in the L.A. area, you're welcome to scan them here.

As an aside, if you're scanning paper documents that also have printing on the back, you can avoid bleed-through by putting black foam core behind it. It makes the scan a little darker (easily fixable in Photoshop) but you won't see any reversed text showing through on your scans.

Edit: As for assembling them - I put together a handout on how to use Photoshop to assemble animation pans (using the Difference layer attribute). Skip to pages 3 and 4 for the pertinent information. The problem with auto-stitching solutions is that they often blur and distort edges of images to align them, which you definitely don't want with schematics. Also, since the edges of a scanner are inconsistent in scanning brightness, the best approach is to do a full scan for each section, but have a lot of overlap for each scan (at least three inches) so you can trim off any areas along the edge which are darker.

Edited by Nathan Strum, Mon Dec 14, 2009 1:37 PM.

[AEX]

RevEng, on Sat Dec 12, 2009 8:45 AM, said:

There's not enough resolution in today's best cameras to manage that feat.

If he takes a picture of a 6'x6' print with a 12MP camera (4000x3000 resolution), it means the resulting image will be 42dpi. That's assuming the lens isn't the limiting factor, which is a huge assumption.

To get up 200dpi he'd need a 276MP camera. To get up to 300dpi he'd need a 622MP camera.

So either way, using a camera or scanner, he's going to be stitching. Scanning is less error prone route.

Not sure that's accurate,

Most print houses assume 300dpi minimum for most applications (the human eye doesn't see much more than about 250dpi). Anything larger consumes memory size and is only beneficial for truly large print jobs or to display exceptional accuracy. A typical 10MP camera can throw out an image easily at 3600 x 2400, and if you were to print that image out at say 16 x 24, the resulting dpi works out at around 325dpi.

DPI is referenced from the 72-points-per-inch standard set by the printing industry - the fact that the camera market keeps pushing up the reference of mega-pixels = more quality is to some degree accuarte, but does not mean that you need the type of MP camera you have mentioned above to print larger ouput effectively (not even hubble has that type of power!).

If you scan the image at 300dpi and stitch the results together (always best done manually because each scan is going to have light and colour changes), the image will be larger, but still 300dpi.

Karl

[Nathan Strum]

There are problems that using a camera has though, which scanners avoid:

• Lighting - You need to use stand lights to get good, consistent results. A flash or room light won't cut it for something like this.
  
• Shadows from folds - Proper lighting can minimize this, but it's going to be nearly impossible to make a document that large perfectly flat in order to photograph it.
  
• Lens distortion - It won't photograph square, unless you've got the right lens/distance combination, or use software to remove the distortion after the fact.

It depends on if you just want a nice picture, or an archival file. A camera will give you the former, a scanner will give you the latter.

[RevEng]

AEX, on Mon Dec 14, 2009 1:42 PM, said:

Not sure that's accurate,

Most print houses assume 300dpi minimum for most applications (the human eye doesn't see much more than about 250dpi). Anything larger consumes memory size and is only beneficial for truly large print jobs or to display exceptional accuracy. A typical 10MP camera can throw out an image easily at 3600 x 2400, and if you were to print that image out at say 16 x 24, the resulting dpi works out at around 325dpi.

He's talking about 6 foot by 6 foot documents. That's 72inches by 72 inches. 2400pixels/72inches=33pixels/inch.

Because MP values are length*width, the values quickly get absurd when you try to approach scanner resolutions with a camera, which is what I was demonstrating.

Working the equations the other way, if you want to capture 72 inches across at 300 pixels per inch, you need a sensor with 72*300 (=21600) pixels in its shortest axis. Since camera manufacturers use 4:3 perspective, the long axis will be 28800 pixels long. 21600*28800=622080000pixels=622.08MP

Put another way, don't you think a 6 foot by 6 foot print is going to be a bit hard to read at 16"x24"? There *was* a reason they printed it so large in the first place.

ALEX said:

If you scan the image at 300dpi and stitch the results together (always best done manually because each scan is going to have light and colour changes), the image will be larger, but still 300dpi.

Once its in the computer, the dpi is irrelevant - it's just an attribute that can be changed at whim. At that point what counts is the absolute number of pixels in the image.

[Curt Vendel]

I'm going to use an HP flatbed scanner after the New Year, laying the print down on the office floor, and moving in a grid across it to capture the entire plot, then will stitch it together.



Curt

[Stephen]

Can't wait to see this beast! Remember the GITA on breadboard thread? Now we have the schematics, someone can go and wirewrap one.

Stephen Anderson

[Zerosquare]

Congratulations on an awesome work, Curt !
(I remember seeing a post from you on DeepChip several years ago, asking for help on how to read those tapes )

But I want to know
You did find a fab which is willing to manufacture custom ICs without charging an arm and a leg ? Do they have a specific process, or a way to create "cheap" masks ?
I've always heard that chip manufacturing had extremely high fixed costs (altough I guess that having a very coarse geometry compared to today's chips may lower the cost).

[Wickeycolumbus]

Stephen, on Mon Dec 14, 2009 4:36 PM, said:

Can't wait to see this beast! Remember the GITA on breadboard thread? Now we have the schematics, someone can go and wirewrap one.

Stephen Anderson

It would be pretty hard to wire wrap a GTIA (or any of Curt's discoveries) from the GDS alone Much easier to get new parts made.

[Stephen]

Wickeycolumbus, on Mon Dec 14, 2009 5:47 PM, said:

Stephen, on Mon Dec 14, 2009 4:36 PM, said:

Can't wait to see this beast! Remember the GITA on breadboard thread? Now we have the schematics, someone can go and wirewrap one.

Stephen Anderson

It would be pretty hard to wire wrap a GTIA (or any of Curt's discoveries) from the GDS alone Much easier to get new parts made.

Yeah - that was the point. Might keep a few folks busy for a while though.

Stephen Anderson

[Wickeycolumbus]

Stephen, on Mon Dec 14, 2009 6:27 PM, said:

Wickeycolumbus, on Mon Dec 14, 2009 5:47 PM, said:

Stephen, on Mon Dec 14, 2009 4:36 PM, said:

Can't wait to see this beast! Remember the GITA on breadboard thread? Now we have the schematics, someone can go and wirewrap one.

Stephen Anderson

It would be pretty hard to wire wrap a GTIA (or any of Curt's discoveries) from the GDS alone Much easier to get new parts made.

Yeah - that was the point. Might keep a few folks busy for a while though.

Stephen Anderson

Oh never mind, I had not seen Curt's post. I thought you were talking about using the GDS to wire wrap a new one (rather than looking at the schematics Curt is going to scan). My bad

[phuzzed]

Curt.... You have done an AMAZING job... I can't stress that enough. I'm very excited about this. Keep up the good work.

[Bryan]

Can we take up a collection to get them scanned professionally?