Jump to content


  • Content Count

  • Joined

  • Last visited

Community Reputation

609 Excellent

About decle

  • Rank

Profile Information

  • Custom Status
    0x10 bits or less
  • Gender
    Not Telling

Recent Profile Visitors

6,550 profile views
  1. OK, so something like this cannot be posted without curiosity getting the better of me... 🔍 ...and as is sometimes the case, this ROM is very interesting for reasons other than gameplay. The memory map of the ROM starts at $4040 rather than the normal $5000, which is strange. The game still has its normal cartridge header at $5000, with 4K of code through to $5FFF. However, it also has over 1K just hanging out between $4040 and $454A. Why would that be? Why not just let the code run on from $6000 through $654x, like every other 6K+ game? The answer to these questions might be on page 41 of the APh Datawidget User Guide. The APh Datawidget had 8K of RAM for program storage, but this RAM has to start at the boundary of an 8K page in memory ($0000, $4000, $6000, etc). So to get the RAM to reside at $5000 it was necessary to set the start address to $4000 and then disable the first 4K of RAM. This bit of the setup is described in step 8 on page 41 of the manual. So the largest program a standard Datawidget could debug is just 4K. Doh! Seeing the strange memory map I wondered whether the chunk of Space Spartans code between $4040 and $454A was relocated there as a workaround to allow it to be debugged on the Datawidget. Disassembling the code I found that was legit, was called during gameplay and did run correctly (i.e. not a ROM dumping artefact). I speculated that by ignoring the instructions in step 8 and reenabling the 1K segments at $4000 and $4400 it would be possible for the Datawidget to use this area. The reason the game code starts at $4040 and not $4000 is to avoid clashing with the repetition of the STIC registers from $4000 - $4021. Up to now this has only been a bit of idle speculation. However, in the course of their book research Tom & Braxton recently turned up the following 1982 memo from David Stifel, which they gave permission for me to share - nice one stifelLinkageMemo.pdf The memo describes how to link games larger than 4K using the APh toolchain, it also covers the memory limitations of the Datawidget and how to sidestep them by relocating code down at $4000, whilst noting the need to avoid the problem addresses from $4000 to $4021 (it also provides a work around for $4800 should the program be bigger than 6K). So whilst it is not guaranteed, I think there is a fair chance that this version of Space Spartans was set up for debugging on an APh Datawidget and for this reason it has a weird memory map. And I for one think this is pretty cool 😀
  2. Sorry, this is a long one and a bit off-topic, hopefully it is interesting 😉 The only two TV POWWW T-carts I'm aware of are the Kempner Football and KSpace T-carts owned by David Chandler's estate. I believe the whereabouts of these are currently unknown - David Chandler's estate sent them to the BSRs, but BSRSteve is not aware of them being received. In what way do you think that the titles on Sharp Shot are not representative of TV POWWW? The videos of two player KSpace and KFoot look to be very similar to the Sharp Shot versions of the games (there are one or two small differences - for example the use of the gun at the end of a KFoot game). I was wondering if this would pique someone's interest 😀 I think there are three pieces of circumstantial evidence that support voice control being used in at least some Fairchild TV POWWW installations. First Marvin Kempner's biography describes the trials of getting the technology working on page 205: "Early in March, we received the news from Fairchild that they were running a little late, but they assured us they would be ready prior to the April start date of the NATPE convention. Normally, this convention, held annually, would start on a Thursday or Friday, go through the weekend and one or two days into the following week. I wanted to be on the air the Monday prior to the start of the convention so that we could get any bugs out of the equipment and have the A.M. Los Angeles show with Regis rolling by the time the general managers and program directors registered. On the planned Monday morning, at 7:00 a.m., engineers from Fairchild and I entered the studios at KABC to set up our spanking new "TV POWWW" equipment. With everything ready to go by 8:30, Regis and Sarah having integrated the contest into the show, things looked great-until we discovered the equipment didn't work! Getting on line through the system was beautiful, but the voice activation was not operational. At 