undefined | Better HN

0 pointsjonhohle1y ago0 comments

The low latency only works when the source is sending an analog signal which is essentially controlling the electron gun. An NES for example, could decide between scan lines what the next scaling to emit would look like.

Most modern systems (and computers) are sending digital signals where entire frames need to be constructed and reconstructed in a buffer before they can be displayed. Many HD CRT televisions have terrible latency because they’ll take the analog scan line signal, buffer it into a frame, scale it to the native CRT resolution, then scan it back out to the screen. A high end PVM might allow a straight path, but there is maybe one Toshiba HD CRT that doesn’t introduce several frames of latency (iirc).

That said, from 1999 to 2008 I ran 1600x1200 on 19in CRTs and except for professional LCDs, nothing had resolution, pitch, and color that came close. For 2008 was the inflection point where cost and quality of LCDs exceeded CRTs.

0 comments

toast01y ago

> Most modern systems (and computers) are sending digital signals where entire frames need to be constructed and reconstructed in a buffer before they can be displayed.

HDMI and DVI-D are a digital representation of a tv signal --- there's all the blanking and porches and stuff (audio data is often interleaved into the blanking periods). You could process that in realtime, even though most displays don't (including CRT HDTVs; I trust your maybe one doesn't).

cmrdporcupine1y ago

What you're describing is fully a product of the display controller architecture, not the digital signal. You can "chase the beam" with HDMI/DVI just as well (or better) as you can with analog. There's no need to buffer the whole screen and be even one frame latent.

I've done it in Verilog on FPGA, for example.

But we do so on our machines because that's how the GPU/display controller pipeline works easiest from a software POV. That latency would be present on a CRT as well. What would be missing is the internal latency present in some monitors.

mabster1y ago

I was doing the "chase the beam" approach, but never got it off the ground (home project, mostly curiosity). But always on the back of my mind was: even if you're doing everything in low latency HDMI signals, there's no guarantee that the display isn't still buffering it, even if you have "game mode" on on your TV.

cmrdporcupine1y ago

yes, absolutely, the monitor vendor can do whatever they want

still, it's usually only 2 or 3 frames in game mode on a good display

1 more reply

tverbeure1y ago

It makes little sense to buffer a full frame before up scaling. Why would you do that? It’s a total waste of DRAM bandwidth too.

The latency incurred for upscaling depends on number of the vertical filter taps and the horizontal scan time. We’re talking order of 10 microseconds.

The only exception is if you’re using a super fancy machine learning whole-frame upscaling algorithm, but that’s not something you’ll find in an old CRT.

j / k navigate · click thread line to collapse

0 pointsjonhohle1y ago0 comments

0 comments

toast01y ago

> Most modern systems (and computers) are sending digital signals where entire frames need to be constructed and reconstructed in a buffer before they can be displayed.

cmrdporcupine1y ago

I've done it in Verilog on FPGA, for example.

mabster1y ago

cmrdporcupine1y ago

yes, absolutely, the monitor vendor can do whatever they want

still, it's usually only 2 or 3 frames in game mode on a good display

1 more reply

tverbeure1y ago

It makes little sense to buffer a full frame before up scaling. Why would you do that? It’s a total waste of DRAM bandwidth too.

The latency incurred for upscaling depends on number of the vertical filter taps and the horizontal scan time. We’re talking order of 10 microseconds.

The only exception is if you’re using a super fancy machine learning whole-frame upscaling algorithm, but that’s not something you’ll find in an old CRT.

j / k navigate · click thread line to collapse