Hybrid Scanlines

Viewing 15 posts - 31 through 45 (of 98 total)
  • Author
    Posts
  • #19566
    James-F
    Participant

    @paulb_nl
    Can you please upload a firmware to test the new low setting?


    @Retrorunner

    Any form of artificially generated scanlines will result in visible gamma change on a display that is calibrated to a gamma of 2.2 with SOLID (full, no holes) greyscale.
    Yes, that implies that removing half the visible area of a certain grey shade will make it visibly darker.
    The human eye is FAR from linear therefor using 100% strength scanlines will NOT result in the same visible gamma curve as without scanlines on the same monitor even though same amount of area was removed from each shade of grey.
    Hybrid scanlines kills two birds with one stone in this instance, it will emulated the larger beam spot with brighter shades AND fix the visible gamma curve,,, IF properly adjusted.
    If the resulting curve is darker looking at the darker shades (which it will because hybrid scanlines [and standard] do not touch this region of the grey scale), you can always bring the brightness slightly up on your TV or lower the Gamma setting if it has one.

    I don’t think the two completely different equations of scanline can be combined in any simple way.
    As said, just bring the brightness on your TV up to brighten the resulting dark region, or lower the gamma.

    #19567
    Retrorunner
    Participant

    The human eye is FAR from linear therefor using 100% strength scanlines will NOT result in the same visible gamma curve as without scanlines on the same monitor even though same amount of area was removed from each shade of grey.

    True but there is still a big difference in gama/color between standard scanlines and bortis scanlines. Using 27% standard scanlines produces visible color change. When using bortis scanlines with 37% the colors looks much more as the original one compared to 27% standard scanlines. With both methods the visibility of scanlines are near the same.

    As said, just bring the brightness on your TV up to brighten the resulting dark region, or lower the gamma.

    I tried but the colors of dark areas sill does not have the same color as with bortis scanlines. I am not an expert but IMO its not a question abouet the brightness because the color tone looks just different because of the way the scanlines are generated. Screenshot[1]
    You can decrease the hybrid scnalines value to 18% to have more original colors on dark areas but the downside is that on bright colors the scanlines are barely visible.

    [1]

    #19576
    paulb_nl
    Participant

    @James-F Here is the firmware with Low, Medium and High settings. https://drive.google.com/file/d/1s2nsIJRXnO_PCNTRWniIhAeW0SC9ZXa8/view?usp=sharing


    @Retrorunner
    Your photos look like your TV loses color with scanlines on. Is everything black and white when you set scanlines to 100%? If so maybe it will help to change scanline position from Top to bottom.

    #19585
    Retrorunner
    Participant

    @retrorunner Your photos look like your TV loses color with scanlines on. Is everything black and white when you set scanlines to 100%?

    With standard scanlines 100% I still have color but its a huge color loss which I thought is a sure thing due to substraction method of the standard scanlines option? With bortis scanlines method I have less color loss (especially when using less than 50% scnalines) because it uses multiplication method.

    Here is the explanation from borti about the difference between subtraction (standard scanlines) and multiplication (bortis scanlines) method translated from german in english:

    Personally, I find the scanlines on the OSSC not optimal because they are generated by subtraction.
    Better I find a generation by multiplication, which I do the rest of the N64A so.

    This has the effect that pixel brightness in the scanline is darkened relative to the reference pixel and not absolute. An absolute calculation even results in a discoloration.

    Simple example:
    Suppose we have a scanline strength of 50%. The reference pixel has RGB values of (255,120,176). The OSSC is stripped 127 everywhere and you get (128,0,49); So a discoloration.
    The N64A is multiplied by 0.5 everywhere and you get (127,60,88); so no discoloration, but a uniform darkening.

    #19587
    paulb_nl
    Participant

    Borti’s method has brighter scanlines with more accurate color mostly at the lower scanline strength. When you set darker scanlines you can’t see the difference anymore.

    However the color loss I am seeing in your pictures should not be so big. The picture of the car is almost black and white. Is that a bad photo or is it really like that? It should be dark but still have good color.

    Did you try to change the scanline position?

