Is there such a thing as an interlaced picture for the +4? It seems that word "interlaced" for the +4 pictures is sheer disinformation. I checked code of programs that were claimed having interlaced graphics - actually they do not have such graphics. The only programs that have something that could theoretically turn on the interlace are generated by p4fliconv but my thorough checking shows me that they are not interlaced too. It seems that we have been used to apply this word to pictures that just toggle two images every frame. It is not a proper interlace. A proper interlaced picture should flicker but it should also show every line of its doubled raster. But C+4 "interlaced" pictures just flicker but don't show every line. I worked with the Amiga 500 which has a proper interlaced mode supported by its OS. So it is easy to use 640x512 interlaced mode instead of plain 640x256 mode. This interlaced mode gives us true 512 lines. So I can suppose that we should use other word instead of "interlaced". Maybe quasi-interlaced is ok? I use it for my new Mandelbrot program that toggles two images. IMHO maybe it is possible to use some TED magic and get the proper interlaced picture on the plus4 but I don't know how to do this.
Re: Is there such a thing as an interlaced picture for the +4? I agree the term was misused a lot and indeed applied to situations when pictures were merely showing alternating frames. However, true interlace IS possible albeit far from trivial and you are correct in saying that not many programs implemented it and pfliconv is one of them. The first one was TLC FLI Collection though. I seem to recall that it had some trouble synching the proper lines with the frame so occasionally it was showing interlaced lines in the wrong frame.
Re: Is there such a thing as an interlaced picture for the +4? The scene always called this "temporal filtering" as interlace. Although it is obviously a misnomer, after all these decades it really isn't worth trying to change the status quo. Call your image flipping version interlace without a second thought. If you manage to come up with a real interlace solution you may call it true interlace or something.
I remember that once someone from the German part of the scene, maybe Bionic, posted a screenshot of a memory dump displayed on screen that suspiciously looked like a true interlace screen. My hazy memory tells me that it actually generated proper interlace output. But I have no idea if it can be still found somewhere or lost for good.
Re: Is there such a thing as an interlaced picture for the +4? gerliczer probably refers to this.
Looking into my precious "+4_previews" folder, a place whispering forbidden secrets that no man in the universe must hear ;) , I've also found am interlace test by Bionic, which I now try to rebuild in some ways here. I remember that his intentions were to keep up the work about his 400x400 ilace stuff, continuing to follow the initial path, and he asked for a very flat and simple 72x128 picture in 3 color only, probably to be placed touching the borders, so I did this girl's bust. Later, he disappeared from the scene and nothing tangible followed :(
what I achieved once I managed to rebuild his work from those mixed files I have is this: he used a C64 multicolor picture drawn by SIT in 1987, set it in some ways as real interlaced picture...but shrinked/squashed/whatever. Didn't have the time to check it on a real machine (but I'm gonna do for sure...), and strangely it doesn't work on plus4emu (did it test on emulators, does it actually work on real machines?), but I got this on YAPE, which I gonna show you with a 32bit interlaced BMP, and a short video from YAPE (OpenGL, default CRT emulation).
Re: Is there such a thing as an interlaced picture for the +4? @Litwr, All, I ran a few tests recently with p4fliconv of IstvanV, and I am fully convinced that he has successfully implemented REAL interlace / interlaced FLI on Plus/4. (he may not be the only one, but this one is easy to test/prove) By real interlace I also mean what you do, using TV tech terminology (576i = PAL CRT TV)
Please look at these pictures, all made with his converter, run on a real +4, and shots taken with a mobile phone. 1st two pictures are the same, they are non-intleraced "normal" Hi-Res FLI (the diff is that for the first photo I used Video port and a decent S-video - HDMI converter, 2nd is via RF cable). These should be 320*248 pixs (using Y border as well) The 3rd one is the real interesing, it is Interlaced Hi-Res FLI, and easy to see the actual double resolution, 320*496. Flashing is barely noticable anywhere on the screen, it stands still. Unfortunately this one displays picture _only_ via the RF cable, could not make it work on S-video. I also made test images in multicolor FLI mode, but they are less convincing to be real interlace.. (still can be though)
Also these 2 photos taken from Dreamtime 2k18, supposed to be Larry's converter. I suspect this is also interlaced, 400(+) lines, but less sure than with the previous example (btw this was the only pic in that slideshow where I saw the higher resolution) This one at least plays on S-video, but as you see a complete new world opens if you look at it using RF cable. The S-video looks blurry, probably due to the flashing + the camera's shutter time, the RF one stood quite still. What do you think?
