Jump to content
Have You Played Atari Today?
2600|5200|7800|Lynx|Jaguar|Forums|Store  

Crazyace

Members
 View Profile  See their activity

  • Content Count

    1,027

  • Joined

  • Last visited

Posts posted by Crazyace

  2. Posted

    Actually the mode that Rybags uses for his New year 2010 picture is pretty interesting - that seems to be combining a line of 16 colour GTIA mode ( 80 pixel low luminance ) with a line of Mode 4 ( 160 pixel 4(5) luminences ) to give a higher res picture. Of course it's not 256 colours, as the mode 4 only offers 8 luminance levels.

    I wonder how some of the images would look in that mode?

  3. Posted

    I did explain to you why I don't have a 256-color paletted picture; the Photoshop that I use maps it using some linear distance formula although manually I can see that A8 palette contains the colors that fit better. As I said, the industry that deals with chroma/luminance (as in JPG/MPG/etc.) will take the luminance over the chroma at even bigger ratios of chroma loss than the ST palette vs. A8 palette.

     

    Sorry chroma loss is a compression artifact ( and it's to do with resolution ) - so you're not really talking about pallettes. ( If you want to bring resolution into it we could just resample the parrot picture to be 4 times coarser in X )

    Anyway the picture I'm more interested in is your monochrome 160x240x16 colour one ( non interlaced ) - or was that a photoshop as well?

  4. Posted

    You don't have more cycles on ST than A8 with the latest Boulderdash code you wrote. I didn't even verify that code by the way to see if it works for all cases-- just taking your word for it.

    And you're not proving ST is superior by providing code that takes tons of CPU time. It was an issue whether ST can fit it into a frame's time.

    No problem, I calculated it at 39K cycles free at 50Hz , which is more cycles than the A8 has in total per frame, or 12K cycles free at 60Hz , which is about 40% of the total cycles per frame on the A8 per frame.

  5. Posted

    Oh dear, I guess it was crazy, after all there's no way the parrot could actually be reproduced on the A8 using all 128 colours at that resolution. Please feel free to provide a version of the parrot picture on the A8 , oh - and I'm also waiting for an A8 bin/image of the Kaliya picture.

    Go re-read what you wrote. You claimed that A8 has 128 colors and that's why it was okay for you to use the 128 color picture. You already drew the conclusion-- you weren't waiting for any 256 color picture; in fact, you denied that A8 has 256 color palette.

     

    Still waiting for your picture - I'm sticking with my view that the ST palette is generally superior to the A8 palette.

  6. Posted

    When I gave you example of Boulderdash, you drew some absurd conclusion that this proves that ST can do A8 games.

    No, actually I wrote code for you to show Boulderdash would be possible ( this is called proof ) - and made a considered statement that any A8 game could be reproduced on the ST.

    ...

    I am talking about the conclusion you drew. It's not proof that any A8 can be reproduced; it's just that you can fit it into the frame's time although still inferior to A8.

     

    As I said, it's my opinion that any A8 game could be reproduced on the ST. The boulderdash code snippet was proof that I could replicate the Boulderdash scrolling, and still have more cpu cycles than there would be free on the A8. ( I gave you another code snippet for Mr Robot as well ... but I dont feel the need to do it for every game there is :) )

    If you want to argue the point further feel free to provide a game that you think cant be replicated.

  7. Posted

    But in this situation we're talking about fixed values. If the A8 had palette registers not indexes (like the ST where you give it the RGB values, not pick them from 512) but in YUV then there'd be some room for utilising the head/toe room of the YUV formulas. As it is they're pretty much classed as errors when converting (from an RGB viewpoint) because they do produce ranges you don't really want to display normally.

    From my experience VGA cannot reproduce the brighter colors of the A8 and C64. Has nothing to do with "formulas" but with $FFFFFF being a nice background color for MS Word instead of a bright white.