8:45, "TV POWWW" was cancelled for the show, and the engineers flew back to Santa Clara to continue working. Tuesday, they were back with several changes only to discover once again it still wasn't operational. By Wednesday morning, when we were still having trouble I was almost at my wit's end. Still, the engineers assured me it was only a simple bug, and they once again returned to Santa Clara. By Thursday morning, with Fairchild's engineers once again hooking up the system, I had visions of never making it for the convention, but lo and behold, everything worked. Regis, with his unusual ego, didn't make our life easier, but Sara Purcell, a tall sinewy blonde, was just precious. She was delightful to work with and had a wonderful sense of humor, even screaming out "POWWW" at the wrong time to see if she could make the equipment work from the studios. We went on the air, and Regis asked viewers to pick up the phone and call to play from home. On Friday morning, again we had no problems. For two hours after the show went on the air, the Los Angeles telephone system was badly taxed with people calling trying to be first to play "TV POWWW" on KABC-TV and asking all kinds of questions." Interestingly, I think that the final section of this video, starting at [18:41], might be a recording of that very first TV POWWW show which, if true, would be incredible: I like the level of explanation given, and how surprised the presenters are that the whole thing works. Also notice that Shooting Gallery has not been altered to add a timer, Sarah is using a stopwatch to time contestants. Sarah Purcell left A.M. Los Angeles later in 1978 and so this ties the recording to roughly the right period. The second piece of evidence is a Kempner advert on page 41 of Broadcasting magazine from February 26th 1979. This clearly states that the game is "Voice Activated!" and that this constitutes an "Electronic Break-thru!". I know, marketing is marketing, but it's not nothing. And the final piece of evidence, is a couple of videos of a Fairchild TV POWWW being demonstrated at the North Carolina State Fair in 1980: The first thing to note is that when presenters are talking to Barney or the crowd [4:30] they use a single mic with a foam wind shield. However, when playing the game this mic is augmented with a second one with no wind shield [10:53]. The additional mic is only used during play, once the game is over the host puts it down on the Fairchild console [11:49]. Later in the video we get a clearer picture of the monitor in the background [13:08]. You can see that the game seems pretty responsive to Danielle, I think a person would be doing well to keep up that fast, that accurately. However, notice that as Danielle makes her POWWWs longer, the shot retriggers, and when she starts to laugh it goes fully auto-fire [13:29]. Something similar happens in this video from Bozo's Circus where a child in the background seems to trigger the shots. I don't think an engineer pressing a button would do this. Instead, this reminds me of an amplitude or "clap switch" type circuit being used to generate input. Not exactly speech recognition as we would describe it today. If this is correct, the main purpose of a secondary microphone at the State Fair would be to isolate sounds made by the player and filter out background noise. Finally, later in the State Fair video Anessa seems to talk rather quietly, which the hosts think causes a couple of POWWWs to be missed [14:29]. As a consequence, they ask her to talk louder a couple of times. Again, if a person was triggering the shots I would not have thought this would be necessary. Oh, and if anyone can shed some light on what the black floppy disk looking object is, protruding from the front of the orange cartridge plugged into the Fairchild console, that would be great: Not TV-POWWW evidence, but somewhat related, Kempner went on to sell other voice services later in the 80s, like the "Telephone Poll" product advertised on page 106 / 107 of Broadcasting. Kempner received a patent for the technology behind this in 1984. This product collated automated polls where viewers called in and responded to a yes / no question that was broadcast over the air. It looks as though the "voice recognition" aspect of this was the caller making any sound at one of two silent points in a message generated by a voice synthesizer corresponding to yes and no. The audio of the caller during these periods appears to have been amplified and converted to a digital signal which was then counted if it was above a volume threshold. I'm not suggesting that any one of these is conclusive evidence that voice control was used at any particular deployment. And