keilbaca Posted July 2, 2004 Share Posted July 2, 2004 So... due to the pics of the screenshots and the vectors, looks like Ben's simple mod is the best... If i read the vectors right that is... Quote Link to comment Share on other sites More sharing options...
+Nathan Strum Posted July 2, 2004 Author Share Posted July 2, 2004 Oh, your poor, poor Atari!!!! One of these days it's gonna kill me by spitting out my Challenge of Nexar cart at my forehead at 300 miles per hour. I told them to adjust it so that the burst/scanline amplitude ratio was 1:1. Maybe that's too much? But it was impossible to find chroma information without paying the $350 for the spec. How are they measuring it? Do you have the ability to measure the chroma amplitude vs. the color burst? Or do you know the proper ratio off the top of your head? If so, I would be forever indebted to you if you could pass along the correct value. That's something I'll have to look into. Those are cool vectorgrams. Are those from a software vectorscope package? Or are they from a hardware 'scope? It's built into Apple's Final Cut Pro. (I have a hardware one here, too, but it's not as easy to get screen captures off it.) Quote Link to comment Share on other sites More sharing options...
+Nathan Strum Posted July 2, 2004 Author Share Posted July 2, 2004 So... due to the pics of the screenshots and the vectors, looks like Ben's simple mod is the best... If i read the vectors right that is... Actually, Ben's is the farthest off. You want to look at the bright spots where the vectors meet. The further from the center of the scope, the more saturated the colors are. If they're out too far, the color is oversaturated. If they're not in the same relative postions clockwise, then the colors will be shifted. (This is all relative to the 2600's RF output, which, while noisy, are the 2600's "real" colors.) Also, if you're familiar with real SMPTE bars on a vectorscope, you should know that these aren't true SMPTE bars. They're 2600 approximations. The CD4050 is the closest to the 2600s "real" colors. The colors are slightly less saturated, and only offset a little bit. RF on the left, CD4050 on the right: RF on the left, Ben's mod on the right: Anyway... next time I'm feeling ambitious, I'll tweak the CyberTech mod and take some more screen captures. I'll also add vectorscope pics, while I'm at it. I think I'm going to need a bigger web page... Quote Link to comment Share on other sites More sharing options...
cwilkson Posted July 2, 2004 Share Posted July 2, 2004 So... due to the pics of the screenshots and the vectors, looks like Ben's simple mod is the best... If i read the vectors right that is... That's not how I read the vectorgrams. An ideal vectorgram would cosist of a set of perfect points, each pair connected by a single line. You'll notice that Ben Heckendorn's mod has a lot more noise (as evidenced by blurrier points and mulitple lines between points) than the CyberTech board. Neither is perfect, but the CyberTech image is much closer to the ideal image. They are both oversaturated. Actually, Ben's is more oversaturated. That's easy to adjust with the CyberTech board. I'm not sure about Ben's...I doubt you have that much control. But Nathan can probably give a better answer to that. -Chris Quote Link to comment Share on other sites More sharing options...
cwilkson Posted July 2, 2004 Share Posted July 2, 2004 Oh' date=' your poor' date=' poor Atari!!!![/quote''] One of these days it's gonna kill me by spitting out my Challenge of Nexar cart at my forehead at 300 miles per hour. Be glad it's not Chainsaw Massacre! I told them to adjust it so that the burst/scanline amplitude ratio was 1:1. Maybe that's too much? But it was impossible to find chroma information without paying the $350 for the spec. How are they measuring it? As far as I know' date=' with an oscope that syncs on video...the same way I was doing it. Do you have the ability to measure the chroma amplitude vs. the color burst? Or do you know the proper ratio off the top of your head? If so' date=' I would be forever indebted to you if you could pass along the correct value.[/quote'] That's something I'll have to look into. Actually, how far out should the vectors point? I see tick marks (100, 80, 60, etc)...I'm assuming that these are percentage points. But percent of what? Or more importantly, what should be the maximum percentage for a fully saturated color? Those are cool vectorgrams. Are those from a software vectorscope package? Or are they from a hardware 'scope? It's built into Apple's Final Cut Pro. (I have a hardware one here' date=' too, but it's not as easy to get screen captures off it.)[/quote'] Ooh....shiny! How much does that cost? -Chris Quote Link to comment Share on other sites More sharing options...
