Ym2413a said...
In a perfect world, every country would agree on the same video standard.
In the real world, every country can't even agree on anything. *lol*
Thank heavens for HDMI, the new messiah! (well at least until the next one comes along!)
@ Baggers, Coley : Excellent work. Looks fine on my Panasonic 16:9 which auto-selects
PAL/NTSC. Doesn't work on my Citizen LCD handheld but that's PAL only, so didn't
expect it to work.
Colours are good and solid so the core problem seems to have been solved / overcome.
I note the display width is ( appears ) considerably less than when running the out-the-box
library Graphics_Demo in NTSC, ie, the big number is twice as far in from the left of the
screen for PAL than it is for NTSC. I don't know if that's because of the driver, mode, or
my TV making a 'best guess'; it unfortunately doesn't report what it's doing.
On my 28" diagonal screen, in NTSC width between left of the number to right of the
Propeller logo is 18", in PAL it's 14", 22% less.
I assume the same PAL60 trick can be applied to TV_Text ? If so, I can live with less rows
than PAL gave but not having a full width text display would not be so acceptable if that is the
case.
I guess the next thing should be PAL60 TV_Text, then 8x8 font drivers. That should overcome
any problem of less vertical lines. Superb work so far.
@ Coley : Untouched and the sizing is fine there but it has really severe strobing on my TV.
Changing all the palette colours in TV_Text to "byte $9e,$0a", yellow on dark blue and it
appears to be rippling left-right on alternating lines, alternating frames, or something like
that. From a medium distance the accents on the characters appear to bounce up and down
but that's an optical illusion.
Don't beat yourself up too much as it may be my TV, it'll be interesting to hear other reports.
I can't see any evidence of strobing on the PAL60 graphics demo.
I've got some PAL testcards which I use ( mainly for checking VCD cropping ) which it may
be possible to convert to image display. They are 352 x 288 GIF's, and I have no idea how
to convert for Propeller use ( time to go and Search ! ). What pixel image size would give
full-width, full-height ? I'm happy to create a similar test card from scratch as that's the
best way to get rid of artefacts caused by resizing.
Ym2413a said...
In a perfect world, every country would agree on the same video standard.
In the real world, every country can't even agree on anything. *lol*
Thank heavens for HDMI, the new messiah! (well at least until the next one comes along!)
OK, I'm about to show my utter ignorance of the latest TV standards (which is true, I know virtually nothing) but I thought I read somewhere that HDMI was encrypted so that people could not tap the line and "pirate" (i.e. record for later playback) the signal. True? If so, that could throw a major monkey wrench into hobbyist use. Hopefully I'm wrong.
You know, we in the USA can benefit from this too. We might ought to use PAL for some of our Propeller projects. The reason being that most digital LCD displays support PAL and NTSC, and in certain situations maybe the PAL display might come out crisper. For instance, my TV to VGA converter suffers pixel crawl with NTSC signals, but PAL might not.
Edit: Yeah Beanie, HD has me worried too, because it's like we finally have an "easy" way to generate PAL or NTSC in our projects, and now the world is switching to a new TV standard that will be two orders of magnitude harder to generate a signal for!!
Post Edited (Dennis Ferron) : 1/7/2008 4:00:08 AM GMT
I will have another look at the driver when I finish work today, there is a lot of minute control available, Parallax did a great job with that driver!
Hopefully we can tweak it to work on your TV.
@ deSilva
It's true that not all PAL TV's can display PAL60, but here in the UK the majority do.
I have tried this output on all my TV's in the house, one is over 12 years old and they all worked fine.
Whilst I am happier that I can now display a quality image on my TV, it is not acceptable for a commercial product to be limited in this way.
We have to aim for an acceptable level of quality for ALL PAL TV's and that will have to be 50Hz, so this is not over by a long way.....
I've shared your concerns about the quality of the PAL output for some time, hopefully we can find a solution to it.
@Ariba
You are most welcome Andy!
@Beanie2K
It's true that there is an encrypton routine available in HDMI called HDCP (High Definition Content Protection). as far as I know encryption is optional.
To clear away one element that can be confusing: DVI and HDMI are exactly the same as one another, image-quality-wise.
The principal differences are that HDMI carries audio as well as video, and uses a different type of connector, but both use the same
encoding scheme, and that's why a DVI source can be connected to an HDMI monitor, or vice versa, with a DVI/HDMI cable, with
no intervening converter box.