of course there are examples of syndicates specifically stating that voice activation was not used, for example TV-PIXXX. Rather, I think that together this is reasonable, if circumstantial, evidence that an amplitude based TV-POWWW voice technology probably existed, and may sometimes have been used. Moving on to Intellivision TV POWWW, the nearest I have found to similar video evidence is the fact that in the Brazilian SBT installation of the mid-80s the presenters seem to fiddle with an Intellivision hand controller around the start TV POWWW games and then don't fiddle with it again during play, much like the hosts of Fairchild installations: However, as the game starting sometimes seems to precede the disc tapping, and no power or A/V leads seem to be coming from the back of the console, I think there is the real possibility that this is an affectation and the Intellivision is a prop, or a shell with a controller linked back to the Kempner rack mounted Intellivision in the control room. Regardless, nothing can be inferred about what may, or may not, be supplying input via another controller port. I think it is also potentially telling that the later Intellivision POWWW advert from Broadcasting, the one that also advertises Telephone Poll as "able to actually recognize words spoken on the telephone!", does not mention voice activation as a feature of TV POWWW. Of course the question is, if Kempner had working voice technology for the Fairchild implementation, why would they not use it in the later Intellivision ones? One possible answer is patents. According to Marvin Kempner's autobiography, Fairchild, who developed the voice input hardware, "requested the right to file for patents on the voice activation and anything else they felt was warranted". Later, when discussing the origins of Telephone Poll, he states "TV POWWW was no longer using the Fairchild Camera Instrument Corporation's patents. The electronic industry had come far and we had found easier ways of producing our interactive equipment". A bit of Googling into Fairchild patents turns up the nattily titled "Electronic device for subtracting signals and associated system employing such device". Not obviously related to TV-POWWW, however, this patent covers technology to prevent the programme presenters' voices or other studio noise activating TV POWWW as a result of it being sent down the phone line and feeding back into the receiver. Without it, sound from the studio could not be fed to the player during game play, and presenters would not be able to coach contestants. Perhaps Kempner was wary of patent infringement and Fairchild's licensing costs were too high? Perhaps it was just easier to drop voice recognition in the move to Mattel? Obviously, much of this is speculation, however, if I was to hazard a guess, it would be that voice recognition did exist and was used in some Fairchild installations, but that it was dropped in the move to Intellivision for patent or contractual reasons. As always, counter evidence and arguments are most welcome. Cheers decle
  3. Yes, I think you're right Steve. However, it seems that Slots did see the light of day, at least for those living in Sacramento... As part of their book research Tom and Braxton uncovered this gem, complete with 80s innuendo, in the KeithR / BSR archive of documents: It seems that this version of Slots did make it to Marvin Kempner at TV-POWWW and a bit of Googling turned up the following video with it being played on KTXL in 1980 (TV-POWWW wasn't just for kids it seems): So there was at least one T-cart with a complete version of Slots on it. Looking at this TV-POWWW marketing document, although the pictures of Basketball and Baseball are not the standard marketing images used in the various catalogues, the games seem to be the same as the cartridge releases. Word Rockets differs to the released game, in that it seems to support four letter words, with letters missing from the beginning (and presumably end) of the word, and it uses an underscore to highlight the missing letter, rather than a space. Football and two player Space Battle appear to be images of the TV-POWWW versions that can also be found on YouTube. So I think there is video evidence of five Intellivision titles being used for TV-POWWW at the moment: Football - https://www.youtube.com/watch?v=CWLCKFMGEdo Slots - https://www.youtube.com/watch?v=cIqgKmY4DDo Soccer - https://www.youtube.com/watch?v=hGXf5V_UH7c Single player Space Battle - https://www.youtube.com/watch?v=weZGElhgEY0 Two player Space Battle - https://www.youtube.com/watch?v=G4tsMQl9fPo Then potentially there are the additional titles identified in the Kempner marketing document: Baseball Basketball Word Fun / Word Rockets Plus the extra titles mentioned in a Kempner advert on page 106 and 107 of this 1983 Broadcasting magazine: Boxing Astrosmash Frog Bog Horse Racing Space Hawk Unspecified Intellivoice game(s) (coming soon!) Marvin Kempner's autobiography lists the following games on page 216: Skiing Boxing Space Patrol - presumably Space Battle World Rockets - probably Word Rockets Astro Smash - he's doing well! 