+Nathan Strum Posted July 2, 2004 Author Share Posted July 2, 2004 That's not how I read the vectorgrams. An ideal vectorgram would cosist of a set of perfect points, each pair connected by a single line. Neither is perfect, but the CyberTech image is much closer to the ideal image. They are both oversaturated. This is a perfect set of color bars: But these are real ones - not Atari 2600 ones. The problem, and I think it's a significant one, is whether or not the 2600 can really output a true set of SMPTE color bars. I don't know enough about the 2600's color palette, or about the color bar cart, to know if those bars are truly intended to be a real set of bars, or just an approximation. It would require someone to disassemble the ROM, and figure out exactly what those colors are supposed to be, in order to determine if it could be used it as a true reference for NTSC colors. I use it strictly for relative comparisons against a stock 2600. That's easy to adjust with the CyberTech board. I'm not sure about Ben's...I doubt you have that much control. You don't... Ben's goes completely out of whack very fast if the picture's even slightly off. Quote Link to comment Share on other sites More sharing options...
+Nathan Strum Posted July 2, 2004 Author Share Posted July 2, 2004 This is like playing some weird text-based version of Pong... Be glad it's not Chainsaw Massacre! Or "X-Man". Actually, how far out should the vectors point? I see tick marks (100, 80, 60, etc)...I'm assuming that these are percentage points. But percent of what? Or more importantly, what should be the maximum percentage for a fully saturated color? Does this help? Ooh....shiny! How much does that cost? FCP is $999 ($499 educational if you know a student or teacher somewhere who can buy it for you). They have cheaper software called Final Cut Express, but it lacks the monitoring capabilities. And you'd need a Mac, too. Your turn. (blip...) Quote Link to comment Share on other sites More sharing options...
cwilkson Posted July 2, 2004 Share Posted July 2, 2004 This is a perfect set of color bars [image deleted] But these are real ones - not Atari 2600 ones. The problem, and I think it's a significant one, is whether or not the 2600 can really output a true set of SMPTE color bars. It can't. The way the 2600 chroma works, is that it gives you 15 colors (+ gray) that are equally spaced around the color wheel. I'll give more details later, but that's really about all there is to it. I use it strictly for relative comparisons against a stock 2600. That's all you can do. Final Cut Pro is about as expensive as I expected. Ouch. So now I've dug out my vectorscope (Tek 520A rackmount == HEAVY!) and I'm trying to make it work. We'll see how it goes. The Tek page is (will be) very helpful. Thanks! [boop!] -Chris Quote Link to comment Share on other sites More sharing options...
+Nathan Strum Posted July 2, 2004 Author Share Posted July 2, 2004 It can't. The way the 2600 chroma works, is that it gives you 15 colors (+ gray) that are equally spaced around the color wheel. I'll give more details later, but that's really about all there is to it. I figured as much. I knew the 2600 was limited, but didn't have any specifics. So now I've dug out my vectorscope (Tek 520A rackmount == HEAVY!) and I'm trying to make it work. We'll see how it goes. The Tek page is (will be) very helpful. Yep, lots of good stuff there. I could stand to go through it for a refresher. Tube scopes are heavy beasts - we've got one of these, instead. Anyway, I hooked up my unmodded 2600 to my TV, and was able to see MindMaster okay, so the mod's going to have to be tweaked. I'm going to put it up on the scope at work, and knock the saturation down a bit, and see what I can get. [dit] Quote Link to comment Share on other sites More sharing options...
cwilkson Posted July 3, 2004 Share Posted July 3, 2004 [silence....your score increments.] Well, I spent all night playing with the vectorscope. And part of the morning. New toys are cool. The good news: it works! And the 2600's video is pretty impressive in all the tests I put it through. I was especially proud of the repeatability of the chroma dots, given what I'm doing to the chroma. The bad news: there is no way to generate primary or secondary colors with a 2600. As a said before, the 2600's color output is generated simply. An internal, analog delay line with 15 taps is used to generate the 15 phases for colored output. When properly adjusted, this should give 15 colors with their respective phases spaced evenly around the color wheel. Color 1, "yellow/gold" has a zero phase delay and is not adjustable. The color delay pot on the 2600's motherboard allows one to adjust the total delay between color 1, and color 15. Ideally, the phase delay between any 2 adjacents colors is the same (= 24 degrees). There's no easy way to make this adjustment the right way, so instead of 336 degrees, most people try to make color 15 have 360 degrees of delay. I.e. they set it to match color 1 visually. Even the videosoft cart does this. It has to. Not quite right, but we live with it. Now you know (probably more than) everything you ever wanted to know about 2600 color generation. Now, the weird news. Although the color burst is 40 IRE in the waveform view, the amplitude was around half what it should have been in the vectorgram. I couldn't figure that one out. Weird. I noticed that when I made the 2600 display a colorwheel, the aplitude (saturation) of each color was less than the one before it. But they were always the same. I.e. no matter what order I displayed them in, they would always be in the same location in the vectorgram. Weird. Any thoughts? Maybe the vectorscope isn't working as well I think it is? I don't know. I don't have a pattern generator unfortunately, so I can't do much rigorous testing for my equipment. Oh well.... -Chris Quote Link to comment Share on other sites More sharing options...