Hippy, thanks [noparse]:)[/noparse] if you can do a 256x192 version, I'll happily do it on a prop for you.
Dennis, let us know how your tests go.
Ariba, thanks for the feedback, and you're most welcome.
Beanie2K, like Coley said, HDMI is CP optional, so I don't think it'll affect us, IF we can get an image out fast enough for HDMI when composite goes, but I think we still have some years left, as most TV's still have composite [noparse]:)[/noparse] well in UK anyway.
I tried the pin 20 version of the demo on a portable DVD player that has a PAL 16:9/4:3 switchable display and a auto-multistandard 29 inch Sony Trinitron 4:3 TV.
As deSilva predicted, the DVD player display didn't work too well because of the 60Hz vertical timing, about 3/4 of the display is pushed up with the rest wrapping around the bottom. The graphics and colors look great though. To see a similar display showing similar results (60Hz NTSC) on this PAL50 display see my picture Mpe07a (Small).jpg in this post:
Thanks for the feedback, glad the colours came through, even if the timing didn't with it being PAL60, we're working on a PAL50, though so fingers crossed [noparse]:)[/noparse]
This is a really great attempt and much qudos for getting this far !!! A testcard is created to test (kind of obvious) all of the aspects within the picture actually mean something.
I have decided to use an external logic circuit that I can certainly be sure will generate a known PAL output, That graphic image will not be allowed to use however. The Propeller is still the back end processor for all of this.
Baggers said...
Hippy, thanks [noparse]:)[/noparse] if you can do a 256x192 version, I'll happily do it on a prop for you.
Many thanks. this is a bit late but everything should be 'pixel perfect'. Not sure about the aspect ratio so squares may not be square, circles not round, but it's usually good enough to show strobing, ghosting and some other undesirable effects.
@ Baggers : Whoot ! You wouldn't happen to have a .binary so I can see how well it
display on my / other TV's ? The great advantage of any testcard is that its a common
reference if nothing else.
I allowed myself to get more than a little carried away with my somewhat harsh and critical comments. I was wrong for doing so. You have pushed the boundaries of what
this device is actually able to do already. I feel sure that Parallax will let you have a play with Prop II before they want to market it as a consumer product.
If you could just mention would be so great to get real UK PAL out of the device, and not a kludge on some NTSC standard. I shall also expect full SECAM - would be tripple
standard then.
@Hippy
Wonder if a EBU style card would be of more use. I believe that anything with a circle will become squashed. Anyway we all know is just for fun and has no technical merit attached to the image displayed.
Hi Mike_GTN,
No worries, I totally understood what you were getting it, it's great progress, but as you use the test cards for professional usage, it would need to be VERY crisp especially for broadcasting etc., but for use for home entertainment / games, it's fine.
So on that point I agree with you completely, which is why we did PropGFX, where the colours are very clear, but at the cost of extra hardware in the form of an AD724 along side the propeller.
and YES, Parallax, any of your employees, if you're reading this, then I'd be honoured to be let to have a play with Prop II before they go to market. tbh, can't wait to get my hands on it lol if the truth be known [noparse]:)[/noparse]
PS, most of my earlier photo's were 16:9, and they look better 4:3, which is what I have my tv set at now [noparse]:)[/noparse] so they don't look as squashed.
@ Baggers : Amazing. Rendered better on-screen than I've seen some VCD/DVD players achieve.
There's some very slight shimmy left-to-right but certainly nothing I'd call strobing and I would
put what there is down to the TV, even TeleText shows slight shimmy. There's some diagonal striping
across the left most frequency bars ( one pixel apart ). That's expected and I see it on all DVD's
I've tested.
From a distance I'd call it "rock solid". Up close it does look like alternate lines are shifted across by
one pixel or so.
So the question must be, why is TV_Text so bad on my TV when this and the first graphics demo are
okay ? Either it's some odd interaction or there's a subtle timing difference I guess.
@ Mike : Not sure about the benefits of the EBU testcard ( just colour bars if I remember right ). The
Philips versions or IBA ETP would probably be better but only really if one has the knowledge or test
equipment to use it. Something to check pin-cushioning and set colour and contrast is all that's really
needed by a home user. Circles should be round not squashed, but it takes some effort to get the
digital image so it appears on screen as it should be rendered. Cropping in the VCD burning chain was
the original motivation and I didn't want to use a copyright image. What I have is a home-brew cross
between BBC J and IBA ETP. There's quite a science to proper test card design and I'm no expert !
