FPGA Based Videogame System

[Andromeda Stardust]

Someone was actually working on an NES flashcart that would also run Gameboy (DMG?) ROMs...I think it was Kevtris?

Maybe you're talking about something else, but bunnyboy does have this on his site:

http://www.retrousb.com/product_info.php?cPath=30&products_id=87

Retrovision (essentially an NES version of the Super Game Boy) has been OOS for years. I think they're a royal pain for him to build, and what with the AVS rollout, he's been very busy lately...

 

It will be interesting to see how the Analogue NT Mini compares to the AVS. Both are running pure FPGA implementations. RetroUSB's AVS is Brian Parker's aka Bunnyboy's own design and the NT Mini will be using Kevtris' design. The Kevtris core seems to support 720/1080p, more/better AV and OC tweaking options, simulates expansion audio, and can access system menu without rebooting the game. The RetroUSB core uses 720p only, built in Game Genie, an ADC to mix in expansion audio, cannot access system menu without rebooting the game, and connects to custom PC interface via USB to dump high score data to NintendoAge using the Scoreboard interface.

 

Overall both are attractive options but the AVS wins in the price department, $185 msrp over the NT Mini at $450! My advice to Analogue would be to drop the expensive Aluminum milling for a plastic enclosure.

 

Another thing I found fascinating was one reviewer posted a teardown photo of the AVS which revealed the NES cartridge connector used a PCIe slot to connect to the mainboard. Kevtris planned on using a similar interface for Zimba3000 cartridge add-ons. Coincidence?

[veelk55]

My advice to Analogue would be to drop the expensive Aluminum milling for a plastic enclosure.

 

That's the first thing that came to mind. I still dont understand the need or appeal of this, even if I do like fancy things for the sake of being fancy sometimes.

This thing couldn't even go through airports without causing trouble. And its obviously not gonna sell to the masses. I assume that that's just part of their core idea for the NT, to be the most expensive NES ever.

Im still waiting to see how the AVS' 720p signal will behave on my 1024x768 TV. That's about the ONLY thing that will tell what is what for me. And if the AVS is bad for me where the Analogue NT would be better, I still wouldn't pay for it because I'm not willing to go the Analogue NT way for the rest of the consoles that would get reproduced if they would, but I WOULD have a system similar to the AVS but for SNES and others, judging by price, design and no aluminum.

Edited August 27, 2016 by veelk55

[keepdreamin]

 

 

9ms is more than half a frame, which is noticeable to me. I can't go from a CRT to any modern display. I can even notice lag on the wireless remotes on the wii when hooked up to a CRT. I'm pretty sensitive to this stuff though.

 

I don't believe this at all.

 

who are you, Barry Allen?

 

What next, you're gonna say you can perceive the lag added by using a 6 foot controller extension?

[narshe]

 

I don't believe this at all.

 

who are you, Barry Allen?

 

What next, you're gonna say you can perceive the lag added by using a 6 foot controller extension?

 

Try it. Get really good at a game, then switch from a CRT to an LCD with very low lag. See what happens. I could even feel the lag in the wireless controller of a Wii when hooked up to a CRT. Wii Virtual Console games are unplayable for me.

 

Now, this is for games I've played before. If I play a completely new game, I have no muscle memory or expected response time, and it doesn't matter much.

 

I speedrun games which is why I'm extra sensitive to this stuff. Strats that involve frame perfect inputs become impossible to do when your input is behind by a couple frames.

[brentonius]

Try it. Get really good at a game, then switch from a CRT to an LCD with very low lag. See what happens. I could even feel the lag in the wireless controller of a Wii when hooked up to a CRT. Wii Virtual Console games are unplayable for me.

 

Now, this is for games I've played before. If I play a completely new game, I have no muscle memory or expected response time, and it doesn't matter much.

 

I speedrun games which is why I'm extra sensitive to this stuff. Strats that involve frame perfect inputs become impossible to do when your input is behind by a couple frames.

