Using Free Sync

[davidelkinsrice]

Hi All,

I'm pretty new to MiSTer fpga. I am a pacman junkie and am trying to get a pacman rom to run at 60.6 hz. @macro commented in another post that you could update two settings in the rtl/pll file to get a true 60.606060hz pacman running. I have a monitor that has free sync running. When I try to run the .rbf from the alternatives directory (that @macro made for me) it just gives me a black screen. @macro said that it works on his CRT monitor. Has anyone here had any success using free sync setting in a monitor? I have tried the vsync_adjust settings in the 0,1,2 and none of them work.

; Set to 1 for automatic HDMI VSync rate adjust to match original VSync.
; Set to 2 for low latency mode (single buffer).
; This option makes video butter smooth like on original emulated system.
; Adjusting is done by changing pixel clock. Not every display supports variable pixel clock.
; For proper adjusting and to reduce possible out of range pixel clock, use 60Hz HDMI video
; modes as a base even for 50Hz systems.
vsync_adjust=2

One thing I wonder about is if I should use custom mode:
; custom mode: hact,hfp,hs,hbp,vact,vfp,vs,vbp,Fpix_in_KHz
; video_mode=1280,110,40,220,720,5,5,20,74250
video_mode=0

I don't know what these settings are. Does anyone here know what they are?

This is the monitor I have: https://www.amazon.com/gp/product/B07HQ ... UTF8&psc=1. When I look at the settings of the monitor it says that it is running 1280x768 60hz (720p).

AMD FreeSync: This 75Hz monitor provides smooth visuals and reduced screen tearing for your gaming experience, thanks to AMD FreeSync technology. Enjoy a seamless gaming experience with minimal stutter and lag.

Thank you for any help.

David

[TheJesusFish]

Scroll down in the ini. There is a VRR setting there. You'll want to set it to 2 and then set the maximum FPS to 75.

[davidelkinsrice]

I tried that and it still didn't work. I also tried to update the vrr_max_framerate to 75 (and vrr_vesa_framerate to 75 - I got an error on the screen with that one). I tried with just refresh_max rate at 75 with vsync_adjust=0,1 and 2(0-match freq of display, 1 match freq of core, 2 low lag) with all of those configurations. I wonder if someone out there with more experience could try to get this rbf file to work on the their monitor with freesync? And if you get it to work how you did it? Here is the .rbf created by @macro:

https://drive.google.com/file/d/1SMyDzt ... sp=sharing

Thank you for any help.

David

[davidelkinsrice]

Am I really the only user on this forum that has a newer monitor with FreeSync. I don't believe it. MiSTer FPGA is made to imitate "exactly" the PCBs with and or gates not emulating like Mame. There isn't a better test for MiSTer to be proved that it is identical to the original PCB than being able to do PacMan patterns that Ken Uston created in his PacMan book from the 1980s. Doesn't this idea intrigue someone out there enough to get the .RBF file that @macro a core developer on this website created to work on their MiSTer system. I would really appreciate it if someone would try.

[Retro-Nerd]

Try video_mode=8

which is your native monitor resolution. The setting vrr_mode=2 is correct for freesync monitors. You can also try to add this:

vrr_min_framerate=50
vrr_max_framerate=75

[rhester72]

tbh I don't think freesync is going to give you as accurate a result on refresh (at least for what the core spits out) than vsync_adjust=2.

I also don't think anyone ever claimed it's "identical" to the PCB. =) The only three cores I'm aware of taken directly from netlists or decaps are Genesis /Megadrive (the new one), Neo Geo, and Jaguar. Everything else (especially arcades) are largely MAME-derived approximations.

[Retro-Nerd]

Freesync surely helps to stabilise the whole image output for arcade games on specific monitors/TVs. On LG displays(maybe others too) it reduces the display lag and triggers a low latency mode, especially with odd refresh rate games. You can measure it with Mister Laggy (got 10ms less lag in e.g. 55-57Hz games).

[davidelkinsrice]

I tried those things:
video_mode=8
vrr_max_framerate=75
vrr_min_framerate=50

@macro's core still isn't working. Funny that this machine that I bought isn't even as accurate as a Mame emulator. Why would anyone want to purchase the hardware if it isn't even as good as an emulated software?! I bought it because I thought I was getting as close to the actual PCB as I could get. I guess not. The guy who got all of the points in Pac-Man used Mame 0.187 Pacman to complete all of the boards and get all of the points. I assumed I could do it with MiSTer. I guess not.