    #19588
    Retrorunner
    Participant

    Borti’s method has brighter scanlines with more accurate color mostly at the lower scanline strength.

    Yep, thats why I would like to see a combination beetwen hybrid and bortis scanlines if possible πŸ™‚

    However the color loss I am seeing in your pictures should not be so big. The picture of the car is almost black and white.

    You mean these two screenshots?

    https://circuit-board.de/forum/index.php/Attachment/98452-DSCN4491-JPG/
    https://circuit-board.de/forum/index.php/Attachment/98453-DSCN4488-JPG/

    Thats the huge difference I have when using bortis and standard scanlines at 37% and even I use standard scanlines at 18% I have still more color loss compared to 37% bortis scanlines.Changing scanlines position has no effect on color neither brightness.

    I am the only one who tested bortis scanlines and see that big difference in color?

    #19589
    borti4938
    Participant

    Especially in the mid-range scanline strength you have false colours in every second line if you simply subtract a constant value. The example above shows that clearly. There not that much different in the low range and high range strength region; that’s also true.

    To be 100% precisely the multiplication has some slight amount of discolouring, too. Ideally you have to go into another color space, e.g. YPbPr, to set the scanlines without discolouring. But then you have to go back.

    However, back your hybrid implementation πŸ˜‰
    I merged it with mine. I kept the choice between generation methods as well as kept hybrid adjustment strength separately. So the user can combine as he likes.
    Could you please take a look into the implementation and look if your logic has been retained?
    Just a remark:
    1. I started with the adjustment of strength value after the reverse LPF
    2. Due to 1., the pipeline length has been increased a bit.
    3. The hybrid contrast on/off is done in stage 1 of the strength modification
    4. I have 100% the same commits in my local upstream branch but hasn’t pushed them as I wait for your feedback first.

    Link to branch with the changes: https://github.com/borti4938/ossc/tree/dev_lcdbl

    #19590
    paulb_nl
    Participant

    Your changes look fine to me πŸ™‚ Nice that the hybrid contrast is separate. I haven’t tested it yet but will do that tomorrow.

    #19598
    paulb_nl
    Participant

    I have tested your firmware and everything is working well! We have a lot of choices now for the look of the scanlines πŸ˜€

    #19602
    borti4938
    Participant

    OK, thanks for the feedback.
    I guess I will add for multiplier variant the hybrid contrast depending on the Y-value to tread every color equally, here.

    #19603
    Galdelico
    Participant

    This all sounds pretty awesome, Paul and Borti.

    Apologies for the newbish question: is this alternate firmware you guys coded meant to be for test purposes only – mostly aimed to the tech heads behind the OSSC project – or it can also be used by your regular Joe, along with all the features included in the latest official firmware?

    And another question: with no means to annoy/push you, is it possible to point your attention to little quircks/bugs in the OSSC settings (specifically, the horizontal mask alignment), and have at least a feedback from you about whether or not – or how much easily – they could be fixed?

    Thanks guys! ^_-

    #19607
    paulb_nl
    Participant

    @borti4938 During my testing I couldn’t get a good result with (ScanlineStrength-Luma). The color saturation was poor.


    @Galdelico
    Borti is preparing a pull request so this will all go into the official firmware. I will check out the mask issue.

    #19612
    Galdelico
    Participant

    @paulb_nl That’s great, thank you!
    Should you need any additional information/test – other than what I’ve already reported here – feel free to let me know anytime.

    #19613
    Retrorunner
    Participant

    Tested: Scanlines 37%, Multiplication vs Subtraction und hybrid contrast: Low. The color is definitly closer to the original with multiplication (bortis scanline solution) as with substraction. Combined with hybrid low contrast its even looking better πŸ˜€ Great work from both of you guys.

    #19626
    marqs
    Participant

    Nice to see multiplication-based and and ‘hybrid’ scanlines getting implemented. At some point I also thought of testing multiplication-based method but that got buried under other things. We could dedicate v0.81 firmware just for scanline improvements – maybe I’ll get that per-line adjustment option implemented and included there as well.

Viewing 15 posts - 31 through 45 (of 98 total)
  • You must be logged in to reply to this topic.