I wonder if there is something "wrong" with my HW setup, or why in the world I don't get any / any reasonable interlaced output on S-video. RF unfortunately blurs the picture a lot and make it full of artefacts. Any idea is welcome. Can someone please run the PRGs on another real machine, I uploaded all PRGs + photos to Google drive. And unfortunately emulators also can't cope with this well, they will just show a flashy image. Try to compare HI (non-intl) and HI2 (interlaced) versions, these show the most obvious difference. [ link ]
And though it might be a native solution on Amiga, but on +4 this obviously requires tinkering. I have not yet checked the code of IstvanV, must be some trick with ff1d/ff1e at the end of the frame. I studied these tricks quite a lot 30+ yeras ago, but was many times uncertain what I really did or saw. IMHO we should dig a bit deeper, as it is truly amazing!!
Re: Is there such a thing as an interlaced picture for the +4? Wow!! This is awesome!! Really looks like some sort of next level gfx on TV. Who would have thought to see such colorful hires graphics from a "normal" 8 bit machine. More or less movie quality I would humbly say.. :)
Re: Is there such a thing as an interlaced picture for the +4? Love that space picture! Would make an awesome title screen for a new Thrust type game.
Amiga related below: On the subject of interlaced it's possible to run a native A1200 in 800x600 (interlaced), I used to do so around 1998. The only downside of that was you need a VGA adapter and an old school monitor like the Microvitec Multi sync CRT to display it. From memory the flickering wasn't very noticeable, I mean back then that screen mode was used for workbench, paint packages and going online with ibrowse so not a lot going on in terms of screen action lol.
BTW This very screen mode 800x600 laced does not display on normal PC VGA screens.
But yeah the unthinkable is sometimes possible with the right gear...
Re: Is there such a thing as an interlaced picture for the +4? Extra credit goes to whom can identify the series I got the picture from And @Mik we can team up for some space / inertia type game, I also loved this Thruster one
Can someone please try to run the Hires interlaced PRG on another +4 and see if can get picture over the video port (s-video or composite)? Would be also nice if @Gaia could make it work on YAPE... Too bad, should the effect's visibility be limited to a real +4 with RF
Re: Is there such a thing as an interlaced picture for the +4? I've tested your pictures on a real plus/4 with a RetroTink2X which is connected to a cheap Samsung monitor. For the hi-res numbered images only black display with either S-Video or Composite. I'm not sure if the RetroTink loses sync or my monitor is not able to display it. All the others are displayed correctly (but flickering as hell - as expected). My experience is that you need a 100Hz display (of course in PAL territory) to remove flickering of interlaced images (or some deinterlace electronics - called FlickerFixer in Amiga terminology).
For you have "good" picture via RF is probably because your TV does some deinterlacing or other image "enhancement" trick that you perceive as "good" picture output.
edit: I'm between moves so I don't have all my equipment with me. Normally I use a BenQ 15kHz capable monitor as the display for RetroTink which allows pass-through mode. This monitor requires the RetroTink to do a frame doubling to convert 576i signal to 576p. This can be a reason why there is no display for those images.
Re: Is there such a thing as an interlaced picture for the +4? Hi @siz, Yes those 2 numbered Hires ones are the interlaced ones, and seemingly we have the same result. :( I also used a Samsung UHD TV and an ODV converter (http://mcbazel.com/index.php/product/odv-composite-rca-s-video-ypbpr-to-hdmi-converter/). I don't think I understand this conclusion: "This monitor requires the RetroTink to do a frame doubling to convert 576i signal to 576p. This can be a reason why there is no display for those images." I am certainly less familiar with these tech than you, I would appreciate if you could test pass-through mode later on that BenQ monitor. Is that CRT or flat panel?
On a flat panel TV with RF, deinterlacing must be obviously done by the TV, I don't know why it looks great, wheteher it's due to enchancement, or it just ultimately displays the interlaced image properly like any normal 576i TV broadcast. (?)
But I think we can agree that the interlace looks real and the resolution is really double. I can post further pictures.
Re: Is there such a thing as an interlaced picture for the +4? In few weeks I may setup my C16 64K system, connected to a 25" CRT TV. It has side panels, and less digitial circuitry than the current flat monitors.
Larry's pictures are really goes into Amiga OCS territory. My previous 20"Dell monitor got crazy from those interlace pictures, but now my 32" Philips flat TV is allright.