[rhester72]

Freesync surely helps to stabilise the whole image output for arcade games on specific monitors/TVs. On LG displays(maybe others too) it reduces the display lag and triggers a low latency mode, especially with odd refresh rate games. You can measure it with Mister Laggy (got 10ms less lag in e.g. 55-57Hz games).

Resolution matters. What res did you test with? (I'm asking because while I acknowledge that freesync could in theory be lower latency, in practice I'd very much like to see that for myself.)

[rhester72]

@macro's core still isn't working. Funny that this machine that I bought isn't even as accurate as a Mame emulator. Why would anyone want to purchase the hardware if it isn't even as good as an emulated software?! I bought it because I thought I was getting as close to the actual PCB as I could get. I guess not. The guy who got all of the points in Pac-Man used Mame 0.187 Pacman to complete all of the boards and get all of the points. I assumed I could do it with MiSTer. I guess not.

What information did you come across that made you think MiSTer would be "as close to the actual PCB as you could get"?

FPGA is primarily good at two things, low lag (because there's no operating system layer in between the implementation and the user to drive lag) and hyperparallelism (for example, when emulating the Amiga custom chipset - while software emulators have to split time to simulate parallel operation, FPGA is actually parallel), which can sometimes lead to lower-powered FPGA solutions outperforming high-powered PC gaming rigs.

All that being said, everything about the accuracy and fidelity of any emulator comes down to the implementation, which can be equally good (or bad) in hardware or software.

This was explored pretty extensively by byuu/Near (RIP) years ago (https://archive.ph/4Umsk), and has been discussed pretty continuously discussed ever since (in fact, many times on this very forum).

The Pacman core for MiSTer itself is derivative of another FPGA core, which describes itself as "A simulation model of Pacman hardware" (https://github.com/MiSTer-devel/Arcade-Pacman_MiSTer), the word 'simulation' being very key.

If you're referring to statements from Analogue, please bear in mind they are a commercial enterprise shilling game systems for profit. They are very frequently rather hyperbolic with their "100% perfection due to use of advanced FPGA emulation", yadda yadda yadda, but to be very frank it's a damned lie and they know it. In fact, a brief perusal of the most popular Analogue support forums demonstrates very easily that, by and large, they experience more incompatibility than most popular software emulators (and certainly so on initial release before they get most bugs worked out, but to this day there are still incompatible NES and SNES titles on their platforms).

In short, if you allowed their marketing to give you a false sense about what FPGA is (and more importantly, isn't), I'm sorry, but MiSTer itself has never laid claim to false statements about accuracy. As I stated earlier, there are cores that are very, VERY accurate...and there are some (Atari 8-bits, for example) that are decidedly...not (at least not yet). Things are always improving, but will very rarely achieve 'perfection' by any definition of the word. Cores are never finished, merely abandoned, just like all implementations.

As to the 'why buy it at all?' question - I can't speak for everyone, but my reasoning was that it's pretty cool LOL, it is low lag and in many cores supports native controllers through SNAC (which drastically improves playability), it has a very wide base of cores for practically any interests out there (and has excellent implementations for obscure machines like the TI-99/4A, near and dear to my heart), and fits two decades of retrogaming and retrocomputing happiness in a box the side of a few decks of playing cards stacked atop one another. What's not to love? It's not cheap, and the barrier to entry in terms of setup and ease of use can be a bit intimidating for newcomers, but overall I think MiSTer has done more to bring FPGA to the 'masses' (tiny group as we may be in comparison) than anything else ever...including the original MiST platform from which it sprang. The brilliance of MiSTer is Sorgelig's visionary approach of using common templates (called a 'framework') that makes it relatively easy for FPGA developers to port cores from other platforms and create new ones from scratch while providing them with common functionality and look-and-feel. There's really nothing else like it.

[rhester72]

Freesync surely helps to stabilise the whole image output for arcade games on specific monitors/TVs. On LG displays(maybe others too) it reduces the display lag and triggers a low latency mode, especially with odd refresh rate games. You can measure it with Mister Laggy (got 10ms less lag in e.g. 55-57Hz games).