+Nathan Strum Posted July 3, 2004 Author Share Posted July 3, 2004 [Woo hoo - score!] Interesting info about color generation. I believe the technical term is "wonky". Something like that. As for your vectorscope, how does it compare with what I've posted on my site? Oh right. Forgot to mention. I added a page about this. (I was up all night playing with a vectorscope, too.) Tested. Disassembled. Adjusted. Re-tested. Screen-captured. Re-assembled. Written up. Over-explained. Updated. Uploaded. Good to go. [bloop] Quote Link to comment Share on other sites More sharing options...
cwilkson Posted July 3, 2004 Share Posted July 3, 2004 [Woo hoo - score!] Interesting info about color generation. I believe the technical term is "wonky". Something like that. I've also heard the word "desperate". As for your vectorscope, how does it compare with what I've posted on my site? That's a great page. Chock-full of good info. Thanks! I'm attaching a photo of my scope's output. I had to take about 20 photos before one turned out. Ugh. I think the vector mode must just be out of calibration. I had to set the gain to "maximum" and then fine tune it. I tuned it so that the burst amplitude was on the 75% line. Now it looks great. Of course, only after I got a photo to compare with your captures, did I remember to adjust the saturation. So I didn't bother. You'll notice in the photo that the phases appear to be the same as yours, but the saturation level varies with the angle. "Green" is the worst... The other thing I forgot to mention before was that the phase delays that I see aren't uniform from color to color. I could explain that by blaming the manufacturing variation between transistors in the TIA. But I don't want to. That seems lame. Do you have the ability to load arbitrary ROMs into your 2600? If so, I could write a colorwheel program for you to look at all the various colors at once. [pong] -Chris Quote Link to comment Share on other sites More sharing options...
+Nathan Strum Posted July 3, 2004 Author Share Posted July 3, 2004 I tuned it so that the burst amplitude was on the 75% line. Now it looks great. Of course, only after I got a photo to compare with your captures, did I remember to adjust the saturation. So I didn't bother. Looks good though. I had an easier time - I just hooked the monitor output of my scope up to my capture card. I've found that for taking low-light pictures like that, I set the camera on a tripod, use a shutter release or a timer, and let it take however long of an exposure it thinks it needs. And since you don't have a pattern generator, the Video Essentials DVD is a good, affordable substitute. There are quite a few test patterns on there. You'll notice in the photo that the phases appear to be the same as yours, but the saturation level varies with the angle. "Green" is the worst... Yeah, that's really off. What kind of 2600 are you testing it on? I noted between my four and six switchers, that the TIA had different part numbers. So I wonder if they changed something, or if it's just due to different suppliers. Six-switch TIA: Four-switch TIA: The other thing I forgot to mention before was that the phase delays that I see aren't uniform from color to color. I could explain that by blaming the manufacturing variation between transistors in the TIA. But I don't want to. That seems lame. It wouldn't be entirely surprising though. Do you have the ability to load arbitrary ROMs into your 2600? If so, I could write a colorwheel program for you to look at all the various colors at once. I have a SuperCharger, so if you can send me something I can load into that, then yes. Although I'll be taking some time off starting Thursday, so after that I won't be able to do any more testing for awhile. BTW - take a look at any game that uses color cycling on the scope. It's pretty cool. [bzzzt!] Uh oh... that didn't sound good. Quote Link to comment Share on other sites More sharing options...
cwilkson Posted July 8, 2004 Share Posted July 8, 2004 Looks good though. I had an easier time - I just hooked the monitor output of my scope up to my capture card. No fair! I guess software is nice like that. The other thing I forgot to mention before was that the phase delays that I see aren't uniform from color to color. I could explain that by blaming the manufacturing variation between transistors in the TIA. But I don't want to. That seems lame. It wouldn't be entirely surprising though. No' date=' not at all. But I'd rather be able to blame it on my vectorscope. I'm sure it isn't working right. Of course it's used, and it's been banged around a lot before I got it. And the last time it was calibrated was in 1991, etc. But still it's disappointing. Do you have the ability to load arbitrary ROMs into your 2600? If so' date=' I could write a colorwheel program for you to look at all the various colors at once.[/quote'] I have a SuperCharger, so if you can send me something I can load into that, then yes. Check your inbox. BTW - take a look at any game that uses color cycling on the scope. It's pretty cool. Yeah, I noticed it on the Cuttle Cart splash sceen. [bzzzt!] Uh oh... that didn't sound good. Yeah, that was my paddle blowing up. I got it fixed finally, but I'll have to concede the match. -Chris Quote Link to comment Share on other sites More sharing options...
EricDeLee Posted July 8, 2004 Share Posted July 8, 2004 Anyone know if the following screen will work with a modded Atari 2600? I'm having problems locating a screen. http://www.buyritegames.com/product_inform...=1119&last=1034 Quote Link to comment Share on other sites More sharing options...
