Found a scan doubler ( who would like one ) ?

[alexh]

-D- wrote:
Alexh I'm not necessarily disagreeing with that, all I'm saying is that I'm not convinced this unit is 16-bit and/or would b0rk the color output.

And I agree, 24-bit color gradients will look messed up switched to 16-bit. But this isn't exactly the same thing, as AGA isn't capable of *outputting* smooth gradients, since it can't simultaneously display the amount of colors necessary for that (and HAM8 produces nasty color artifacts).

Sorry but you are just thick if you cannot understand it.

The 16-bit scandoubler ignores the 3 LSB's of Red and Blue and 2 LSB's of green. That means that an AGA colour with a blue value of 0 is displayed the same as an AGA colour with a blue value of 7! (assuming the other components are the same)

AGA can display smooth gradients. If say the colours in a program used 256 shades of red for example (highly unlikely but possible) it would have ALL the possible RED shades of a 24-bit screen and if you use a 16-bit scandoubler you would only be able to distinguish 32 of them from the others.

You get a BANDING effect as what was supposed to be a subtle gradient of colours appears as one.

[alexh]

-D- wrote:
Millions of devices around the world use these chips for video playback

Ah, yes but they DONT use the RGB input they use the Y/C input which doesnt need 24-bits for a reasonable colour reproduction

-D- wrote:
It will be at least as capable as any earlier "amiga specific" scandoubler.

Bollox, the Arxon and petsoff scandoublers are both 24-bit RGB scandoublers.

-D- wrote:
Furthermore, if it is indeed using the Averlogic chips which alexh mentioned, they both have 8-bit DACs each for RGB.

Absolutely true. But the AL250a only has 5/6-bit inputs! You cant magically get data back once you've thrown it away!

-D- wrote:
HAM8 mode is essentially a 15-bit *external* output, anyhow.

You've read too much crap off the internet.

[Damion]

Sorry but you are just thick if you cannot understand it.

The 16-bit scandoubler ignores the 3 LSB's of Red and Blue and 2 LSB's of green. That means that an AGA colour with a blue value of 0 is displayed the same as an AGA colour with a blue value of 7! (assuming the other components are the same)

LOL, I'm "thick" because because the datasheets for the chips you mentioned state they have 8-bit DACs for each color, yet you keep saying it's 16-bit. That's pretty funny.

This has nothing to do with me not understanding the difference between 16 and 24-bit output. Any dumbass can switch from truecolor to highcolor and see the difference (and I've already stated my awareness of the AGA 24-bit *internal* palette).

AGA can display smooth gradients. If say the colours in a program used 256 shades of red for example (highly unlikely but possible) it would have ALL the possible RED shades of a 24-bit screen and if you use a 16-bit scandoubler you would only be able to distinguish 32 of them from the others.

Yes, using 256 colors it's possible for a smooth gradient, but as you've stated, highly impractical. Try viewing the image you linked to on an AGA screen, and it will look like ass.

You get a BANDING effect as what was supposed to be a subtle gradient of colours appears as one.

Due to the very nature of how HAM works, you get fugly looking images, gradients included. This even has nothing to do with HAM8 not being a 24-bit output in the sense of displayable colors, regardless of the internal palette.

[alexh]

quote]
-D- wrote:
LOL, I'm "thick" because because the datasheets for the chips you mentioned state they have 8-bit DACs for each color, yet you keep saying it's 16-bit. That's pretty funny.[/quote]
You know what is funnier... the same datasheet with 5-bit red input, 5-bit blue input and 6-bit green input!

-D- wrote:
Yes, using 256 colors it's possible for a smooth gradient, but as you've stated, highly impractical. Try viewing the image you linked to on an AGA screen, and it will look like ass.

Look at the RED bit in one AGA screen, BLUE bit in another and GREEN in another and they look IDENTICAL to the PC in AGA!

-D- wrote:
Due to the very nature of how HAM works, you get fugly looking images, gradients included. This even has nothing to do with HAM8 not being a 24-bit output in the sense of displayable colors, regardless of the internal palette.

No HAM8, just regular AGA screen mode, any number of colours you like. 256, 32, 8.

Choose a different 24-bit palette entry value for each colour where only the LSB's are different.

With an AGA paint package, paint the background in one of these colours and write your name using the different colours.

Look at it on your XRGB monitor... yep cool... "-D-"

Now use a 16-bit scandoubler... but wait... I cant see anything... everything looks the same colour!!

ARRRGHHH!!!

[alexh]

One other thing I should point out is that the scandoubler is not (by default) a frequency shifter, and so PAL screens are still 50Hz and may not display on some LCD panels.

[Damion]

OK, I will try it... and if I have to eat shoe, I will :-p

And I agree, a 16-bit scandoubler might not be ideal for AGA, if indeed it effects the internal palette and not just the "amount" of displayable colors, admittedly I'm not an expert on this.

BUT, I can't find anything in the datasheet for the AL250a indicating a castrated input, however, I'm open to correction.


 

[JosephC]

AGA isn't capable of *outputting* smooth gradients, since it can't simultaneously display the amount of colors necessary for that

Dead wrong.  100% incorect, false and untrue.  You should be ashamed of yourself.