I believe you. Though I'm not sure I can feel 9ms of input lag, I am a sound editor and can definitely notice a frame or 9ms lag between sound and image (especially with dialogue). So it doesn't stretch the imagination at all to believe that someone can feel 9ms. For a while I had both a CRT and an HDMI TV. Playing the original NES was pure joy on the CRT. It felt so good. Going back to my HDMI there was definitely lag. I hadn't noticed it until I got the CRT. After doing a test to see how much lag there was I discovered there was zero lag with my CRT and 35ms lag with my HDMI. Although I no longer have a CRT, and 35ms doesn't in my opinion make a game unplayable, it is sure better to have zero lag. Really it's amazing how differently the game responds with such a "small" amount of lag. And people have different sensitivities to lag. A friend of mine can't believe that I can feel 35ms lag... But I most definitely do. My ability to play NES games is never as good on my HDMI tv. I even remember thinking "I was so much better at this when I was a kid." Well, not really. On my CRT I was just as good. My silly friend uses a PC emulator that goes through a projector... 200ms lag minimum there. I was astonished... Yet he thinks he can't feel any lag at all... But for me, that would make a game totally unplayable. For him he somehow is used to it. Edited August 27, 2016 by brentonius

[veelk55]

 

Now, this is for games I've played before. If I play a completely new game, I have no muscle memory or expected response time, and it doesn't matter much.

 

And then you play the original game in its original lagless environment and you feel ripped off and never want to trust anything ever again but a CRT with original hardware .

[brentonius]

And then you play the original game in its original lagless environment and you feel ripped off and never want to trust anything ever again but a CRT with original hardware .

I'm hoping the RetroUSB AVS turns out to be as lagless as original NES with CRTs. But I doubt it. Around 10ms or under for 720p to 1080p would make me a little sad but I could definitely live with it

Edited August 27, 2016 by brentonius

[keepdreamin]

 

Try it. Get really good at a game, then switch from a CRT to an LCD with very low lag. See what happens. I could even feel the lag in the wireless controller of a Wii when hooked up to a CRT. Wii Virtual Console games are unplayable for me.

 

Now, this is for games I've played before. If I play a completely new game, I have no muscle memory or expected response time, and it doesn't matter much.

 

I speedrun games which is why I'm extra sensitive to this stuff. Strats that involve frame perfect inputs become impossible to do when your input is behind by a couple frames.

Sorry, I don't believe you.

 

Do you have any idea how fast 9ms is? At a peak, the brain can decipher an image shown for just 13ms. http://www.dailymail.co.uk/sciencetech/article-2542583/Scientists-record-fastest-time-human-image-takes-just-13-milliseconds.html

 

Yet you claim to be noticing lag even faster than that. I also don't believe your reflexes are so highly tuned to fall within a 9ms window of error.

[Andromeda Stardust]

My silly friend uses a PC emulator that goes through a projector... 200ms lag minimum there. I was astonished... Yet he thinks he can't feel any lag at all... But for me, that would make a game totally unplayable. For him he somehow is used to it.

I wouldn't call your friend stupid. Sometimes you don't notice what you're missing. Your friend hasn't played on anything but laggy PC emulator, and he probably abuses save states so if he screws up, he can just rewind until he gets it right. If he had the opportunity to feel how responsive the NES hardware is on a CRT, he would likely ditch the emulator crap. Probably also fuss that he can't "redo" something when he screws up!

[keepdreamin]

and just to put this in perspective, the guy claims he is perceiving lag almost 2x faster than the actual refresh rate of the screen.

Edited August 27, 2016 by keepdreamin

[narshe]

and just to put this in perspective, the guy claims he is perceiving lag almost 2x faster than the actual refresh rate of the screen.

 

It's not a matter of how fast my brain can react. It's that I've been playing a game with a certain amount of lag, we'll say 0 for a NES console hooked up to a CRT. I've built up muscle memory for a game. You introduce 1 frame of lag. When a character, say mario, is running, he's moving more than 1 pixel per frame. If you're doing things that require a couple pixel precision, you won't be able to do them, until you get used to the slight lag that was introduced. It has nothing to do with your brain perceiving it. You're just used to jumping from a specific point, and now you need to jump slightly earlier (as an example). It's purely an issue of muscle memory.

 

What is your experience with switching from a CRT to an LCD?