Resolution matters. What res did you test with? (I'm asking because while I acknowledge that freesync could in theory be lower latency, in practice I'd very much like to see that for myself.)

Actually, I did find this thread:

viewtopic.php?t=8083

but my experience is closer to KennyL's than your own. I get about 9ms (center) on my Acer XF243Y at 1920z1080 with vsync_adjust=2 and Ultra-Low Latency enabled on the display, and effectively no change at all with FreeSync enabled (tried with VRR=0 and VRR=2).

[akator]

In short, if you allowed their marketing to give you a false sense about what FPGA is (and more importantly, isn't), I'm sorry, but MiSTer itself has never laid claim to false statements about accuracy.

True, but there have been a lot of MiSTer fans that have made that claim. I paid attention to MiSTer and advocates for years before finally buying it. One of the things that I remember vividly is several people saying they got rid of most of their original hardware because they didn't need it anymore because MiSTer was just as good, if not better.

Like an idiot, I had doubts but allowed those claims to boost my expectations. When I finally had a MiSTer assembled and setup, I was very disappointed. In no way did I feel that MiSTer could replace any of my original systems. It actually made me repurchase original 2600 hardware and gradually some other systems. But that was my fault for believing those who made that claim. There are some cores that definitely replace original hardware for me, and some not.

That original disappointment was replaced by love for MiSTer and what it does well. It's truly and incredible system and I am very thankful for all of those who put in the hard work to make it a reality.

It was my fault for listening to and believing even some of those proclamations, but I can't help but wonder how many others have had a similar experience and been falsely disappointed because of unrealistic expectations created by the over-exuberance of MiSTer's online fans. (Of course, that can happen with anything these days.)

[akeley]

It was my fault for listening to and believing even some of those proclamations

I can't say I agree with this sentiment. In my opinion it's totally on all the people who made these statements. Which, more often than not, were made with conviction so zealous, fanatical, and downright hostile to other forms of emulation, that it has actually put off many folks off trying FPGAs to start with. You can't really blame people who try to learn about something highly technical for believing stuff when a large amount of people tells them something.

I was in a similar situation to yours, doing a lot of research before buying in, because at the time the 200-250 USD needed for the most basic setup was a lot of money for me. And the assorted forums were full of this kind of proselytizing. I'm quite sure anybody who had been around at the time saw the infamous "duurrr, it's not emulation, it's Hardware Simulation 1!!11" exchange here and there at least once.

Also, the media were enamoured with FPGA, and not even only because Analogue itself, though for sure they have had a big part in building the myth as well. SmokeMonster et al propagated the "magic FPGA" claim all around, no doubt in good faith, but the truth of the matter is that whenever you tried to even ask some simple questions you risked an argument with one of the cultists.

I don't really follow this anywhere else any more, so perhaps the situation has changed a little bit for the better at least, and maybe there are some more somber and realistic takes out there. But even if, it’s probably constrained to some individuals and scattered niche sources, with language impenetrable to a layman. I'm pretty sure though it hasn't been a part of some wider conversation in the manner rhester72 suggests above.

[rhester72]

I don't really follow this anywhere else any more, so perhaps the situation has changed a little bit for the better at least, and maybe there are some more somber and realistic takes out there. But even if, it’s probably constrained to some individuals and scattered niche sources, with language impenetrable to a layman. I'm pretty sure though it hasn't been a part of some wider conversation in the manner rhester72 suggests above.

Maybe I missed it. I don't really follow the YouTube hype/advertainment train.

[davidelkinsrice]

Resolution matters. What res did you test with? (I'm asking because while I acknowledge that freesync could in theory be lower latency, in practice I'd very much like to see that for myself.)

I tested at 60hz I'll try a higher resolution if I can! Thank you for this idea. @rhester thank you so much for your thoughtful response to my question. I guess I too bought into the hype and just wish it was true. I see your point of it being a very cool way to port systems into one system and agree that it is so versatile! I agree there is nothing like the original - like buying the original Atari2600, etc. I do think it's cool. I just wish the hype was true that I "bought" when I spent $300+ for my MiSTer. I appreciate all of the comments here. You guys are cool! Have a nice day.

[akator]

Maybe I missed it. I don't really follow the YouTube hype/advertainment train.

As awesome as our 21st century is, it's difficult to get unbiased/less-biased info about things. It used to be you could go into any number of retail stores and try something for yourself to get a clue about it, but that's much more difficult these days.

I also try to avoid the YouTube/advertainment content, but there's really a vacuum of hands-on information out there about open source products unless they are being evangelized.

Anyone have any suggestions/recommendations for alternative and reliable information sources (preferably not social media)? It would be greatly appreciated.

[Retro-Nerd]

A bit of drama here, eh?

Could somebody send me this 60.60 Hz version of the pac-man rom here? I would like to try it on my Mister and LG OLED.

[davidelkinsrice]

Here's the ROM. Thank you for trying:

https://drive.google.com/file/d/1SMyDzt ... sp=sharing

If you get it to work, will you tell me what settings you used? I would really appreciate it. It was very interesting to learn many things about my MiSTer fpga. I haven't meant to cause any drama.

[Retro-Nerd]

I've connected my Mister via HDMI (in direct video mode) to the Retrotink 4K scaler. It's works perfectly fine in 60.60 Hz/Freesync here. I will test it later via Mister HDMI, directly connected to the LG OLED. I don't use any other settings in the Mister.ini as i already mentioned. Sometime the display itself can be the issue.

[davidelkinsrice]

Yeah, I don't have a retrotink 4k upscaler. I see one on ebay for $1000. I'm just using the hdmi from the Mister fpga. I like all of the data you get on your screen. That is awesome.

[Retro-Nerd]

Works fine with HDMI direct connection too in 60.6 Hz/VRR (no external scaler used).

Mister.ini attached too.

[davidelkinsrice]

Thank you! I'll use your MiSTer.ini and see if it works for me. If not, then something is different in my monitor. I am running the rom from the alternatives directory. That's not a problem, is it (where the .rbf file is stored)?

/media/fat/Arcade/alternatives/_newPacMan#
-rwxr-xr-x 1 root root 2855044 Feb 14 16:37 Arcade-Pacman.rbf

I have Free sync running on my monitor: https://www.amazon.com/gp/product/B07HQ ... =UTF8&th=1 Is there anything specific that I need set for my monitor? I believe it is running 720p 1280x768 @ 60hz.

[Retro-Nerd]

Native resolution of your monitor is 1920x1080p, which is video_mode=8.

You can't use the vrr_min and max settings from my Mister.in since the LG OLED is a 120 Hz display. You have to use your VRR range settings according to your monitor manual (which should be 50-75 from what i've read).

[davidelkinsrice]

I took your MiSTer.ini and changed things to 50-75 with video_mode=8. My monitor is in Game Mode, HDMI 1080p 60Hz model ASUS VP228. With Adaptive-Sync/FreeSync On. HDMI is the input. I verified that I'm using a HDMI cable - (not a vga cable with an adapter). I took your MiSTer.ini and took out the stuff at the bottom for specific things did a sort|uniq to it and did the same thing with my MiSTer.ini file and compared values and they were all pretty much the same. I put [MiSTer] at the top. I updated it to not use the alternate menu (it was messing up my menu system). I tried it with the rom and it didn't work. Are there any settings on the actual hardware that could affect this from not working? I put changed osd_rotate=1 because my monitor is vertical. I noticed that your file has vga_mode=rgb (my didn't have it at all - seems strange to me).

Online I saw this: Check the Connection: Ensure that the monitor is connected to the graphics card using a DisplayPort cable, as FreeSync typically requires a DisplayPort connection to function properly. I don't think my mister has a display port connection. Are you using a DisplayPort? Do I maybe need to run a script to make this all work? I have run the fast_polling_usb script and that seems to help with speed. Thank you for any help.

[Retro-Nerd]

You may have a point here. I'm sure that some monitors have to use a display port connection for Freesync use, some others may only work via HDMI connection. My PC monitor (a 27 Inch 2021 Dell e.g. can use both connection types with freesync).

But for my test i've used my LG OLED C9 TV via HDMI. Display Port is (for now) a pure monitor thing i guess. Your monitor supports display port. Get a HDMI to display port adapter (or a complete cable with HDMI on one end and display port on the other) and try it. Worth a shot.