@Charles Please do not forget, that the Plus/4's so-called S-video output is not really S-video compatible, just they separated the two signals from eachn other, but it's 1V is just too much for some of the TVs and monitors (it was created years before the S-Video standard). So you may need to add a 330 Ohm resistor to the luma and chroma to get the proper picture (it could explaint the blackc&white picture: your equipment cannot handle the signal as it far is beyond the standard)-
Re: Is there such a thing as an interlaced picture for the +4? @MMS, I know it's not fully S-video compliant, but don't really know the details or differences. You seem to be the expert here However composite video should be standard signal, should not it? I get the same result using that, no picture from these interlaced PRGs.. My HDMI converter works great most of the time, actually this is the only exception. And just to be fair this interlace is also a not-std feature on the +4. At the end I am looking for a way to use it reliably - this could greatly extend the capabilities of our beloved 8 bit computer - I think we will need to look at the code to see if we can finetune sg, altering HW is another league - at least for me. Please let us know once you managed to test it w CRT. Thanks
Re: Is there such a thing as an interlaced picture for the +4? @Charles Sorry I have noticed your additions to this topic only today. Your pictures are great. But IMHO you assumption that the IstvanV's great picture converter produces real interlaced pictures is rather not true. I checked the code, I am almost sure, it doesn't contain anything that can support the real interlace. Please check your ideas on simpler images, just try to produce 50-70 lines of texts - such things were standard for the Amiga 500... The Plus4 abundance of colors can easily be misleading. The flicker effect only gives us more colors, not more raster lines.
Re: Is there such a thing as an interlaced picture for the +4? Litwr wouldn't interlace apply only if it is possible to shift the signal by 0.5 pixels every odd 50hz frame? I mean in x this is possible (for multicolor) but it's not possible in y? Or am I just completely on a wrong track with this?
https://www.rigacci.org/docs/biblio/online/pal_video/ seems pal uses two fields of 312.5 pixels which add up to 625 full scanlines. On a C64 blog they wrote that the C64 just "ignores" the halve pixel offset and thus just displays 312 lines every 50hz frame (without the 0.5 pixel offset, thus no interlace possible).
Re: Is there such a thing as an interlaced picture for the +4? Sorry, I am not an interlace expert. However IMHO an interlaced picture is a normal picture for analog tv and monitors, we just need to generate a video signal which is similar to the standard signal for tv. You know that a lot of the super-popular Motorola 6845 variants have interlace mode not realized. This is true, for instance for the BBC Micro or Amstrad CPC. IMHO the Amiga was the first PC which could actually generate a proper signal for interlaced images. IMHO we need to analyze the Amiga video output, maybe we need to ask Amiga experts.
Re: Is there such a thing as an interlaced picture for the +4? Degauss once showed me that it is possible to alter the pal phase on Plus/4.. I think you know this trick, too, to get even more colors (around 200 then).. I am also no expert, but maybe it's possible on Plus/4 to modify the Pal signal in that way to display a real interlace picture.. I heard on several places that the Plus/4 is the only computer where you can alter every part of the Pal "protocol", maybe your guess about possible real interlace is true then, too! I think you are right with this! (You know more than me about this topic!)
edit: I think the crucial part could be to shift the video signal by 0.5 pixels each second frame.
Re: Is there such a thing as an interlaced picture for the +4? @Mad: Unfortunately, that is not how PAL signal generation works. PAL phase inversion will not help at all as it only affects the colour burst part of the signal. And that so called .5 pixel shift is encoded in the signal with the frames. Which is identified by the decoder circuit based on the structure of the signal received during vertical retrace. Now, TED generates that in a form which will force the display to always show the same (even?) half frame. This part is where my memory gets very hazy, but I think even and odd frames have the opposite phase of their V-Sync signals which TED cannot generate, AFAIK. The other problem, which would be a bit easier to remedy, is that a PAL half frame consists of 287.5 raster lines, not 288 as TED generates it. Yet another problem, if you are using things like HDMI converter, that TED's frame rate is not exactly 50 Hz, which may cause frame detection errors in some devices, resulting in skipped frames or no picture at all.
Re: Is there such a thing as an interlaced picture for the +4? Thanks for clarifying.. It's completely reasonable to show the same halve frame all day long. Seems like the shift depends on the frame type (even/odd), then? And this frame type (odd/even) is determined in the VSync? And Plus/4 always sends even,even,even..? Nice to know.. I really never thought to finally understand this part at some day.
Re: Is there such a thing as an interlaced picture for the +4? @mad The PAL color inversion is a very easy trick, just change line evenness in $FF1C-1D and get it. Plus4emu supports this - there was a minor error in the emulation but it works for most cases. Every line for a PAL picture has color information (burst) and for odd and even frames it is created in a different way. When we change the line evenness this forces the TED to think that it must generate the color signal for another evenness. @gerliczer Thank you. I have only slight doubts about the TED inability to generate a proper V-sync. We need exact information in what way even and odd PAL frames differ.
Re: Is there such a thing as an interlaced picture for the +4? > The PAL color inversion is a very easy trick, just change line evenness in $FF1C-1D and get it. Plus4emu supports this - there was a minor error in the emulation but it works for most cases. Every line for a PAL picture has color information (burst) and for odd and even frames it is created in a different way. When we change the line evenness this forces the TED to think that it must generate the color signal for another evenness.
Are you sure about this? I have to admit that I'm not that familiar with PAL standard but IMHO TED (and VIC too) generates the same frame twice for the odd and even PAL lines. $FF1C-1D just changes the line number within the current half-frame and not between frames. But that's just my opinion.
Re: Is there such a thing as an interlaced picture for the +4? I second Litwr. It is very easy to trigger the inversion inadvertently when you do raster line manipulation. I did that too during development of Sweet DFLI C16.
Re: Is there such a thing as an interlaced picture for the +4? Sorry I didn't want to change the topic to color phase inversion. I just wanted to say with my post above, if it is easy to invert the color phase maybe it's easy to change the ".5 pixel shift", too. Nice to know that it's just $ff1d to be placed on odd/even lines. I think I got this color inversion effect for myself by triggering $ff1d on the wrong place in an actual scanline (at least that is what I thought has happened). Good to know that's even more easy, maybe someone will take advantage of this effect some day.
edit: haha.. if you combine this effect with "flicker interlace".. You got like "even more colors than just 200", albeit flickering.. Maybe it's better to just alter the color phase line by line or even just if it's more appropiate on a line, without flickering..
Re: Is there such a thing as an interlaced picture for the +4? @siz The PAL inversion was tested on real hardware... It is sad that our small community misses a man who know PAL signals well.
Re: Is there such a thing as an interlaced picture for the +4? Nice to see this topic revived guys.
I definitely understand quite little of PAL picture generation, but I trust my eyes. The pictures I posted earlier in fact seem to have more distinct lines, it might be the result of some flickering removed by my TV's digital processor though...
"Color inversion" - if my understanding is right - is just when you shift the screen by an odd number, is not it? Then it's indeed a completely different story. Btw is there any real good / visible use of it?
Back to interlace: IstavnV's FLI routine has some interesting code in the raster interrupt part. It alters the horizontal scanning ($ff1e) a few lines before frame switching, and just upon frame switching (with quite a sick precision timing ) then a few lines later it does a line shifting upwards. I don't really get what it does and how / why it works, would be really great to test these examples on a CRT monitor, preferably an original Commodore one and have a close look at the results. Cannot someone do this here?
Re: Is there such a thing as an interlaced picture for the +4? For the color inversion, see Litwr's 231c technology demo. It's a really cool effect, and it could be used to display some logos in some "new" colors for example, but AFAIK it hasn't been used elsewhere.
(I really don't want to add any more noise to this, so I'll just very quietly mention that TLC's fantastic part in GOTU Megademo might have something similar. His scrolltext mentions the following: "Hogy lehetseges ezen a [...] compy-n az, hogy a kék szín fokozatok nélkül menjen át zöldbe??? a valasz röviden: így ..." Rough translation: "How is it possible to make this computer change blue into green without steps??? In short: like this". I think that part does some weird trick that might not be emulated on YAPE.)
Some write up to Pal. The last paragraph features the Pal hack "all" computers and game consoles use.. I was of course totally wrong by saying that a 0.5 pixel shift is needed, that's of course distinguished by the field, as stated already..
btw.. agree to Charles with seeing more resolution on these TV screenshots. Even there is this typical interlace x shift halve line after halve line (looks like missing deinterlacing (on moving pictures) or something), at least it seems so.. You can ?clearly? see that the pixels on these places have doubled the y resolution. Maybe it's this "It alters the horizontal scanning ($ff1e) a few lines before frame switching" for odd/even field select stuff (wild guess)???
edit: judging from the synchronization diagrams for the fields in the link above, it seems to be impossible to generate odd/even addressing via ff1e tricks.. :( Maybe I am wrong on that, but it really looks like it's impossible..
Are the around 8 pal vsynch rasterlines also "rooted" in the ted and is it thus modifiable by registers on the right rasterlines? If yes, maybe it's doable to change the refresh behaviour [odd/even fields] (with ff1e/ff1d).. Or??
Re: Is there such a thing as an interlaced picture for the +4? Just to point out what Csabo did refer to, this is a comparative analysis of the TLC's part in GOTU Megademo.
My reaction has been like: oh cooome ooonnn... plus4emu is capable to emulate even the artifact colors in the horizontal borders of the giant letters, whereas YAPE follows by far far away. After 15', the ETC's FRQ converter started to play the same track on both channels...oh well, at least now we can count him into the demo's credits aswell
Re: Is there such a thing as an interlaced picture for the +4? Let me present a small program that shows the PAL color inversion effect. It shows also that this effect is quite easy to get.
@mad Thanks for the link. I checked the Istavan's interrupt routine again. It seems that it does something that could turn PAL interlace on but it does this not the same way as it is documented on the PAL video timing specification site. The Istvan's routine does all $ff1e manipulation in line 1 and this is difficult to match with information on the site. Moreover the routine is the same for odd and even frames and it doesn't add/remove any delay. So I doubt very much about the routine ability to produce a true interlace picture. However IMHO it may be real to program the TED to generate the standard PAL interlace signal: for odd frames, we must push the TED to skip one short sync at 311, then set the line number to 1 and then again skip one short sync... IMHO there might be other variants to tern interlace on - we need more information.