[keepdreamin]

 

It's not a matter of how fast my brain can react. It's that I've been playing a game with a certain amount of lag, we'll say 0 for a NES console hooked up to a CRT. I've built up muscle memory for a game. You introduce 1 frame of lag. When a character, say mario, is running, he's moving more than 1 pixel per frame. If you're doing things that require a couple pixel precision, you won't be able to do them, until you get used to the slight lag that was introduced. It has nothing to do with your brain perceiving it. You're just used to jumping from a specific point, and now you need to jump slightly earlier (as an example). It's purely an issue of muscle memory.

 

What is your experience with switching from a CRT to an LCD?

 

I'm not debating that there will be varying degrees of change going from a CRT to LCD.. What I am strongly not believing is your claim to notice a delay as fast as 9ms. And that 9ms has detrimental effects on your gameplay. That is absolutely absurd. You don't have super powers. Your muscle memory isn't that insanely accurate. 9ms is faster than a CRT is even putting up pictures on screen. I've seen a few of these similar super reaction skills claims from other speed runners before. I think you guys are drinking a bit too much of your own kool-aid.

 

Personally, I don't find myself noticing input lag until it's up around 50ms.

Edited August 27, 2016 by keepdreamin

[veelk55]

I don't exactly notice such small input lag (like wireless lag) unless I have a CRT next to it with a console running the same game that has said input lag elsewhere, wired (assuming the one playing on a CRT is running perfectly and not through another emulator of course).

 

On a CRT with a game like Super Mario Bros. or any other game which responds instantly to your button inputs, I feel like "Hey, the game is reacting before I even press the button!",

It almost feels as if the game is predicting what I want to press. When it feels like this, I know that's the smallest amount of input lag possible. This is what I call "0 lag".

Would there still be like less than 5 ms? I wouldn't know. The only real way to know how much input lag it takes for me to feel it, is if an NES cartridge was made to simulate adjustable response rates by MS.

 

I can tell you that I can feel input lag from PC emulators, but I have no idea how much they produce, and they seem to vary, too, depending on the controller and screen resolution and other options and emulator itself. I can at least feel some respond faster than others, but none ever as quickly as the "0 lag" above, even if they can be close enough to not complain.

 

I can play Super Smash Bros WiiU on an HDTV through the GamePad, And while I know that the GamePad has wireless lag, it's never enough to detect it ever. It can pass for "0 lag" to me.

Edited August 28, 2016 by veelk55

[Andromeda Stardust]

GamePad is 16ms lag. This has been confirmed by Nintendo. Anyone who claims they can detect less than one frame of lag has been "drinking their own koolaid," as keepdreamin put it.

[narshe]

GamePad is 16ms lag. This has been confirmed by Nintendo. Anyone who claims they can detect less than one frame of lag has been "drinking their own koolaid," as keepdreamin put it.

 

How much does the WiiMote have? I can definitely feel that compared to having a controller hardwired to the Wii. I would assume the GamePad has less lag than a WiiMote. Then there something you can try. Plug a controller into the Wii and try it, then use the WiiMote. I'd be interested to see if you guys can feel the difference, because I think it's fairly prominent.

[Andromeda Stardust]

 

How much does the WiiMote have? I can definitely feel that compared to having a controller hardwired to the Wii. I would assume the GamePad has less lag than a WiiMote. Then there something you can try. Plug a controller into the Wii and try it, then use the WiiMote. I'd be interested to see if you guys can feel the difference, because I think it's fairly prominent.

Some non-scientific empirical evidence: In the western shooter minigame on Kirby's Adventure, I typically get .13 seconds on a real NES. On the Wii-U virtual console played through the gamepad, I get .15 seconds reaction time. The 5th boss you need to get .15 seconds or less, so it's a crapshoot if I beat him or not on Wii-U VC, but I can get him every time on NES.

[keepdreamin]

Some non-scientific empirical evidence: In the western shooter minigame on Kirby's Adventure, I typically get .13 seconds on a real NES. On the Wii-U virtual console played through the gamepad, I get .15 seconds reaction time. The 5th boss you need to get .15 seconds or less, so it's a crapshoot if I beat him or not on Wii-U VC, but I can get him every time on NES.

