This is an amazing project and is exactly what we need. Although I doubt we’ll ever perfectly reproduce these screens (unless humanity decides to make tubes again) we are able to get close. One thing it’s important to note on your images and project overview are the camera settings used to capture them. We then have a chance of ‘reverse engineering’ the photo so to speak.
Itll be great to go through these photos! Thanks again!
Yeah youve got to take into account youre dealing with emissive surfaces and cameras aren’t particularly good at capturing the physical surface as the sensor is getting overloaded with photons and the photons arent coming out directly at the camera and so youre getting bloom effects - the brighter the phosphor the more bloom it produces. The same thing of course will happen if you photographed a set of pixels so its a case of ignoring the photo somewhat and just modelling the underlying physical phosphors.
Fantastic - what display is this? Indeed what game is this?
Belated “thanks!” to everybody for nice comments regarding the collection. I’ve been rather busy with IRL and my other projects, also away from my hardware for nearly a year so didn’t have time to update and expand it properly, as originally intended. I’ve been back at it for the last two months though and hopefully the new version will be available somewhere early next year (with more TV set variety and more close-ups)
Here’s a sneak preview:
As for noting the settings, you can always check the exif data of individual images (that said I’ve recently settled for mostly using ISO 160 / f22/ 0.3s template using Canon 6D + EF 50mm 1.8). I also include a lot of RAW files, so you can try to develop them better. Aiming for colour accuracy is a nightmare (especially taking full screen photos) and the whole calibration thing way over my head/wallet, so I usually just settle for the “eyeballing-nearest” kinda style.
I’m also always on the lookout for contributions, especially the machines/TVs which arent’ in the collection already (even though I’ve acquired much more hardware since v0.1). Most wanted atm are photos of Commodore and arcade monitors, als original NeoGeo consoles (and some more obscure ones).
That could be due to a combination of deconvergence, overexposure and lack of focus.
So in a photo that had lower exposure and was sharper in focus, the phosphors might appear a lot tighter with possible separation in between each of them.
Sega Rally 2 Arcade CRT:
Sorry for the late reply! These are fantastic images - its really great to see high quality pictures that both show the overall screen and detail down to the phosphors. Its a great resource for future developers to really get the look of these old displays. One thing Id like to see is some how record what inputs were used into the display - maybe recorded in the image name i.e RF, composite, rgb, s-video etc. This is really important information in figuring out what is the display and what is the signal for a cettain artifact. Thanks again for these great shots!
EDIT: actually it looks as though you have recorded the signal - appologies great work!
Fantastic shots if these Sega arcades - Im going to try Sega rally on an emulator with the Sony Megatron Virtua Racing preset - see the first post for that shader for screenshots. Theres probably a bit of eork I need to do to fully match these screenshots though.
From my old, scary and power hungry cab. I manually did the shots and modified exposure and temperature; what you see is almost the same as you would see on the real screen.
–AND FINALLY I can show you the infamous pattern seen by my eyes when i don’t wear my glasses when shaders emulate slotmask on 1080p; it turns out those shots show it when zoomed out, and in turn is the proof shaders are pretty accurate lol!
Look for disturbing vertical - slightly brighter - lines:
shadow mask phillips 107e5 17" monitor 640x480 resolution for MS-DOS
600x240 at 120hz on Retroarch running on Linux Mint on the same monitor. No shaders.
running Ocean Hunter at 496x384
Gzdoom+Vkquake+YamagiQuake
These are fantastic! Thanks! Its not something Ive thought too much about connecting a pc to a crt and playing retroarch through that. Something to try thanks! If you get the chance can you take some close up shots that show the phophors in a bit more detail
thesw are great pics - really good to see the close up shots that show what the phosphors are up to! thanks!!!
My time to shine , these are 2560x240 in resolution sent through scart cable to rgb scart thomson crt tv as follow in the diagram
For this method , the switchres should be off and fullscreen should be windowed fullscreen
Great shots!
Just noticed this: https://forums.libretro.com/uploads/default/original/3X/1/2/12178cb7f022014e296160a8101fb9a1c1cb4776.jpeg
Hence the question, why does the slotmask gap disappear when the phosphor is not excited?
Maybe it has something to do with electron beam “mask” rather than with phosphor coating.
I’m afraid I don’t understood.
Something specific to the model and how “much” the phosphors are separated, in other words the phosphors gap is not really pitch black here?
In my shots, I can spot the mask almost everywhere:
Because there is no slotmask gap on the phosphor stripes to begin with. The phosphor stripes are vertical stripes of continuous phosphor lines on the inside of the glass. In front of the glass on the inside there is a metal plate with holes in it through which the electrons pass. The slotmask pattern is actually a “shadow” of the metal plate where the electrons cannot pass because they hit the metal. That’s why the concept is called a “shadow mask”.
See below picture from the excellent book on CRTs from Kenneth Compton Image Performance in CRT Displays:
This picture shows the “dotmask” shadow mask and aperture grill principle. Slotmask is like dotmask shadow type but different shape metal plate to increase brightness versus dotmaks. If you want to see an example of the metal plate for slotmask look up Chromaclear.
Before the electron beam hits the phosphors, it’s guided trough a grid.
The grid looks something like this with slot mask crt’s:
