Advertisement
Advertisement
⚡ Community Insights
Discussion Sentiment
72% Positive
Analyzed from 1538 words in the discussion.
Trending Topics
#latency#crt#games#game#input#more#snk#things#timing#oled
Discussion Sentiment
Analyzed from 1538 words in the discussion.
Trending Topics
Discussion (32 Comments)Read Original on HackerNews
Then problem #2 with the display (mostly) is latency. Those CRTs were fast. Even 50ms of rendering latency is noticeable on a some of the console games that require very-precise input timing.
You get emulation latency (this may avoid that by using ASICs, at least); input latency above what the original hardware had, if you're not using the real thing (bluetooth...); any picture-conversion latency (this might avoid that, but I wouldn't bet on it) to digitize the signal into HDMI if you're working with real hardware with analog outputs; TVs that struggle to get under 50ms of latency, especially without making the picture look a ton worse; and then shader-induced latency if you're trying to make it look semi-correct. Like, getting it down to where it doesn't feel wrong is tricky as hell.
If you have an HDR TV, preferably OLED, and miss the CRT look, check out the RetroTink 4K https://www.retrotink.com/
With something like the MiSTer, you can also enable high speed USB polling, which I believe is roughly 1000hz. My understanding is that it doesn't work with all controllers, but it has worked with all the controllers I have tried it with.
The composite video artifacts are definitely noticeable though; I noticed the weirdness of the waterfalls in Sonic when I was playing it recently. It doesn't bother me that much but I could see why it bothers other people.
That’s an oversimplification. Many retro game consoles don’t use a frame buffer. Instead they render the game state to the screen on the fly, one scanline at a time, and they’re able to process input mid-screen because they read the controller input many times faster than 60Hz (on the order of 2kHz). In practice, this means input lag is way below even 1ms.
Lightgun games, for example, rely on very precise timing of the control input vs the CRT raster and simply do not work without a CRT.
Perhaps the most famous light gun game of all time (Duck Hunt on the NES), does not rely on especially precise timing. It draws one white rectangle per frame over each duck when you pull the trigger and checks if the Zapper can see it. LCD latency will probably still break this, but it's not like the later Super Scope for the SNES that actually does track the precise raster position. I expect it would be possible to patch the timing in software to make it work for a specific model of LCD. But even if you did this, the Zapper also includes a bandpass filter at the CRT horizontal retrace rate (about 15kHz) to better reject other light sources, so you'd need to mod it to bypass that, or mod the LCD to strobe the backlight at the right frequency.
Retro arch has run ahead latency reduction etc, I'd like to see some comparisons of that Vs mister. I could do it myself but I've never got round to it. I've noticed that fiddling with latency reduction in retro arch really works, but it is a lot of fiddling.
As John Carmack said once, we can send a data packet across the atlantic faster than we can get a pixel out the back of a computer nowadays.
Displays are no longer the problem anymore, we're back to CRT speeds again.
That only worked because they expected to run over composite. Arcade cabinets used RGB which doesn't have the bandwidth limitations of composite.
And that's the problem with current pixel art artists: they have no idea of what actual pixel art looked like. Hint: look at Garou with at least scanlines (or maybe a bilinear filter) enabled. That's what's Garou almost meant too look in CRT, far closer than raw pixel art.
If you're talking about the waterfalls, I'm not convinced blurring was necessary or intended. RGB support was rare in televisions in the USA, but it was common in PAL regions via a SCART cable, and the Mega Drive had native RGB output. Furthermore, the waterfalls are drawn as vertical lines, which I interpret as representing individual streams of water. If it was purely a pseudo-transparency effect it would make more sense to use a checkerboard pattern, e.g. as in the spotlight effect in Streets of Rage 2.
I think I'll pass giving them a single cent.
Makes me feel a little conflicted having bought one of those SNK bundles a couple years ago on Steam or Humble Bundle or something. Don't love the idea of giving money to someone like that.
He won't get a cent when I play on my MVS carts on actual hardware, so there is that.
It’s not. It sounds much more like a vanity or passion purchase. Like how Neil Young once invested in the Lionel model railroading manufacturer.
The biggest concern I have is accuracy. SNK wouldn't be able to just start manufacturing their old chip designs from the 90s again because a lot of the chips in the Neo Geo from other companies are no longer made, such as the 68000, Z80, and YM2610. This means that they'd have to make a new SoC that incorporates the IP from those chips. At that point there's no real benefit to the ASIC over an FPGA. It means the system costs less to produce, but if they find any inaccuracies in their new SoC design they won't be able to release an update to fix them. I'm cautiously optimistic, but I'm not going to place a pre-order until they release more details about what exactly is going on under the hood.
If instead one want the real HW, a working MVS PCB can be found for less that €100. A JAMMA cabinet would be the perfect place to use it, but with a supergun a CRT TV can give the same results. Cheap superguns can be found or built for around €30, fancy models can cost between €200 to €300 tho.
The carts aren't cheap tho, specially if original. I've only a battered "Puzzle Bobble" cart, it's a really fun game in PvP, but probably not the more iconic NeoGeo title.