     

    In this context, the ST doesn't use RGB via VGA , but via a TV - so that's where the comparision should be made

  8. Posted

    Add that to the list of "crazy" things you did. Compare 128-color based parrot to ST's 512-color based palette and claim A8 palette is 128 colors.

     

    Oh dear, I guess it was crazy, after all there's no way the parrot could actually be reproduced on the A8 using all 128 colours at that resolution. Please feel free to provide a version of the parrot picture on the A8 , oh - and I'm also waiting for an A8 bin/image of the Kaliya picture.

    • Like 2

  9. Posted

    When I gave you example of Boulderdash, you drew some absurd conclusion that this proves that ST can do A8 games.

    No, actually I wrote code for you to show Boulderdash would be possible ( this is called proof ) - and made a considered statement that any A8 game could be reproduced on the ST.

     

    You were suppose to do that but you only went one way and only showed the ST interlaced picture.

    In my replies I've always given 2 cases - one interlaced(photochrome) and one not.

  10. Posted

    Yes, I should have written "to emulate ACSI in firmware". I don't know what are the exact differences between ACSI and SCSI (maybe they're totally different, I don't know).

     

    No problem, I never really thought about the hard drive too much ( it was a bit too expensive, so I tended to do all of my programming on a 2 floppy ST )

  13. Posted

    You miss one bit: the hard disk driver must be booted. So it is either BIOS which actually has to support IDE (in predefined routines, like TOS 2.06), or the disk controller (= "separate computer" I wrote about) which has to emulate SCSI in firmware. Sorry.

     

    Isn't the Atari hard drive interface some variant of SCSI? How easy would it be to wire a SCSI drive up directly?

  14. Posted

    Just as a comparision I've prepared 3 versions of that C64 picture ( taken from the original image on CSDb rather than the processed C64 one - which does look pretty good )

     

    STterm is a non interlaced 60Hz image.

     

    PhotoTerm is the interlaced version.

     

    A8Term is an A8 GTIA version, ( mono only - there are other people who could convert it to better A8 colour formats )

     

    Again it's a pretty easy comparision to make :)

    post-4839-126264186017_thumb.png

    post-4839-126264189396_thumb.png

    post-4839-126264198349_thumb.png

  15. Posted · Edited by Crazyace

    It's related. The main reason to even switch color spaces when compressing in MPEG, JPG, etc. is because it's based more on visually appearance rather than what's better or convenient for the monitor and hardware. I have a bunch of hardware here like a video digitizer board for PC/Amiga from 1980s that uses 16bit in another color space to show 24-bit RGB images. It's faster to upload/download to the board, uses less space, and looks like 24-bit images. It's not my opinion that I'm giving edge to A8-- I gave you several reasons already. One of the reasons was that an application that does lighting/ray tracing would already pick the best color and only the luminance would be the major player. TVs already have erroneous chroma when you compare two of them (so they are more tolerable). MPG/JPG trashes chroma over luminance. You are mixing things up with resolution-- that's why you would give edge to ST. If all things are kept same, A8 palette gets the edge over ST's palette. You haven't given a single reason why St gets the edge except for just saying that it lacks in reds. And this is debatable-- whether you would take a hit on the reds or luminance values.

     

    I think you're a little confused, MPEG/JPG and PC/Amiga video digitizers don't actually have anything to do with the A8 pallette ( perhaps you need to read up on your research material )

    Ray tracing apps just generate pictures - the palette on ST or A8 is only how the final image is displayed.

     

    I've said it's my opinion - but please feel free to make some more 'real world' examples of images quantised to A8 and ST palette, and we can compare. ( and I'm still interested in your Kalya image on A8 )

  16. Posted

    I would say it's easier to take a RGB image (as most are in modern age) and convert to ST's color space which is also RGB whereas converting to a color image using A8 palette requires some heuristics and manual touching up.

    Also the A8 is lacking in reds, which can be a lot more noticable than the 8 shade limitation on the ST.

