VOGONS


Reply 200 of 211, by Benedikt

User metadata
Rank Member
Rank
Member
root42 wrote on 2020-08-30, 22:23:

You can use some software to re-encode your capture as 4k and upload that to YT. It will force YT to use higher bitrates and resolution, leading to better visual quality. Even if your original material wasn't 4k to begin with.

I can't get over the wastefulness that video platforms provoke with their lack of configuration options.

Anyway, has anyone given pixel shaders a thought? While I have no experience writing pixel shaders, they could potentially make the filtering process lightning fast even on slightly older systems.

Reply 201 of 211, by root42

User metadata
Rank Oldbie
Rank
Oldbie

Yeah, color conversion possibly could be done in a pixel shader... we could even support RGBrgb that way.

By the way, my OSSC also has significant jitter on my Tseng VGA when running in 800x600 and a little bit when running in 640x480. I think the digital sampling really shows the imperfections of the old video cards. A CRT probably smoothes this out with its sync circuits.

YouTube and Bonus
80486DX@33 MHz, 16 MiB RAM, Tseng ET4000 1 MiB, SnarkBarker & GUSar Lite, PC MIDI Card+X2+SC55+MT32, OSSC

Reply 205 of 211, by Predator99

User metadata
Rank Oldbie
Rank
Oldbie

SIGROK2EGA at its best...now I am happy with the Youtube-Quality...select 1080p HD 😉
https://youtu.be/FvCjBuS9NBc
Sound is from a Crystal/Dream Wavetable card I am currently testing, also sounds amazing.
SC8500 sound card - what does it do?

Reply 206 of 211, by Pierre32

User metadata
Rank Member
Rank
Member
Predator99 wrote on 2020-09-05, 20:35:
SIGROK2EGA at its best...now I am happy with the Youtube-Quality...select 1080p HD ;-) https://youtu.be/FvCjBuS9NBc Sound is fro […]
Show full quote

SIGROK2EGA at its best...now I am happy with the Youtube-Quality...select 1080p HD 😉
https://youtu.be/FvCjBuS9NBc
Sound is from a Crystal/Dream Wavetable card I am currently testing, also sounds amazing.
SC8500 sound card - what does it do?

Looks spectacular! Amazing progress.

Reply 207 of 211, by Predator99

User metadata
Rank Oldbie
Rank
Oldbie
Pierre32 wrote on 2020-09-06, 07:55:

Looks spectacular! Amazing progress.

Thank you!!!

2 more Fullscreen videos wih sound:
Commander Keen 5
https://youtu.be/0PQi3vlXOyw

Wing Commander 1
https://youtu.be/apsMpAtMLMI

Reply 208 of 211, by kdr

User metadata
Rank Newbie
Rank
Newbie

I've recently ordered a couple of cheap Cypress FX2LP eval boards from China (so they'll take ages to arrive) because I need a 16ch logic analyzer to debug an old Turbo XT motherboard with glitchy DMA. I'm also hoping to be able to play with Predator99's awesome video capture code!

In preparation I have been digesting the FX2LP datasheet and reference manual. Gosh it's a complicated device! But it's given me some thoughts on how to improve the video capture, provided you have access to some of the extra pins on the FX2LP chip (RDY0/1). Basically the idea is to route HSYNC to one of the RDYx inputs and use this as a hardware trigger to begin/end the capture of a single scanline. During one scanline you are still reliant on the internal 30/48Mhz clock to capture the pixels [so there will be some drift] but the hardware can automatically resync at the start of each new line, so the drift would be roughly identical for all lines.

As a bonus, there would be no need to transmit any data during horizontal blanking, reducing USB bandwidth requirements and letting the capture speed get cranked up...

Reply 209 of 211, by root42

User metadata
Rank Oldbie
Rank
Oldbie

That would be excellent news, if possible. Could become a very nice capture device with little visible jitter.

YouTube and Bonus
80486DX@33 MHz, 16 MiB RAM, Tseng ET4000 1 MiB, SnarkBarker & GUSar Lite, PC MIDI Card+X2+SC55+MT32, OSSC

Reply 211 of 211, by Benedikt

User metadata
Rank Member
Rank
Member
kdr wrote on 2020-10-15, 01:44:

In preparation I have been digesting the FX2LP datasheet and reference manual. Gosh it's a complicated device! But it's given me some thoughts on how to improve the video capture, provided you have access to some of the extra pins on the FX2LP chip (RDY0/1). Basically the idea is to route HSYNC to one of the RDYx inputs and use this as a hardware trigger to begin/end the capture of a single scanline. During one scanline you are still reliant on the internal 30/48Mhz clock to capture the pixels [so there will be some drift] but the hardware can automatically resync at the start of each new line, so the drift would be roughly identical for all lines.

In that case, it's probably time for custom boards.
A small PCB with an edge-mounted DE9 connector on one end and micro USB on the other would fit nicely in a shielded standard enclosure for D-Sub adapters.