It would be great if IstvanV could join our discussion. IMHO he is the best PAL signal expert among us. He can also completely clarify details about his great picture converter.
@Charles It would be very interesting to get something like this Amiga 500 screenshot. Nobody doubts that there are more than 400 raster lines.
Re: Is there such a thing as an interlaced picture for the +4? @Litwr, All,
For me the biggest question is the testing method / test setup. Emulators except for IstvanV's plus4emu are out of league for interlace testing. My home TV setup with the s-video/composite to HDMI converter seemingly won't display these interlaced modes. In RF works, but with all the artefacts and some blur making it difficult to properly judge. A CRT monitor would be desired for the testing, preferably a Commodore one. DOES NOONE HERE HAVE ONE who would help us with taking some screenshots / short video?
Re IstvanV's code:
definitely there is difference between odd and even frames - you are wrong in this regard -, it has the "normal" FliIRQ at each frame (at raster line $03 $36), but InterlaceIRQ 1-2-3 only every second frame (at raster lines $02 $f9, $02 $ff, $03 $05 respectively).
So _every second_ frame at: - $02 $f9 InterlaceIRQ in line $02 $fa after some precise timing puts $6d into $ff1e instead of $1c (I've replaced sta $ff1e w lda $ff1e to get the reading) - $02 $ff IntlerlaceIRQ2 in line $03 $00 (First line of new frame!) puts $ad into $ff1e instead of $c4 - $03 $05 IntlerlaceIRQ3 still in line $03 $05 pulls the screen up by 1 pixel line (dec $ff1d) btw, this also does the color inversion in principle - it might just be there for that purpose (?)
At _every frame_ at - $03 $36 (just before the start of visible screen area) fliIRQ routine displays FLI image (plays with $ff1d,$ff1e,$ff1f as well but imho this has nothing to do with the interlace)
Not sure what the $ff1e trick really does - I am not familiar with using it - but if its behavior is similar to $ff1d and if its effect spans over multiple lines, comparing to the table shared by @Mad it suggests to me it can bring those 5 long syncs forward, to initiate an interlaced Frame 2. Does anyone here have better understanding / experience with horizontal scanning effects?
Indeed this implementation might not be 100% complaint with the PAL specs, or I don't exactly know why my HDMI converter gives up trying to display it, might be what @gerliczer advised about (not proper 50 Hz any longer) The routine btw has different timings for NTSC, it's a true masterpiece of a crazy mind. I wrote a mail to @IstvanV asking for some help / clarification but he did not respond yet..
Re: Is there such a thing as an interlaced picture for the +4? @Charles Thank you. You are right, there are different handlers for different frames, I made an embarrassing mistake. So I am now persuaded that the @IstvanV 's converter actually makes interlaced pictures. I just must more thorough analyze his code. I have got the next picture from the converter and this picture definitely has more than 400 lines. BTW this converter can be improved in one way. It misses pixel by pixel mode, it always keeps aspect ratio instead. For example, I have an image 320x400 - it can be perfectly fitted into the plus4 screen but p4fliconv re-scales it and makes it 160x400.
Re: Is there such a thing as an interlaced picture for the +4? @Litwr, @Charles: So IstvanV really seems to be the first guy to double the resolution on Plus/4.. Maybe no computer from that era had "native" 320x400 or even more resolution.. Really crazy!
Re: Is there such a thing as an interlaced picture for the +4? @Mad It seems that Gaia was right, please check his comment at the top; he mentions that TLC was the first one to do this (as with several other things ).
Re: Is there such a thing as an interlaced picture for the +4? Oh my.. Sorry I just didn't read that comment before.. props to TLC! Gaia even wrote that p4fliconv does it right already!
Re: Is there such a thing as an interlaced picture for the +4? So there is one ultimate question, if we can improve this to work on more systems / setups. 1. I still don't exactly know why the HDMI converters fail to display these interlaced images and if we can fix this 2. Could YAPE also handle it /display it properly? @Gaia, please..?
If we can fix these we should just start using interlace more, costs barely no rastertime and could be used not only for FLI but all kinds of applications. Can someone possibly draw a 8x16 hires charset by chance? Even lines to be stored in one char and odd lines in another (at diff offset)
Does noone here understand what exactly the $ff1e adjustment - revealed in my prev post - does?
Re: Is there such a thing as an interlaced picture for the +4? Great idea with the charset.. Maybe this could also be implemented in someway into a kernal, would need another $800 then though..
Re: Is there such a thing as an interlaced picture for the +4? @Charles: I think, I understand what those Horizontal Raster Position Register writes do. What I don't understand is why you don't understand it. You yourself linked the image of how the PAL frame is structured. Please compare the number of short and long synchronization pulses between the half frames and read all the horizontal timing legends. Row widths and timing values do scale in the image.
On a different note, those $02 $f9 or $03 $05 and other such notations baffled me. I think, you should have written raster line $0f9, $105 and so on. Mixing the Raster IRQ Enable Bit into the values did not help understanding too much. Neither did Big Endian notation.
Re: Is there such a thing as an interlaced picture for the +4? @Charles >1. I still don't exactly know why the HDMI converters fail to display these interlaced images and if we can fix this IMHO this means that the signals are not 100% correct, so we can improve something. I want to do this but I need an oscilloscope...
Re: Is there such a thing as an interlaced picture for the +4? Maybe @gerliczer can help with fixing the routine Saturday at 4 am I was glad to be able to extract something which makes sense to bring the discussion forward. But you are right, it is ways more appropriate to use $0f9 vs $02 $f9 :) I think raster line $100 = 313. PAL line (for second frame), can we agree on that?
I am not afraid of admitting what I don't understand or not experienced with. As I wrote I suspect this is indeed "bringing those 5 long syncs forward" but I don't fully understand what a short sync and a long sync truly mean, how these are generated by TED, and how these can be altered by writing to horizontal scanning. I'm quite experienced w vert sync tricks, but not yet with the horizontal counterpart.
I would appreciate if you explained what writing $6d @ pos $1c and then later $ad @ pos $c4 is ($ff1e) supposed to do. "Shifting" is into the same direction. My _guess_ is the first (6 lines before the new frame) brings the long sync fwd and the second (first line of new frame) does the same w a short sync, but timing wise these are hieroglyphs to me. And most of all do you have any idea how to improve to make it more PAL std compliant?
@litwr, I wish I could help with an oscilloscope, but I can help with testing. Next week if I can again find some time I will do some changes to the timing to see what happens. But again would be nice if someone also tested the effect on a CRT monitor. Most hdmi converters today use line doubling (“bob de-interlacing”) so the output would lack the additional vertical resolution anyway.
Re: Is there such a thing as an interlaced picture for the +4? It would be really great if somebody could publish plus4 video signal from oscilloscope for two frames. The same information from Amiga interlaced picture would also help a lot. IMHO the Amiga interlaced picture works everywhere with every monitor... The Amiga interlaced pictures are much more stable that those from the IstvanV's converter...
Re: Is there such a thing as an interlaced picture for the +4? @Litwr "IMHO the Amiga interlaced picture works everywhere with every monitor..." Not exactly. I remember reading on Amiga forums, that they had VGA converters, but only a handful of monitors supported this special frequencies used by the interlace pictures. Certainly they could buy extra scandoublers, but only Multisync monitors were able to properly handle the Amiga's picture.
Amiga guys highly praised the NEC Multisync monitors, also most of the Sony special color Trinitron monitors (security, broadcaster, etc) had 450-500 line resolutions, and could easily show any special signals came on composite input. (VHS resolution was approx 240 lines, SVHS was 400 lines, so 450-500 line was pretty awesome)
Re: Is there such a thing as an interlaced picture for the +4? Hi @MMS, I am getting a bit confused, did not all PAL CRT sets have 576 (visible) lines and should not they all be able to display 576i picture? Is not this more about tolerance against deviations from the standard PAL signal..? You consistently write monitor, but if not talking about TV sets, I wonder if these can be found anywhere Today. (You can pretty much only find PC monitors on the market) I consider buying a CRT TV for testing, but finding one with s-video input or Scart which works properly with an svideo-scart cable seems to be a challenge at this point.
Re: Is there such a thing as an interlaced picture for the +4? Actually, almost all Philips TVs (at least the medium ranges and up) had a side panel contains the AV and S-Video connectors, and all of these TVs had a scart socket with Y/C signals connected. (upmarket TVs had even the R,B,G all connected, typically called Matchline).
Usually you can find the "leaflets" of them on Philips page. (on the second page at Connectivity you may find all important information). If you find a nice one somewhere, you can check the technical details in advance. One example, I think we produced it in Hungary: https://www.p4c.philips.com/files/2/29pt8640_12/29pt8640_12_pss_.pdf
The basic, simple TV models sometimes missed these features, had only composite (CVBS) on scart and on the side panel too.
Re: Is there such a thing as an interlaced picture for the +4? Actually I still check this forum every couple of months. But it seems I had not signed into my account for years.
It appears I still have some source code form the 400x400 pixels demo part I worked on in 2005. You can also see it in the screenshots above.
Manipulating the PAL signal should be easier today than it was ever before. You can get fairly cheap scopes that can capture an entire PAL frame at once to verify all the timing. I implemented the interlace code with an analog scope.