+Nathan Strum Posted July 8, 2004 Author Share Posted July 8, 2004 Probably. It might depend on the mod used. Ben uses LCD screens with his mod on his 2600 portables. The CyberTech and CD4050 mods should work, since the DVCAM deck I captured the tests with has a small LCD monitor that worked just fine with them. But I'd make sure it had a money-back guarantee before buying it, just in case. Quote Link to comment Share on other sites More sharing options...
keilbaca Posted July 9, 2004 Share Posted July 9, 2004 I just did Cracknel's mod on the 6-switch, but its shaky on most games. I did it on a 4-switch and everything was stable. Just wanted to let you guys know that basically there is a difference between the two. Quote Link to comment Share on other sites More sharing options...
EricDeLee Posted July 9, 2004 Share Posted July 9, 2004 Nathan Strum: Well... I already have a PsOne screen... but I can't get the Atari 2600 to run off of it through the AV jacks. Any suggestions? They say that some PsOne screens are finicky. Any help on getting that to run will save me a headache! Quote Link to comment Share on other sites More sharing options...
cwilkson Posted July 10, 2004 Share Posted July 10, 2004 Nathan Strum: Well... I already have a PsOne screen... but I can't get the Atari 2600 to run off of it through the AV jacks. Any suggestions? They say that some PsOne screens are finicky. Any help on getting that to run will save me a headache! How is your 2600 producing A/V? (which mod?) -Chris Quote Link to comment Share on other sites More sharing options...
EricDeLee Posted July 10, 2004 Share Posted July 10, 2004 Ben's Hack.... To tell you the truth I haven't tried anything yet. Except to hack the Atari. My PSone Screen is in Tulsa right now and I'm out of town. But I was hearing a lot of people say that it will not work with a PsOne screen for some reason. Quote Link to comment Share on other sites More sharing options...
mos6507 Posted July 11, 2004 Share Posted July 11, 2004 If there is enough interest I can run a Cybertech signal through the hardware vectorscope I have inside my Toaster/Flyer system and compare the results. DPS Waveform/Vectorscope cards aren't particularly expensive anymore on the used market. I saw one guy asking $399 for one off of google. Pro standalone hardware is an order or two of magnitude more expensive. Sample Image off the web: Quote Link to comment Share on other sites More sharing options...
cwilkson Posted July 11, 2004 Share Posted July 11, 2004 Ben's Hack.... To tell you the truth I haven't tried anything yet. Except to hack the Atari. My PSone Screen is in Tulsa right now and I'm out of town. But I was hearing a lot of people say that it will not work with a PsOne screen for some reason. I've heard that too. The signal Ben's mod produces is far from the NTSC spec and it's unbuffered, and I'm guessing that there's no buffer inside the PSone screen. So it's probably loading the circuit down beyond recovery. Adding a transistor buffer to the circuit might fix the problem. Try the following: +5v ___ | | | / |/ Input O------| 2N3904 | | /| /_| | 75 Ohm +---///--------O Output | | / 220 Ohm / | | --- The output resistor should be 75 Ohms. But I just made up the emitter resistor value...feel free to change it for better results. I wouldn't go below 75 or 100 Ohms though, just to save power. -Chris Quote Link to comment Share on other sites More sharing options...
cwilkson Posted July 11, 2004 Share Posted July 11, 2004 If there is enough interest I can run a Cybertech signal through the hardware vectorscope I have inside my Toaster/Flyer system and compare the results. That would be cool. I sent you a rom that displays all 16 colors (at 50% luma). Check your inbox. There is a definite problem with my scope. There's no reason for the signal amplitude to change. It's basically a digital output. There's no change in frequency, so there wouldn't be a changing reactive effect (capacitive or inductive). It has to be the scope. But it would be interesting to see the colorset displayed properly. DPS Waveform/Vectorscope cards aren't particularly expensive anymore on the used market. I saw one guy asking $399 for one off of google. Pro standalone hardware is an order or two of magnitude more expensive. Heh. Even that's out of my budget. I tend to not buy equipment until the magic $150 mark even when I have money. But it's certianly something to keep in mind if I want to do more video stuff in the future. (Oh, the HORROR!!!) -Chris Quote Link to comment Share on other sites More sharing options...
EricDeLee Posted July 28, 2004 Share Posted July 28, 2004 I will be trying your suggestion, cwilkson, as I am unable to get anything on my PSone screen. (Screen works well with every other system... just not my hacked Atari. Besides... I've got issues to work out with the display on my Atari. I almost got it right... but not quite.) Quote Link to comment Share on other sites More sharing options...
Like us on Facebook
Follow us on Twitter
AtariAge on Instagram
Watch us on YouTube
Chat on Discord
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.