Blurinator 9000 - Blur simulation software

Advanced display talk, display hackers, advanced game programmers, scientists, display researchers, display manufacturers, vision researchers & Advanced Display Articles on Blur Busters. The masters on Blur Busters.
MCLV
Posts: 43
Joined: 04 Mar 2021, 15:04

Blurinator 9000 - Blur simulation software

Post by MCLV » 30 Apr 2021, 14:29

I just published first version of a software for LCD blur simulation which I named Blurinator 9000. My goal was to create an educational tool for demonstrating basic effect in LCD motion blur physics rather than an industrial grade LCD simulator. I hope that you will like the result :) It is available via this link https://github.com/marianvlcek/blurinat ... /tag/v9000 and it looks like this:
Blurinator_9000_screenshot.png
Blurinator_9000_screenshot.png (245.51 KiB) Viewed 19674 times

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: Blurinator 9000 - Blur simulation software

Post by Chief Blur Buster » 30 Apr 2021, 16:50

MCLV wrote:
30 Apr 2021, 14:29
I just published first version of a software for LCD blur simulation which I named Blurinator 9000. My goal was to create an educational tool for demonstrating basic effect in LCD motion blur physics rather than an industrial grade LCD simulator. I hope that you will like the result :) It is available via this link https://github.com/marianvlcek/blurinat ... /tag/v9000 and it looks like this:
Blurinator_9000_screenshot.png
Oh, fantastic work!

This does what I wanted to do in year 2014 but never had the time to: a software-based display emulator package. TestUFO is gradually getting there, but it's in bits-and-pieces in different tests rather than a unified display emulator.

Related Thread: Proposed TestUFO Display Emulator in 2014

This can provide the basis for collaboration on improved software-based display emulators, which I've been dreaming of doing for years (it's a very programming heavy work) and this is something I have been wanting to start as an open source project.

Hopefully I can collaborate some source code later this year to this and/or port this to JavaScript. PixPerAn had a very basic display emulator at the bottom edge even twenty years ago, but I have been wanting to build this into TestUFO eventually. The math is quite simple in my mind, but it's just a lot of unpaid programming. Thanks for stepping up to the plate to doing this.

Long-term (2021-2025) enhancements I'd love to see in commits / forks / inspired spinoff projects, crediting your original project of course!
  • Color sequential emulation, like this beta TestUFO emulator
    [WARNING: Epileptics using 60Hz displays, bad red/blue/green flicker!] ...)
  • PWM emulation (to generate the ugly serrated artifact at www.blurbusters.com/lcd-motion-artifacts ...), like this TestUFO emulator
  • DLP temporal dithering emulation
  • Plasma temporal dithering emulation (tightly grouped pulses)
  • CRT electron beam emulator (rolling scan emulation)
  • Artifically fast/slow simulated LCD GtG using the formulas written by ScepticMatt ([urlhttps://forums.blurbusters.com/viewtopic.php?f=7&t=418]thread[/url])
  • Optional ability to playback animation in real time on ultra-high-Hz displays (e.g. emulating a 60Hz CRT on a 360Hz LCD) to watch original artifact in realtime
  • Etc
As we get to a 1000Hz display, it will be able to software-emulate a lot of retro displays including its temporal artifacts (temporal dithering, color sequential, PWM artifacts, GtG, CRT impulsing, rolling scan). Although your project mainly focusses on sample and hold behaviors, it seems extensible to emulate other kinds of displays in the future by modelling sub-refresh behaviors.

Blur Busters a great prestigious place for this type of collaboration work, so I am honored to be the first website to get knowledge of your open source project!
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

User avatar
Discorz
VIP Member
Posts: 999
Joined: 06 Sep 2019, 02:39
Location: Europe, Croatia
Contact:

Re: Blurinator 9000 - Blur simulation software

Post by Discorz » 26 May 2021, 11:33

MCLV wrote:
30 Apr 2021, 14:29
I just published first version of a software for LCD blur simulation which I named Blurinator 9000.
Ooohhh my! this is soo good!
I immediately loaded this one. Had so much fun with it :D :D :D
Image
Compare UFOs | Do you use Blur Reduction? | Smooth Frog | Latency Split Test
Alienware AW2521H, Gigabyte M32Q, Asus VG279QM, Alienware AW2518HF, AOC C24G1, AOC G2790PX, Setup

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: Blurinator 9000 - Blur simulation software

Post by Chief Blur Buster » 11 Jun 2021, 13:41

Discorz wrote:
26 May 2021, 11:33
MCLV wrote:
30 Apr 2021, 14:29
I just published first version of a software for LCD blur simulation which I named Blurinator 9000.
Ooohhh my! this is soo good!
I immediately loaded this one. Had so much fun with it :D :D :D
Want to post more feedback?
Do tell!
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

User avatar
Discorz
VIP Member
Posts: 999
Joined: 06 Sep 2019, 02:39
Location: Europe, Croatia
Contact:

Re: Blurinator 9000 - Blur simulation software

Post by Discorz » 12 Jun 2021, 17:03

Chief Blur Buster wrote:
11 Jun 2021, 13:41
Want to post more feedback?
Do tell!
I tried to get exactly the same sample-and-hold ufos like on my VG279QM (@280Hz OD80) and the results are extremely close. But I couldn't replicate things like slight overshoot mixed together with ghosting. It was either or one or other. So it feels like overshoot and GtG slider should have wider range (slower GtG, higher level of Overdrive), and at one point those two should overlap.
I was unable to replicate strobed and strobed vrr ufos properly. There is no option to move the crosstalk effect (up-down/left-right) to match double imaging and make it fainter. So attached strobed and strobed vrr ufos are the most inaccurate.