    ...

    Nope, that's your speculation. Luminance definitely has more impact visually than having a red being off unless for some special cases where image is more biased towards reds which was probably case of the Parrot.

     

    Notice I said 'can' - the example of the parrot shows a marked inferiority of the A8 pallette compared to the ST palette. Therefore it's not speculation on my part.

  18. Posted

    I would say it's easier to take a RGB image (as most are in modern age) and convert to ST's color space which is also RGB whereas converting to a color image using A8 palette requires some heuristics and manual touching up.

    Also the A8 is lacking in reds, which can be a lot more noticable than the 8 shade limitation on the ST.

     

    I doubt your photochrome has an option to convert image to A8s palette.

    It's not 'my' photochrome - just an ST application for showing images.

     

    The point is about the palette as a whole which includes both luminance and chrominance-- I give the edge to A8 despite it having more quantization on the chroma side. As I said, you can subsample chroma-- not even approximate with dithering but completely trash 93% (15/16) values and still get image looking almost same as original.

     

    You seem to be talking about image compression here - which is completely unrelated to A8 and ST palettes.

    ( The only link I could think of is the interlacing of a low res colour only GTIA mode with high res luminance mode. )

     

    As you say - you give the edge to the A8 here, that's your view. I give the edge to the ST.

  20. Posted

    It is 160*240*16 but it's pixel replicated to 320 width so all the images have same width.

    Lot's of that picture seems to be 80 pixels replicated to 320 - only small sections look like 160 , which is why I'm not classing it as 160*240*16.

    There's not interlaced stuff there. The new images are Kaily565.bmp and Kaliy411.bmp

    These aren't actually A8 images though, so they're not actually relevant in any way to the topic.

  21. Posted

    That is an advantage - but I think the way forward was to move away from NTSC style encodings - RGB ( even though it only have 8 bits per channel on the ST ) is a better general colour encoding system than the older atari scheme.

    The ST uses a 9-bit RGB palette, right, with 3-bits per channel, not 8-bits which would be 24-bit RGB? (unless I'm misunderstanding your statement)

     

    Sorry, slip of the keyboard - I meant to say 8 levels, which is 3 bits per channel. Thanks for pointing that out.

  23. Posted

    I have no means of getting interlaced stuff of my A8 into BMP yet. I'm working on it. Here's one I posted previously that uses resolution enhancement to get 60Hz imagery at 160*240*16.

     

    Your picture is pretty nice - How does it look on an A8? - It's not really 160x240x16 though - although you have done a really good job on the face.

     

    Here's a version for normal ST , and a photochrome interlaced version to compare with. ( I cut out the middle 320x200 from the picture , as that shows most of the colour range )

    post-4839-1262546825_thumb.png

    post-4839-126254685956_thumb.png

  24. Posted

    I might do that if power management and other factors are affecting their high resolution timers.

     

    ...

     

    That was my point as peteym5 also stated. If you had an A8 during 1970s,1980s, and wanted to do DTP or CAD-stuff, you had many choices not just ST. During 8-bit era, A8 ruled supreme except for maybe 640*yyy modes which only CGA had.

     

    Well if you do, make a new topic for it - this one is about A8 and ST :)

  25. Posted

    I am waiting for someone to do a "Atari 8-Bit" vs "Modern 64-bit Microsoft Windows 7 PC".

     

    Don't do that - Atariksi will be there lambasting the Windows 7 PC for not having 1.79MHz timers and superfast joystick ports :)

     

    Hate to mention it, but the PC success was mainly because of Microsoft. Even with all the GUI OS blunders, Windows dominated and became the world standard for personal computers.

    I think the PC was winning even in the Dos era - VGA and SuperVGA were the best graphics avaible, and the cpu's were getting fast and cheaper at the same time.

    • Like 1
AtariAge
Forums
Store
General
Follow AtariAge

Copyright ©1998-2021 AtariAge

×
×
  • Create New...