News:

Printed Amstrad Addict magazine announced, check it out here!

Main Menu
avatar_mr_lou

Do you prefer CTM simulation in videos or not?

Started by mr_lou, 19:32, 06 June 16

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Do you prefer CTM simulation or not?

Yes, I prefer this CTM simulation of yours
12 (70.6%)
No, I prefer the raw pixels, even though I know that's not how it looked back in the day
5 (29.4%)

Total Members Voted: 17

Gryzor

I honestly can't tell the difference between the last two vids! Maybe it's that I'm in a well-lit room, but I froze them both and was switching between them... couldn't tell them apart!

||C|-|E||

The second one is sharper!  :D

I will go for the first one, in my opinion it is closer to what you could see in a CRT (at least in the center, since the CRTs do not behave the same in all the areas of the screen...)

mr_lou

The difference is the amount of glow-blur.
Logically the 60 one should be sharper than the 80 one.

seanb

Quote from: mr_lou on 13:32, 08 June 16
I can't produce a better pattern when being limited to 720p (and I am).
It's the closest thing I can get to the real thing, faking the placement of the RGB diodes (or whatever they're called).

Here's another rendering with glow-blur set to 60 and 80 and no gamma change from the original recording.

http://www.dewfall.dk/Bombjack-glowblur60.mp4

http://www.dewfall.dk/Bombjack-glowblur80.mp4

Applying the PNG file with semi transparent black lines obviously makes the whole picture a bit darker. Then adding the glow-filter (because diodes glow) brightens it back up. Some colour is changed in this process.

The question is, what do we prefer then? I'm thinking this project will represent the CPC in some way. How do we want to present it?



I really like the Blur 60 ine a lot. It keeps the colours bright and vivid as CPC colours should be and doesn't blur the image out too much.
Thou shall not question Captain Wrong!

mr_lou

Quote from: seanb on 10:47, 09 June 16I really like the Blur 60 ine a lot. It keeps the colours bright and vivid as CPC colours should be and doesn't blur the image out too much.

You don't think it's a bit too bright? I was thinking about lowering brightness a bit actually.
But yea, I do think this is the closest I can get to simulating a feeling of a CTM screen.

seanb

Thou shall not question Captain Wrong!

mr_lou

Quote from: Gryzor on 13:00, 08 June 16I think the amount is fine; it's the pattern that's not perfect...

Here's another attempt.

Different pattern, bigger process.
Two problems: Because the resolution isn't high enough, I have to blur a bit more than I should in order to hide the wrong looks. And the colours are a bit dusty right now.


http://www.dewfall.dk/caves.mp4

Have an idea for another pattern to try out too, but right now it's bedtime.

Gryzor

Hmmm I think I like this less - too fuzzy and the pattern doesn't look too right...?

||C|-|E||

#33
Quote from: mr_lou on 20:37, 10 June 16
Here's another attempt.

Different pattern, bigger process.
Two problems: Because the resolution isn't high enough, I have to blur a bit more than I should in order to hide the wrong looks. And the colours are a bit dusty right now.


http://www.dewfall.dk/caves.mp4

Have an idea for another pattern to try out too, but right now it's bedtime.

I like the previous ones more, to be honest  :) Maybe this changes from person to person, but when I stare at my CRT what I see more is the vertical pattern of the grid. Something like this:



However, I do not easily see horizontal patterns, like the scanlines that are emulated everywhere.

TFM

Last one is like the original monitor!  :)
TFM of FutureSoft
Also visit the CPC and Plus users favorite OS: FutureOS - The Revolution on CPC6128 and 6128Plus

Gryzor

Didn't we have some close ups posted somewhere?

Skunkfish

Quote from: ||C|-|E|| on 14:54, 16 June 16
I like the previous ones more, to be honest  :) Maybe this changes from person to person, but when I stare at my CRT what I see more is the vertical pattern of the grid. Something like this:



However, I do not easily see horizontal patterns, like the scanlines that are emulated everywhere.

You're obviously sitting far too close to the screen, your eyes will go square!  :o
An expanding array of hardware available at www.cpcstore.co.uk (and issue 4 of CPC Fanzine!)

mr_lou


||C|-|E||

I like spind3.mp4 more :) The colors are brighter and they blend better together, in my opinion.

TFM

TFM of FutureSoft
Also visit the CPC and Plus users favorite OS: FutureOS - The Revolution on CPC6128 and 6128Plus

robcfg

@mr_lou , Have you tried taking a macro picture of the CPC screen?