Re: Is there such a thing as an interlaced picture for the +4? Bionic finally, it's so great you come back here, hope to meet you there more frequently Having this thread rich of infos and tech details, it's probably time to release a little techdemo to show off that stuff on real harware
Re: Is there such a thing as an interlaced picture for the +4? Hi Bionic,
Welcome back and Thank You for sharing the routine! It's quite amazing that you made it in 2005.. The code looks very similar to IstvanV's FLI interlace routine, though using slightly different timing. Have you played with multiple different timings by chance? I wonder as I cannot get picture using an HDMI converter using your routine either (signal/freq is out of range?). It also does not seem to work properly on any of Today's +4 emulators. Let me ask was that a CRT monitor you used for your screenshot, or should it be LCD, what model is it? However on RF this (also) gives solid picture on my TV. Sending two photos, I made the effect a bit more visible should someone still have doubts (the first one shows an alternative screen in the odd lines (orange bg, blue text), the second uses only different FF07 value)
Re: Is there such a thing as an interlaced picture for the +4? Wow! This thing really exists! It would be fantastic to get 72 lines of texts... It is sad that my C16 has become non-functional recently and emulators can't help. :(
Re: Is there such a thing as an interlaced picture for the +4? >The code looks very similar to IstvanV's FLI interlace routine, though using slightly different timing. Have you played with multiple different timings by chance? I wonder as I cannot get picture using an HDMI converter using your routine either (signal/freq is out of range?). It also does not seem to work properly on any of Today's +4 emulators. sad Let me ask was that a CRT monitor you used for your screenshot, or should it be LCD, what model is it?
I used an analog scope to replicate the PAL timing. I actually found a photobelow. Using a modern DSO would be much better today.
The monitor is a Dell TFT with composite input. It's possible that it is more tolerant of timing than other converters.
The best solution would really be to make sure that the timing is correct and not adjust it to a single monitor. Maybe it's also necessary to allow for some options to adjust the timing, in case there is still variation between different monitors.
Re: Is there such a thing as an interlaced picture for the +4? @Bionic: thank you for the great news! It looks great!
OFF: Please let me ask one personal question: how you made your Dell monitor accepting the Commodore signals? My Dell 2007FP Ultasharp is simply does not accept any of them ()has composite and Y/C inputs too. Did you made some hacks, like adding extra resistor to video signals) (I tried a 270 Ohm, but no success, still only black screen) Or your monitor accept the signals as they are? Thank you for your answer!
Re: Is there such a thing as an interlaced picture for the +4? OFF: Just some silly maybe useless idea..
In the past (around 1996) I was forced to invent a new graphics mode for an upcoming game. It was Tunnel B1 by Neon Software.. https://www.youtube.com/watch?v=Q6wOnNVpaVk..
The thing is the graphics back then didn't support Hicolor (16 bit = 65536 colors). There was a boss who said, "you invent something or I don't let you out of your working room"... So I was forced to be in this room for 2 weeks..
I came up with the idea to double the resolution of a normal 256 color VGA signal by the so called ModeX. So you got a 320x400 picture.. And choose to display 6 bit of green in one line and 3.5 bit of blue and 4 bit of red in every other line, so the color information was spread over two lines. I mean we got this PAL phase inversion somehow.. Maybe it would make sense to build a converter featuring real interlace + PAL phase inversion on every other line? Not sure of this, since I am certainly not the one to build a converter for this mode..
But if it worked for a commercial game in the 90s maybe it is also not detestable to the viewers today.. (A german mag titled that "we wanted to spare performance" by just rendering every second line, though)
Re: Is there such a thing as an interlaced picture for the +4? @Mad: Problem is that what you did in Mode-X has nothing to do with PAL phase inversion. PPhI gives you artefact colours that are not part of the normal PAL palette of TED. AFAIK, they are falling into the opposite positions of the standard colours in a colour wheel. I think they don't match any of those normal ones. Which may not be too much of a problem for a converter for static screen content, but you are probably badly screwed with anything dynamic.
Re: Is there such a thing as an interlaced picture for the +4? I just meant that "we" may use the 400 lines of the two PAL fields. One with the one 121 colors and the second with the other "new 121" colors. So every other line of the 400 we got another "palette". Thus increasing color count. So to say 121 colors mixed with another "pal phase 121" colors. It was just an idea..
Like going from 256 colors to around 12000 using Mode-X in VGA. Just now going from 121 colors to artificial 20000 or even more (or less :)).. So to say flicker interlace without flickering. Maybe the visible line raster is not so much of a problem, like on VGA.
What's the problem with dynamic content if you use 25 fps? I was just talking about still pictures, maybe that was not clear.. Anyways, it was just an idea. Maybe phase inversion is not needed at all for this, though..
Let's see what the pros will do with this double resolution modes later..