 

More non-scientific testing. Running my SNES through a Framemeister (already around 20ms of lag) to a tv with lag in the mid 20's, using 8bitdo's just released bluetooth wireless receiver and their SNES30 pad... I can get between 6 and 10ms of lag on the manual lag test using random timing. Go figure.

[Andromeda Stardust]

 

More non-scientific testing. Running my SNES through a Framemeister (already around 20ms of lag) to a tv with lag in the mid 20's, using 8bitdo's just released bluetooth wireless receiver and their SNES30 pad... I can get between 6 and 10ms of lag on the manual lag test using random timing. Go figure.

Any test that relies on human response time on top of actual display + input lag is going to be inaccurate. With the Kirby test, the game measures in 1/100 seconds but the NES outputs 60fps. It likely does not poll the controller frequently enough to measure response any finer than .016 seconds, so with Kirby the difference between the .15 and .13 being 20 milliseconds could easily be rounding error for 16ms per frame. Nintendo is serious about reproducing classic game experiences so it is likely their emulator could be faster on dedicated hardware compared to traditional PC emulators which rely on video buffers.

 

Anyway I have no reason to doubt the 16ms statistic when my unscientific tests confirm it.

[SnoopKatt]

One thing to keep in mind with input lag is you have to know how the electronics handle input in the first place. For my senior design project, I was planning on making a wireless NES controller from scratch, with the idea that the final design would simply be a drop-in replacement PCB with a battery and a receiver that would plug in and essentially feel and control identically to the wired controller. I didn't get to finish it because I ended up doing a different project, but the first step was to understand how the controller port works, so I probed it with a logic analyzer. What was surprising was that the NES actually polls at 60 Hz! I thought it would be much quicker, but it's not instantaneous or even on the order of hundreds of micro-seconds.

This means that the average delay is half of a 60 Hz cycle, which is 8.33 ms. Best case is close to 0 ms (pressing the button right before it polls), and worst case is 16.7 ms (pressing the button right after it polls). If the display adds some lag, it can be noticeable (and usually is!), but if you're talking about NES (and anything else that polls at 60 hz), polling is always going to keep you behind.

If your display adds 10 ms, then going from 16.7 ms to 26.7 ms might be noticeable, but do keep in mind that you're never starting from zero.

[veelk55]

So these are the 16 ms in a frame:

- - - - - - - - - - - - - - - -

 

So while a frame is happening on screen, if I press a button when the frame's screen time is here:

- - x - - - - - - - - - - - - -

that means I'm experiencing 14 ms of lag already?

 

So in the case that someone claims to feel 9 ms of lag, it would mean that only an average of half the time the game is 1 frame behind?

 

And a third question: what advantage would anything higher than 60hz polling rate offer if the game is still running at 60 fps? You're still not gonna see the change until 16 ms have passed. That is if you can slowdown time to notice that! XD (or use a high speed camera?)

[SnoopKatt]

So these are the 16 ms in a frame:

- - - - - - - - - - - - - - - -

 

So while a frame is happening on screen, if I press a button when the frame's screen time is here:

- - x - - - - - - - - - - - - -

that means I'm experiencing 14 ms of lag already?

 

So in the case that someone claims to feel 9 ms of lag, it would mean that only an average of half the time the game is 1 frame behind?

 

And a third question: what advantage would anything higher than 60hz polling rate offer if the game is still running at 60 fps? You're still not gonna see the change until 16 ms have passed. That is if you can slowdown time to notice that! XD (or use a high speed camera?)

1. Exactly.

 

2. Almost, on average the game responds half a frame behind. But the variation is crazy! If a display adds 9 ms, then you're getting between 9 ms and 25.7 ms. Granted that other things in the signal chain could add lag, but for this particular implementation, I'd say anyone claiming they can "feel" 9 ms or any small number like that are just under the placebo effect. Less is always better, and a CRT is always going to be king, but if a display only adds 9 ms, I'd find it hard to believe people can actually detect a difference. Same kind of thing happens in hi-fi audio as well people can claim whatever they want, but this kind of stuff is really easy to measure, and unless they can prove it, then there's no real basis for their claim (not trying to be a dick though).