@Hippy,
could you perhaps check this code with the PAL60 driver? As it does not work on my PAL screens I should like a feedback wrt to the quality... The difference being that it works in non- interlaced mode...
@Hippy : EBU was a good one as removed anything other than colour bars. Would see no other issues
then and was a serious statement as would show how good the PAL colours were. Circles that look like
circles, I don't see this happening. I don't put this down to your positioning of the original either. I will
be guessing that the image has been compressed further to get this coming out of the Propeller.
I wanted to simply caution people against whipping the backs off their TV sets and messing about with
the controls thinking this is a suitable instrument to drive some patterns. It is a good example (me even
impressed) it is however not a technical instrument. Nobody here has said that it is - but I think we would
all be amazed at what some people would do!
@Baggers - I've now gone in to Testcard Retirement. It's safe to come out!
@ Mike : Okay; I misunderstood what you were saying. Yes, for colour, EBU is better. On my
proper VCD cards circles come out okay. They have to be distorted so when the TV dot-size
takes effect they go back to being round but I've managed to get round circles on both 4:3
and 16:9. It is hard work though.
And you're right ... please people don't take the backs off your TV and start twiddling ! Not
only is it dangerous, but what may look pretty as a testcard isn't meaningful unless the card
is calibrated, and this certainly isn't.
@ deSilva : About to test your code ...
Out the box, so presumably running with Parallax TV.spin, it worked okay. Mikado draw Lines are extremely badly strobing, both yellow and black lines ( shaking rapidly left to right ). 'Menu text' is rock-solid. Text test, same shaking as with Mikado lines.
Using the PAL60 TV.spin, the height changes drastically( to be expected, top of the 2.1 of "Vers 2.1" touches the top of the screen ). Mikado lines were solid, no strobing / shaking. Text test no strobing / shaking.
And an amazing coincidence / rare sight ... BBC2 transmitted Testcard J for a few minutes at 03:00 along with a digitised speaking clock. So at least I now have the feel for what the TV does on a genuine reference ! It did suffer some strobing on that so I guess the TV is inherently susceptible to it.
Comments
Thank heavens for HDMI, the new messiah! (well at least until the next one comes along!)
Post Edited (Coley) : 1/6/2008 10:37:35 PM GMT
PAL/NTSC. Doesn't work on my Citizen LCD handheld but that's PAL only, so didn't
expect it to work.
Colours are good and solid so the core problem seems to have been solved / overcome.
I note the display width is ( appears ) considerably less than when running the out-the-box
library Graphics_Demo in NTSC, ie, the big number is twice as far in from the left of the
screen for PAL than it is for NTSC. I don't know if that's because of the driver, mode, or
my TV making a 'best guess'; it unfortunately doesn't report what it's doing.
On my 28" diagonal screen, in NTSC width between left of the number to right of the
Propeller logo is 18", in PAL it's 14", 22% less.
I assume the same PAL60 trick can be applied to TV_Text ? If so, I can live with less rows
than PAL gave but not having a full width text display would not be so acceptable if that is the
case.
I guess the next thing should be PAL60 TV_Text, then 8x8 font drivers. That should overcome
any problem of less vertical lines. Superb work so far.
Post Edited (hippy) : 1/6/2008 10:51:32 PM GMT
Don't worry about the lack of width, we've probably altered the 'hx' horizontal expand.
You can lower it back to the default one in the graphics demo.
The reason I did that, was to cover more screen, leaving less of a border [noparse];)[/noparse]
Cheers again,
Baggers.
give this a try, it works fine of my widescreen TV, vertical and horizontal offsets might need tweaking a bit.
Coley
Changing all the palette colours in TV_Text to "byte $9e,$0a", yellow on dark blue and it
appears to be rippling left-right on alternating lines, alternating frames, or something like
that. From a medium distance the accents on the characters appear to bounce up and down
but that's an optical illusion.
Don't beat yourself up too much as it may be my TV, it'll be interesting to hear other reports.
I can't see any evidence of strobing on the PAL60 graphics demo.
I've got some PAL testcards which I use ( mainly for checking VCD cropping ) which it may
be possible to convert to image display. They are 352 x 288 GIF's, and I have no idea how
to convert for Propeller use ( time to go and Search ! ). What pixel image size would give
full-width, full-height ? I'm happy to create a similar test card from scratch as that's the
best way to get rid of artefacts caused by resizing.