[alexh]

-D- wrote:
BUT, I can't find anything in the datasheet for the AL250a indicating a castrated input, however, I'm open to correction.

http://www.averlogic.com/admin_en/product_en/pic8/540738.pdf

Look at the number of bits on the input on the diagram on page 1. 16-bits

-D- wrote:
if I have to eat shoe, I will :-p

Nowt bad about being wrong mate, happens to me several times a day, you just pick yourself up, learn from it and get on with it. Makes you a stonger engineer.

[Damion]

alexh wrote:
http://www.averlogic.com/admin_en/product_en/pic8/540738.pdf

Look at the number of bits on the input on the diagram on page 1. 16-bits

OK -- you are correct. I agree this is sub-optimal...


@JosephC

Dead wrong. 100% incorect, false and untrue. You should be ashamed of yourself.

Get over yourself. 256 color gradients, yes. HAM8 gradients look fugly with an excess of color fringing/bleeding.

[alexh]

-D- wrote:
Get over yourself. 256 color gradients, yes.

No, no, no.

Not just 256 colour gradients. (Although these are the most obvious)

Any AGA screen.

If the colours use the lower 3-bits of each colour component, and there are several of them separated by only these bits. They will appear as one colour.

It is particularly noticable on AGA 3D demo's with light shading, AGA monochrome demo effects, AGA 2D pictures with flesh tones or halo's (like say from a candle).

[Damion]

alexh wrote:

-D- wrote:
Get over yourself. 256 color gradients, yes.

No, no, no.

Not just 256 colour gradients.

Any picture.

If the colours use the lower 3-bits, and there are several of them. They will appear as one colour.

Ugh... there seems to be a bit of misunderstanding going on here.

Issue one -- In the above reference (@JosephC), my intention wasn't to say that a 256 color gradient would look as it should with a 16-bit SD, only that while AGA itself (viewed normally) is definitely capable of 256 color smooth gradients, HAM8 gradients don't look right because of the way HAM works.

Issue two -- I figured that since AGA is only capable of displaying 256 simultaneous colors (for the most part, outside of HAM modes), that therefore a 16-bit SD should be fine... not taking into consideration that a 16-bit SD could actually mess with the palette capabilities (it must "round off" certain colors, so to speak). I clearly understand the difference between 16 and 24-bit in regard to displayable colors, but didn't consider that the palette (in the case of AGA, 24-bit) might also be affected. Therefore, I stand both enlightened and corrected in this regard.

Also, I determined (from a quick browse of the datasheets) that the Averlogic SD was 24-bit, after noticing the 3 8-bit DACs. I stand corrected here also, as alexh is right, the input is indeed 16-bit.

My apologies to the thread.

[Damion]

alexh wrote:

-D- wrote:
BUT, I can't find anything in the datasheet for the AL250a indicating a castrated input, however, I'm open to correction.

http://www.averlogic.com/admin_en/product_en/pic8/540738.pdf

Look at the number of bits on the input on the diagram on page 1. 16-bits

-D- wrote:
if I have to eat shoe, I will :-p

Nowt bad about being wrong mate, happens to me several times a day, you just pick yourself up, learn from it and get on with it. Makes you a stonger engineer.

Cheers. I came out guns blazing, only to feel like an ass. Sorry guys :lol:

[nBit7]

Maybe someone can find or make up a graphic that shows a 256 8bit single colour gradient next to the equivalent using only 4 and 5 bits.

That way people can see the difference for them self without having to resort to fiddling with screen modes.

[Darrin]

nBit7 wrote:
Maybe someone can find or make up a graphic that shows a 256 8bit single colour gradient next to the equivalent using only 4 and 5 bits.

That way people can see the difference for them self without having to resort to fiddling with screen modes.

I'd be more interested in someone using one to run Zool2 AGA, Gunship 2000 AGA or UFO Enemy Unknown AGA and letting us know if they are "aware" of any problems when playing a game.

I'm sure that squinting at hi-res static screens will reveal obvious differences, but I wonder if the naked eye when confronted with scrolling backdrops and fast moving sprites will detect any problems.  I use a SDFF (internal) in my A1200T and I haven't noticed any problems.  Granted, most games are ECS and I haven't tried to compare the cutput of the SDFF on a VGA monitor with my 1081 connected to the usual video out.

[JosephC]

HAM8 gradients don't look right because of the way HAM works.

You have it all backwards.

HAM8 gradients are the only thing that do look right because of the way HAM8 works.

When the color changes wildly from 1 pixel to the next you get fringing on HAM8.  When the colors are a smooth gradient you get beauty and elegance on HAM8.

[alexh]

Darrin wrote:
I'd be more interested in someone using one to run Zool2 AGA, Gunship 2000 AGA or UFO Enemy Unknown AGA and letting us know if they are "aware" of any problems when playing a game.

Probably not. You'd just shrug off having fewer colours as being part of the game. DCE FlickerMagic/ScanMagic internal scandoublers are I think 16-bit and they are widely used. I think only recently did it come to light that they were (but I am probably wrong and when they came out everyone hated them, but then forgot the limitation over time ;-))

I would imagine that using WinUAE with the aforementioned games and switching to a 16-bit direct colour mode may have the same effect as a 16-bit scandoubler. Try it.