 

3. If you have some sort of internal game timer that's measuring something in between frames, a higher polling rate would be effective. It'd be tough to justify why most games would need it, as I've never heard someone say the NES responds too slow.

[Andromeda Stardust]

1. Exactly.

 

2. Almost, on average the game responds half a frame behind. But the variation is crazy! If a display adds 9 ms, then you're getting between 9 ms and 25.7 ms. Granted that other things in the signal chain could add lag, but for this particular implementation, I'd say anyone claiming they can "feel" 9 ms or any small number like that are just under the placebo effect. Less is always better, and a CRT is always going to be king, but if a display only adds 9 ms, I'd find it hard to believe people can actually detect a difference. Same kind of thing happens in hi-fi audio as well people can claim whatever they want, but this kind of stuff is really easy to measure, and unless they can prove it, then there's no real basis for their claim (not trying to be a dick though).

 

3. If you have some sort of internal game timer that's measuring something in between frames, a higher polling rate would be effective. It'd be tough to justify why most games would need it, as I've never heard someone say the NES responds too slow.

Brian Parker or RetroUSB developed a high speed camera (1000fps) and a directly wired interface which sent controller input exactly when the controller was polled by a test ROM. The test ROM run on a Powerpak immediately flashed the screen allowing the high speed 1000fps camera to record the timing difference between the controller input (represented by bright LED) and corresponding screen changes. His AVS tests were comparable to results published by screenlag.com

 

The very best displays, typically PC monitors, can refresh the screen almost immediately following the last scanline. In fact the AVS FPGA only buffers a single scanline worth of content during 720p HDMI output scan with no screen buffer at all. This is on the order of microseconds behind a true NES composite signal because each scanline needs to be line tripled to output 720p. That is loads faster than Retron5, Retro Freak, or even any Nintendo Virtual Console service.

 

http://nintendoage.com/forum/messageview.cfm?catid=31&threadid=166023

[SnoopKatt]

Interesting... But that'll still depend on how the controller polls it too. It's been 60 hz on each game I've tested (Megaman 2,SMB, and some others), but the way they handle the data looks a little different. Might be possible that he has it polling faster.

[veelk55]

Oh wow, interesting

I talked to RetroUSB with questions about the AVS' 720p possible lag on TVs, and he sent me that same link to the thread.

 

Yeah, I think that someone claiming they can feel 9ms of lag is actually claiming they can feel 32 ms of lag since that's the only time you're gonna notice being a frame behind.
NO ONE will ever notice being 16 ms (a frame) behind if you press a button exactly after it polls on an NES.

And considering that 0 - 16 ms of lag equals my "0 lag" scenario, and my WiiU passes for "0 lag" as well (for me) after being confirmed to have an additional 16 ms of lag (which could make it 32 ms in the case you press the button exactly after it polled) then that's still insane if someone claims to feel 32 ms (when they said 9 ms)

 

But wait, there's more... isnt the WiiU's GamePad screen further behind an HDTV? (though Im sure we're talking about HDTV here) but I still think playing off of the GamePad passes for 0 lag as well (for me).

 

3rd edit:

The claim about muscle memory is not entirely scratched off either. We need to find a game that exists in an old console that also exists non-emulated (for no lag) in an HD console.

I'd say MegaMan Legacy Collection, but I think since that's emulated (high level emulation if that makes a difference) it can't be used to compare true CRT to true HD.

Only having a WiiU game non emulated (like Shovel Knight, maybe?) in a CRT and then using the HDMI cable, thats the best way to know if one can detect this 9 ms of lag difference (32 ms?)

Edited August 28, 2016 by veelk55

[Great Hierophant]

I have read that some games do not poll the controller every frame. It is really up to the programmers to determine how fast their game needs to respond to input. There is nothing to prevent a programmer from reading the memory address for the controller port more than once during a 16ms period. However, given that the screen will only change every 16ms and that people cannot process changes on the screen that quickly, there is little reason to do so. Even the motions involved in a fighting game can be captured as one frame=one button press.

 

The latency reported for the AVS is very impressive, averaging around 1.4 frames behind a CRT. The Framemeister seems to be in the same ballpark and that costs much more than an AVS and you still need an RGB-modded console to get similar visual quality.