That will surely show the true grid size of the screen so you can create the adequate filter.

mr_lou

Quote from: robcfg on 20:13, 17 June 16
@mr_lou , Have you tried taking a macro picture of the CPC screen?


That will surely show the true grid size of the screen so you can create the adequate filter.

Yes I have. That's what I'm doing. But in order to create the best possible filter I will need even more than a 4k resolution.
I'm limited to 1280x720 pixels so I have to come up with some alternative within this limitation.

The grid on the CTM isn't that complex. It is vertical lines, each one shifted by a half in relation to each other. To reproduce this with 1280x720 pixels you can either filter out every 2nd pixel, or every 4th pixel. It really only comes down to which of those two you prefer. I'm leaning towards the 4th pixel one because my project is designed to run on a Blu-ray player, meaning the viewers will/should be sitting a good distances from the screen, and then get an impression of a CTM. It's not exactly the distance between LEDs that the CTM has, but from a distance it gives the impression of a monitor - if applying some blur too.

At least that's what I think. But that's why I'm posting here of course, to find out what everyone else thinks.

Don't you think the spind3.mp4 one could pass?

||C|-|E||

I think that it could definitely pass, but it would be great if you could show us many different games to better evaluate. For example, I am very familiar with Dinamic arcades and seeing these will make easier to judge in my case  :)

roudoudou

Quote from: robcfg on 20:13, 17 June 16
@mr_lou , Have you tried taking a macro picture of the CPC screen?


That will surely show the true grid size of the screen so you can create the adequate filter.


Last time i saw a CPC screen macro view, i notice that the grid was kind of 7 physical pixels for 8 logical pixels, this is noticable in mode 2 where one single pixel is almost never on the grid, but slide a little more than a physical pixel




spind2 video is a little smooth, very nice and respect the main philosophy


When i see a real monitor, i think the main difference is when color change from one pixel to another, then the color isn't correct but a mix of neighboured colors


There is indeed more calculations if you want to do this


My pronouns are RASM and ACE

mr_lou

#44
Another attempt, again using a different approach, and also a different pattern.

http://www.dewfall.dk/spind4.mp4

This one is very time consuming. What I'm doing is actually to scale the 720p video up to 4k resolution. Then applying an overlay with RGB-elements using multiply filter. Then adding some effects. This produces a 4k CTM simulation rather well I think. Looks cool. I think I will save these for a future 4k release of the project.  :)

But for the current project I'm limited to 720p, so I scale down this 4k video again and end up with the spind4.mp4 result.

I think this is the approach I'll end up using. I've already hired a FFmpeg expert to create a script for me that executes these steps.

The problem with the other attempts is that the actual CTM monitor will give a different pattern depending on which colour it's showing. Depending on which channels of the RGB-elements are lit up, it'll give a different pattern. Therefore any static pattern will never be satisfying.
By using 4k to apply the filter, and scale down to 720p again, we will get a varying pattern and looking close we won't recognize the CTM as such. But from a distance I think this is the best we can do with 720p.

Opinions?

roudoudou

still prefer the spind2 video!
My pronouns are RASM and ACE

||C|-|E||

Well... with the audio I must admit that I always want the cleanest possible output  :) however, I actually like how low resolution things look in CTM screens, more round and so on  :)

Executioner

Quote from: roudoudou on 14:18, 18 June 16
Last time i saw a CPC screen macro view, i notice that the grid was kind of 7 physical pixels for 8 logical pixels, this is noticable in mode 2 where one single pixel is almost never on the grid, but slide a little more than a physical pixel

I'm not sure it's even 7. Seems to be about 5 pixels on my macro images. This from Plustest:



mr_lou

When I take a close look at my CTM monitor, the 8-pixel cursor seems to light up about 10½ RGB elements horizontally.

So that's what I'm trying to reproduce.

I've made so incredibly many experiments at this point now. I'll be using an approach that (obviously) doesn't give a 100% representation, but yet one that'll give an impression of a CTM monitor.
While it's not 100% like the CTM, flat colours aren't either. And then I prefer this "old look" rather than the flat colours.

seanb

Seeing them all together on my TV I'm going to go for spind3 because the colours are brighter.
It misses some of the emulation effect compared to spind2 but not much and I like that one more than the rest.
Thou shall not question Captain Wrong!

Powered by SMFPacks Menu Editor Mod