Re: Is there such a thing as an interlaced picture for the +4? Hi Mad/MMS,
Tunnel B1 rings a bell - wasn't Mr. Pet involved in this? Also, I certainly remember the screenmode from Alto Knallo. It was a nice hack at it's time. Today, people will probably guess it is meant to achieve something other than simply increasing the number of colors
The monitor is a Dell 2001FP, afair, and I just used a Y/C cable. No special tricks. Was a great monitor back then.
I guess apart from increasing the number of lines, one could do many colour tricks by manipulating the PAL signal. Phase inversion surely is an option, I recall that Graham also managed to do that on a C128.
Also, I wonder whether there is any way to fix the color signal when using NTSC timing to improve the horizontal resolution. As long as the burst is correct, everything should be interpreted as a PAL signal. So it's mainly a question of generating something meaningful.
Re: Is there such a thing as an interlaced picture for the +4? Yes Mr.Pet did the "Playstation emulation" for the PC, he did the ultra fast polygonroutines there.. If I remember right these where his first works on PC, even heavily Pentium optimized stuff. I worked in several companies with him. Basically I had to improve the visuality of his routines back then, he did just some realtime floyd steinberg 320x200 256 color dithering, which looked not presentable, albeit beeing a cool idea at that time point. If I remember right he had 3 or 4 versions to get from HiColor to 320x200 256, which all looked not so well for the normal viewer. The screenmode in Alto Knallo is actually "my invention". The guy who caged me for two weeks later claimed that the screenmode was his idea, what was very embarassing to me (It wasn't Mr.Pet). But now there is a lot of grass grown over that part.. There was a demo which I found out later, I think it was from Orange which used almost the same trick using 3 lines for RGB. However I think the 2 lines where bit better optically.
If you got results with the PAL signal I am really interessted, it's somehow a possibility no one whould have thought of back then.. That all sounds a bit like that there is actually not a real boundary when it comes to PAL signal manipulations.. You guys rule!
Re: Is there such a thing as an interlaced picture for the +4? @Mad, >> I just meant that "we" may use the 400 lines of the two PAL fields. One with the one 121 colors and the second with the other "new 121" colors. So every other line of the 400 we got another "palette". Thus increasing color count. So to say 121 colors mixed with another "pal phase 121" colors. It was just an idea..
Just for the record, IstvanV's FLI routine is seemingly just doing this. In every second PAL field he also does a single line shifting, which produces this "color inversion" effect, altering the color palette.
@Bionic, should we have a digital oscilloscope, would you have time/will to improve the timings of your routine by chance? If only the lack of equipment is hindering us, I am glad to donate to this noble cause
Re: Is there such a thing as an interlaced picture for the +4? @charles
Actually i got a nice DSO, so that is not limiting. It's rather lack of time, I Guess. The interlace example should also work fine as it is above, it's more about adjusting to different monitors, maybe.
Seems tempting to do a bit more PAL hacking, though Let's see whether I get around to it someday.
Re: Is there such a thing as an interlaced picture for the +4? Hi @Bionic,
Yep, you don't even need to explain that to me.. (time constraints) But we all would appreciate here if you could allocate a bit of time to further testing. In principle the routine works fine indeed, it's perfect on my TV w RF, it's only the HDMI converter which fails to display anything at all when interlace mode is on. I've just bought a Dell monitor w S-video input but it struggles to display the standard video signal of the Plus/4 as well, so will look further. And the other target is to make an emulator - preferably YAPE - to be able to display this, that would open the door for more applications. As per my tests not even IstvanV's plus4emu can display the picture generated by his own interlace routine properly.
I've compared the 2 interlace routines, they are quite different timing wise. However they seemingly achieve the same result (proper interlace on RF) but no image at all w HDMI converter. (I'm still looking for a usable screen w S-video input)
Bionic interlace routine (every second frame) at line $0fa: $FF1E $30 -> $54 at line $100: $FF1E $c4 -> $ca
IstvanV's interlace routine (every second frame) at line $0fa: $FF1E $1c -> $6d (InterlaceIRQ at $0f9) at line $100: $FF1E $c4 -> $ad (IntlerlaceIRQ2 at $0ff)
Do these timing differences make any sense to You? Do you think you can adjust it further if it is not yet fully compatible w PAL standard?
ALL, please find the source code and executable PRG of both routines following this link. It would be great if you tried on different screens using the Video port of the Plus/4. Thank You
Is there such a thing as an interlaced picture for the +4?
I remember that his intentions were to keep up the work about his 400x400 ilace stuff, continuing to follow the initial path, and he asked for a very flat and simple 72x128 picture in 3 color only, probably to be placed touching the borders, so I did this girl's bust. Later, he disappeared from the scene and nothing tangible followed :( 
IMHO we should dig a bit deeper, as it is truly amazing!!
Saturday at 4 am I was glad to be able to extract something which makes sense to bring the discussion forward. But you are right, it is ways more appropriate to use $0f9 vs $02 $f9 :) I think raster line $100 = 313. PAL line (for second frame), can we agree on that?