OK, I'm about to show my utter ignorance of the latest TV standards (which is true, I know virtually nothing) but I thought I read somewhere that HDMI was encrypted so that people could not tap the line and "pirate" (i.e. record for later playback) the signal. True? If so, that could throw a major monkey wrench into hobbyist use. Hopefully I'm wrong.
Edit: Yeah Beanie, HD has me worried too, because it's like we finally have an "easy" way to generate PAL or NTSC in our projects, and now the world is switching to a new TV standard that will be two orders of magnitude harder to generate a signal for!!
Post Edited (Dennis Ferron) : 1/7/2008 4:00:08 AM GMT
It is more or less coincidence when this modification works fine on many "high end" TV sets...
Now I can use much more color combinations without flickering.
Thank you Baggers and Coley
Andy
I will have another look at the driver when I finish work today, there is a lot of minute control available, Parallax did a great job with that driver!
Hopefully we can tweak it to work on your TV.
@ deSilva
It's true that not all PAL TV's can display PAL60, but here in the UK the majority do.
I have tried this output on all my TV's in the house, one is over 12 years old and they all worked fine.
Whilst I am happier that I can now display a quality image on my TV, it is not acceptable for a commercial product to be limited in this way.
We have to aim for an acceptable level of quality for ALL PAL TV's and that will have to be 50Hz, so this is not over by a long way.....
I've shared your concerns about the quality of the PAL output for some time, hopefully we can find a solution to it.
@Ariba
You are most welcome Andy!
@Beanie2K
It's true that there is an encrypton routine available in HDMI called HDCP (High Definition Content Protection). as far as I know encryption is optional.
To clear away one element that can be confusing: DVI and HDMI are exactly the same as one another, image-quality-wise.
The principal differences are that HDMI carries audio as well as video, and uses a different type of connector, but both use the same
encoding scheme, and that's why a DVI source can be connected to an HDMI monitor, or vice versa, with a DVI/HDMI cable, with
no intervening converter box.
There is a simple FAQ here www.octavainc.com/HDMI%20deciphered.htm
Regards,
Coley
Dennis, let us know how your tests go.
Ariba, thanks for the feedback, and you're most welcome.
Beanie2K, like Coley said, HDMI is CP optional, so I don't think it'll affect us, IF we can get an image out fast enough for HDMI when composite goes, but I think we still have some years left, as most TV's still have composite [noparse]:)[/noparse] well in UK anyway.
You might be able to use something like :- www.oodletuz.fsnet.co.uk/soft/tcmaker.htm to design your required
cards.
Mike.
I tried the pin 20 version of the demo on a portable DVD player that has a PAL 16:9/4:3 switchable display and a auto-multistandard 29 inch Sony Trinitron 4:3 TV.
As deSilva predicted, the DVD player display didn't work too well because of the 60Hz vertical timing, about 3/4 of the display is pushed up with the rest wrapping around the bottom. The graphics and colors look great though. To see a similar display showing similar results (60Hz NTSC) on this PAL50 display see my picture Mpe07a (Small).jpg in this post:
http://forums.parallax.com/showthread.php?p=699285
It looks great on the Sony TV.
Nice work, would like to see source available with PAL50 Hz though.
Thanks!
David
Thanks for the feedback, glad the colours came through, even if the timing didn't with it being PAL60, we're working on a PAL50, though so fingers crossed [noparse]:)[/noparse]
Cheers,
Jim.
Baggers.
This is a really great attempt and much qudos for getting this far !!! A testcard is created to test (kind of obvious) all of the aspects within the picture actually mean something.
I have decided to use an external logic circuit that I can certainly be sure will generate a known PAL output, That graphic image will not be allowed to use however. The Propeller is still the back end processor for all of this.
Mike.
Many thanks. this is a bit late but everything should be 'pixel perfect'. Not sure about the aspect ratio so squares may not be square, circles not round, but it's usually good enough to show strobing, ghosting and some other undesirable effects.
Edited : Correct file attached
Post Edited (hippy) : 1/8/2008 2:26:22 AM GMT
Hippy, how's this?
display on my / other TV's ? The great advantage of any testcard is that its a common
reference if nothing else.
Edited, to give info on the binary for others, who are looking for it with forum search
PAL60 demo, showing a test card created by Hippy
5Mhz Clock, Pll16x, using pins 12-14 for TV out.
Post Edited (Baggers) : 1/9/2008 7:49:46 PM GMT
I allowed myself to get more than a little carried away with my somewhat harsh and critical comments. I was wrong for doing so. You have pushed the boundaries of what
this device is actually able to do already. I feel sure that Parallax will let you have a play with Prop II before they want to market it as a consumer product.