😉 Slots The March 2002 Intellivision Newsletter talks about: Astrosmash Skiing Word Fun / Word Rockets And finally, there are the games released on Sharp Shot. The BSR website describes how APh submitted four games programmed for TV-POWWW as what became this title. So this would add the cut down versions of: AD&D Sea Battle As a consequence, there could be as many as 16 Intellivision titles adapted for TV-POWWW. Presumably all on T-carts, as detailed in the marketing doc. Now that would be some collection! Perhaps a New Year challenge to CMart? 😉 It's also interesting that there is no mention of voice activation in the Kempner sales spiel. Evidence for voice control of Intellvision deployments seems to be very thin on the ground (basically non-existent). This is in comparison with the earlier Fairchild implementations where there is reasonably strong circumstantial evidence for it being used. Nevertheless, I feel a project coming on... Happy (early) New Year! decle
  4. This is very interesting. Could I ask a favour of those that own PlayCable manuals? Could you message me a list of the manuals and part numbers you own please (images work if you can't be bothered typing)? I'd like to try to build a list of the games broadcast and add it to the PlayCable Research and Development thread. Unfortunately the Voltron's attempt to do this back in 2013 seemed to peter out. Ahh, ignore me, the list is already on INTV Funhouse (as an aside, I think it is unlikely that either Bump 'N Jump or Master of the Universe were actually broadcast, as both games would need to have been cut up or cut down in order to fit into the PlayCable's 8K memory). I would still be interested if anyone other than CMart owns the PlayCable schedules, so we could reconstruct the dates on which games were broadcast. Assuming that all the titles in the auction highlighted were on PlayCable this would double the number of known 8K games broadcast to two (Motocross and Buzz Bombers). Thanks for any help. decle
  5. Sadly I don't think there are any remnants of Craps or Slots in Roulette. Here is a text file of the lower 8 bits of the game expressed as "graphics": roulette.txt The main graphics start at about $50e8, with some more at $57f8. Sadly no dice or cherries to be seen Taking a similar approach with Space Battle and Sharp Shot also fails to show the Cylon Basestar seen in the TV POWWW version of the game: spaceBattle.txt sharpShot.txt Is it just me, or does the Basestar look more like a Battlestar? And is it not a bit weird that you would be shooting a Battlestar? The TV POWWW videos of Space Battle (or KSPACE as the game was known inside Mattel) might show an interesting development titbit. You will notice in the single player version of the game the centre of your crosshair is very similar to the Space Battle one, and the Cylons sometimes turn and leave the screen, such that you can see the top or bottom of the spaceship, just like they do in Space Battle: However, in the (later?) two player version of the game, the Space Battle like target is missing and the Cylons seem to have more limited movement, never turning such that your can see their top or underside: This more limited movement is also seen the Sharp Shot version of the game, and might suggest that the two player version of KSPACE is a significant rewrite, more closely related to the Sharp Shot version of the game than its Space Battle based single player sibling. Anyway, if you fancy looking through game ROMs for graphics, here is the Perl script I used to generate the files above: stdBin2Gfx.pl Cheers decle Having found a better video of the single player version of KSPACE it also has a some interesting Basestar graphics that are momentarily garbled as it comes onto screen at 1:00:
  6. I thought it might be amusing to create a spoof Intellivision VIC-20 "emulator", showing the famous one-line maze program. This is pretty easy to do in IntyBASIC: ROM and BASIC source code: vic20.zip However, I had hoped to implement this in ECS BASIC. Surely this is doable? Clearly I had forgotten quite how bad ECS BASIC is. This is as near as I could get: Implemented with the following "one-liner": 10 CLR 20 SET A$="///////////////////" 30 FOR A=1 TO 19 40 IF(RN(0)>49)A$(A)=60 50 NEXT A 60 PUT A$ 70 GOTO 20 For a language implemented on a machine with reasonable graphics capabilities (for the day) ECS BASIC is poor. The list of problems I encountered is as follows: As we know ECS BASIC is dog slow because of the synchronisation with VBLANK. As it turns out this is the least on my concerns... You can't change the border colour!? It takes 18 seconds to set the background of the screen to a consistent colour (white in this case) using the following code: 10 FOR A=0 TO 239 20 BK(A) = 7 30 NEXT A All tiles can either have a printed character or a background colour, but not both - brings a whole new meaning to foreground / background mode 10 CLR 20 BK(0) = 1 30 PRIN "HELLO" 40 PRIN "HELLO" 50 BK(20) = 1 60 BK(40) = 1 70 GOTO 70 All printed lines are terminated with a newline, so you can't print in the 20th column without generating an empty line 10 PRIN "12345678901234567890" 20 PRIN "12345678901234567890" 30 PRIN "1234567890123456789" 40 PRIN "1234567890123456789" This also means you can't build a line incrementally (so faking someone typing is impossible) 10 PRIN "H" 20 PRIN "HE" 30 PRIN "HEL" 40 SET A$="H" 50 PUT A$ 60 SET A$="HE" 70 PUT A$ Oh well, back to the drawing board having learnt some things. 🙂 Cheers decle
  7. Hey Guys, Through October I took a look at the Intellivision debugger written by Rick Koenig back in 1982 / 1983 when he worked for Mattel as a Blue Sky Ranger. A twist of fate led to a version of this software being left in the production ROM of World Championship Baseball, where it can be found to this day. I blogged about this investigation as a RetroChallenge entry. If you've been following along, you will have seen that I had some success in getting Rick's work up and running. You can find a summary of my findings in the following document which, thanks to the kind permission of the Intellivision Productions, also contains the full source of Rick's debugger with my annotations by way of explanation. rickKoenigIntellvisionDebugger-20191107.pdf If you fancy playing with the debugger you can find a version of David Rolfe's demo game Killer Bees with the Rick's work embedded within it here: killerbees.zip Instructions for how to use the debugger can be found in my blog, in the document above and in the following video which shows it in use: As always any feedback is most welcome (other than the fact that I got the addressing mode of the MOV instruction at 3:40 wrong - it should be direct, not immediate mode). 😉 Cheers decle
  8. Love it! I did not realise that the test carts split the ROMs up in that way (I guess it's the usual combination of me being stupid, and never having opened one up). Of course, this really throws down the gauntlet in the category of "original things to do with the cart splitter". 👍 Challenge accepted! (success far from guaranteed 😉)
  9. Hi Steve, Thanks for sharing this it is very cool! For those that are interested, it looks as though NYIT Computer Graphics Lab was heavily involved in the early graphics scene and their alumni went on to places like Pixar and SGI. Looking at the video, there are some interesting bits and pieces. I'm not sure that it necessarily adheres to the rules it sets itself. For example, in the racing game with 2 colours per card: There seem to be a number of areas (notably curves and diagonals) which seem to have 3 or more colours in very close proximity... I wonder whether this was a piece of early analysis work designed to fix the specifications of the Super-STIC by providing a "feel" for how the different specifications will impact the user experience, rather than examples actual working hardware? As such they could be rendered on more powerful hardware and need not necessarily follow the rules exactly. On a slight tangent, did the items from David Chandler's estate include the T-carts described as containing Kempner Football and KSPACE, as described toward the bottom of this page?... https://papaintellivision.com/gIntv.php I ask because as Marvin Kempner was the man behind TV Powww, and therefore, I believe that these may well contain examples of the TV Powww versions of Football and Space Battle: and it would be great if curiosities these could be preserved, dumped and shared. Cheers decle
  10. Very cool. I wonder if the X-10 controller would be compatible with the ECS too? It seems that the Aquarius tape interface runs at ~600 baud (this is a bit data dependent as spaces are twice as long as marks) based on carrier frequencies of 1667Hz and 833Hz. I think the ECS is similar, but runs its cassette interface at half this speed, ~300 baud, using different carrier frequencies (4800Hz and 2400Hz). It doesn't look good. However, you mentioned in the YouTube comments that the hardware was later released as the Radio Shack Appliance and Light Controller for use with the TRS-80 Color Computer. This is interesting, as the Color Computer uses a different baud rate and set of carrier frequencies for its cassette tape again (1500 baud and 2400Hz / 1200Hz). So perhaps not all is lost? Paging the ECS Guru, @Lathe26