Its missing panel type emulation like Chief said because each panel has different type of ghosting - different transitions-different response. This seems to be closest to IPS or OLED. For example VA panel black level smearing can't be replicated here. I can imagine how difficult it is to code/simulate that. But overall very impressive work for demonstrating basics. I enjoyed it :D.
If some of the above were taken into account this could be a very useful tool for demonstrating specific monitor motion performance or more. But it would require the right person/reviewer to do it accurately.

This is the closest result I could get for VG279QM:

1. sample-and-hold _______2. strobed ______________3. strobed vrr
Image Image Image
VG279QM, 280Hz, OD80, 960pps, middle screen - numbers 2 and 3 are inaccurate, reference here | Simulated with Blurinator 9000 software (by Marián Vlček)
Compare UFOs | Do you use Blur Reduction? | Smooth Frog | Latency Split Test
Alienware AW2521H, Gigabyte M32Q, Asus VG279QM, Alienware AW2518HF, AOC C24G1, AOC G2790PX, Setup

MCLV
Posts: 43
Joined: 04 Mar 2021, 15:04

Re: Blurinator 9000 - Blur simulation software

Post by MCLV » 13 Jun 2021, 15:47

Thank you very much for positive feedback! I'm glad that you like the tool :)

There are definitely some limits since I tried to program Blurinator in a quite simple way. The source code has just around 200 lines and the blur simulation is basically done in about 20 lines. The rest is just an auxiliary code and a code necessary for GUI.

All GtG transitions are handled in a same way so faithful simulation of VA panel is not possible. On the other, it greatly simplifies the code since one can generate response curve beforehand and apply it to whole image with one line of code.

Similarly, subpixel blur cannot be replicated and blur can be set only in integer number of pixels. This includes setting of strobe length which always result in an integer blur trail. Strobe phase is another thing that is fixed at the optimum timing, i.e. just before next refresh occurs.

Overdrive at the setting without overshoot halves the response time and there is no way to adjust this factor in a built executable. To model it, I used a sum of two exponential functions which is obviously not identical to a behavior of a real display.

It's definitely possible to create a more accurate simulator capable of simulating different GtG transition speeds depending on the initial and final levels of grey. I wouldn't say that is too difficult, I guess that initial version would take a week of full time development. However, the effort to make a polished product with a nice GUI would be clearly larger. In addition to that, I would expect that the computation demands would be higher so such simulator would not be as responsive unless some optimization effort would be spent. Therefore, I don't intend to go in this direction now.

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: Blurinator 9000 - Blur simulation software

Post by Chief Blur Buster » 13 Jun 2021, 17:11

Being an open source project -- if any developers are reading this -- keep in mind the PixPerAn style adjustments can be somewhat easily added:

Image

And/Or the proposed GtG math formulas provided by SkepticMatt made in Mathematica (see these images)

Perhaps a settings or user-entry box to allow defining a custom formula to generate the GtG/overdrive curve (would have to be formatted in one line). A little more generic than supporting OD LUT, but this is one idea for any enterprising open source developer wanting a way to incubate simple support for alternate GtG pixel response curves...
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

User avatar
Discorz
VIP Member
Posts: 999
Joined: 06 Sep 2019, 02:39
Location: Europe, Croatia
Contact:

Re: Blurinator 9000 - Blur simulation software

Post by Discorz » 08 Jul 2021, 07:51

I tried to simulate strobing in photoshop (basic online https://pixlr.com/e/) but couldn't get the frame blending right. Yours is much closer to real life. Notice how previous frame of mine has less pronounced colors because I only used transparency to get the effect. The ufo black lines from previous frame inside current frame is less pronounced in Blurinator, which is accurate. It seem to have some type of exposure applied to it, exposure that doesn't destroy colors. What was your approach to this? How to achieve this effect in ps?

Image Blurinator 9000 - accurate
Image Pixlr E - inaccurate
Image reference
Compare UFOs | Do you use Blur Reduction? | Smooth Frog | Latency Split Test
Alienware AW2521H, Gigabyte M32Q, Asus VG279QM, Alienware AW2518HF, AOC C24G1, AOC G2790PX, Setup

MCLV
Posts: 43
Joined: 04 Mar 2021, 15:04

Re: Blurinator 9000 - Blur simulation software

Post by MCLV » 12 Jul 2021, 06:38

Discorz wrote:
08 Jul 2021, 07:51
What was your approach to this?
RGB values have to be converted to linear scale at first. Then one can process image and sum the luminance correctly. Afterwards, gamma curve has to be applied to return to original encoding. Blurinator uses hardcoded gamma value of 2.2.

I think that the difference that you see likely comes that the other software does not operate on the linearized image but directly averages RGB values.

User avatar
Discorz
VIP Member
Posts: 999
Joined: 06 Sep 2019, 02:39
Location: Europe, Croatia
Contact:

Re: Blurinator 9000 - Blur simulation software

Post by Discorz » 12 Jul 2021, 07:26

MCLV wrote:
12 Jul 2021, 06:38
Guess I can't improve it further then.

Do you have any future plans with this software, updates? If yes, can we expect improvements like wider/slower response times range and a possibility of mixing ghosting with inverse ghosting (overshoot)...?
Compare UFOs | Do you use Blur Reduction? | Smooth Frog | Latency Split Test
Alienware AW2521H, Gigabyte M32Q, Asus VG279QM, Alienware AW2518HF, AOC C24G1, AOC G2790PX, Setup

Post Reply