If you could just mention would be so great to get real UK PAL out of the device, and not a kludge on some NTSC standard. I shall also expect full SECAM - would be tripple
standard then.
@Hippy
Wonder if a EBU style card would be of more use. I believe that anything with a circle will become squashed. Anyway we all know is just for fun and has no technical merit attached to the image displayed.
With regards
Mike.
No worries, I totally understood what you were getting it, it's great progress, but as you use the test cards for professional usage, it would need to be VERY crisp especially for broadcasting etc., but for use for home entertainment / games, it's fine.
So on that point I agree with you completely, which is why we did PropGFX, where the colours are very clear, but at the cost of extra hardware in the form of an AD724 along side the propeller.
and YES, Parallax, any of your employees, if you're reading this, then I'd be honoured to be let to have a play with Prop II before they go to market. tbh, can't wait to get my hands on it lol if the truth be known [noparse]:)[/noparse]
PS, most of my earlier photo's were 16:9, and they look better 4:3, which is what I have my tv set at now [noparse]:)[/noparse] so they don't look as squashed.
There's some very slight shimmy left-to-right but certainly nothing I'd call strobing and I would
put what there is down to the TV, even TeleText shows slight shimmy. There's some diagonal striping
across the left most frequency bars ( one pixel apart ). That's expected and I see it on all DVD's
I've tested.
From a distance I'd call it "rock solid". Up close it does look like alternate lines are shifted across by
one pixel or so.
So the question must be, why is TV_Text so bad on my TV when this and the first graphics demo are
okay ? Either it's some odd interaction or there's a subtle timing difference I guess.
@ Mike : Not sure about the benefits of the EBU testcard ( just colour bars if I remember right ). The
Philips versions or IBA ETP would probably be better but only really if one has the knowledge or test
equipment to use it. Something to check pin-cushioning and set colour and contrast is all that's really
needed by a home user. Circles should be round not squashed, but it takes some effort to get the
digital image so it appears on screen as it should be rendered. Cropping in the VCD burning chain was
the original motivation and I didn't want to use a copyright image. What I have is a home-brew cross
between BBC J and IBA ETP. There's quite a science to proper test card design and I'm no expert !
could you perhaps check this code with the PAL60 driver? As it does not work on my PAL screens I should like a feedback wrt to the quality... The difference being that it works in non- interlaced mode...
http://forums.parallax.com/attachment.php?attachmentid=51286
Post Edited (deSilva) : 1/9/2008 2:22:40 AM GMT
then and was a serious statement as would show how good the PAL colours were. Circles that look like
circles, I don't see this happening. I don't put this down to your positioning of the original either. I will
be guessing that the image has been compressed further to get this coming out of the Propeller.
I wanted to simply caution people against whipping the backs off their TV sets and messing about with
the controls thinking this is a suitable instrument to drive some patterns. It is a good example (me even
impressed) it is however not a technical instrument. Nobody here has said that it is - but I think we would
all be amazed at what some people would do!
@Baggers - I've now gone in to Testcard Retirement. It's safe to come out!
Regards
Mike.
proper VCD cards circles come out okay. They have to be distorted so when the TV dot-size
takes effect they go back to being round but I've managed to get round circles on both 4:3
and 16:9. It is hard work though.
And you're right ... please people don't take the backs off your TV and start twiddling ! Not
only is it dangerous, but what may look pretty as a testcard isn't meaningful unless the card
is calibrated, and this certainly isn't.
@ deSilva : About to test your code ...
Out the box, so presumably running with Parallax TV.spin, it worked okay. Mikado draw Lines are extremely badly strobing, both yellow and black lines ( shaking rapidly left to right ). 'Menu text' is rock-solid. Text test, same shaking as with Mikado lines.
Using the PAL60 TV.spin, the height changes drastically( to be expected, top of the 2.1 of "Vers 2.1" touches the top of the screen ). Mikado lines were solid, no strobing / shaking. Text test no strobing / shaking.
And an amazing coincidence / rare sight ... BBC2 transmitted Testcard J for a few minutes at 03:00 along with a digitised speaking clock. So at least I now have the feel for what the TV does on a genuine reference ! It did suffer some strobing on that so I guess the TV is inherently susceptible to it.
Post Edited (hippy) : 1/9/2008 3:18:12 AM GMT