  11. Cheers, Things have been moving on apace. I think we now have a test driver that fires up the debugger and allows you to fiddle with its UI (instructions for the controls can be found towards the bottom of the latest blog post) Rick seems to have done a cracking job in such a short period of time; I believe during an interview with the Intellvisionaries (episode 26) he suggested he wrote the debugger in 2-3 weeks between working on Motocross and Masters of the Universe. The debugger has some cool features that aren't seen in the other Intellivision tools of the period (e.g. disassembling the current opcode) and one that isn't seen in JzIntv (trace to non-EXEC code). There is still quite a bit of work to do understanding how the debugger is put together. You can follow my progress here.
  12. Hey Guys, I thought I might give RetroChallenge a shot this October. Not done it before, should be interesting. Anyway, I think I'm going to take another look at World Championship Baseball. This time to see what we can find out about Rick Koenig's Intellivision debugger. As I find stuff, I'll blog about it here in a suitably RetroChallenge way. Everything should kick off at the beginning of next week, so if you have time, why not follow along at home? Cheers decle
  13. Hi everyone, I have what I hope is an interesting update of the PlayCable technical description to share this evening, which you can download here: playCableTechnicalSummary-20190918.pdf I suspect you will like the story behind this piece of equipment: The big ticket items are: p38 - A more detailed description of Joe Jacobs' and Dennis Clark's Development PlayCable adapter p6 - A summary of the DCX11A DataChannel card p37 - An overview of Jerrold development engineers using hacked PlayCables to play games ripped to audio tapes p19 - Resolution of a potential discrepancy in the understanding of transmission frequencies Enjoy decle
  14. Hi everyone, Just one update today, but it's a really interesting one. On page 26, Section 5.5 has been expanded to include an overview and pictures of the PlayCable based development system put together by Joe Jacobs and Dennis Clark. Here you go: intellivisionDevelopmentBackInTheDay-20190727.pdf And a bit of a teaser: Enjoy! decle
  15. This is great news, I'm looking forward to the results. I'm intrigued to see the complete font and how the real POPlexer handles "Sphinx Of Quartz, Judge My Black Vow". I think it's fair to say that POPlex-a-like does not do the best job I couldn't work out a way to get a full 4x4 tile font to fit into 64 GRAM characters, even if bits of similar glyphs like P and R are reused. It was also necessary to invent a number of glyphs, as they do not appear in cylondr's video: However, a look at the Intellivision catalogue suggested that, of the released Mattel titles, the most complex seemed to be "Blackjack & Poker" with 17 characters, and ignoring spaces, 11 unique letters. So, to keep things simple, POPlex-a-like applies a set of restrictions - a title can have no more than 20 characters, of which no more than 16 can be unique. It will be interesting to see what the real thing does with more complex titles like "Sphinx Of Quartz...", or The Dreadnaught Factor (180 letters long, with 28 unique characters!). It would seem that I spoke too soon about the completeness of POPlex-a-like, pride before a fall and all that. On watching cylondr's video again, the original Poplex-a-like was missing the jingle on starting a game, the "message from Marketing" when the demo cart is not plugged in and the all important "key clicks" when changing games. The following update adds all of these "features" and reduces the delay before kicking off the demo to 10 seconds: poplex-a-like.zip As you can see in this short clip: I think this gets close to what is shown in cylondr's video. The font and sound are probably a bit off, and there are some unknowns, for example whether the game launch jingle plays when starting the demo cartridge, but I think it's a reasonable first second stab. Either way, it's an item off the rather lengthy list of half finished projects. Unfortunately, the PlayCable's menu works in a different way to the POPlex-a-like. Essentially it is more like a game than a hypervisor. Happily it is just software, and with a bit of work it can be transformed... Cheers decle
  • Create New...