D
Homebrew RetroArch Switch
- Thread starter ShadowOne333
- Start date
- Views 1,710,983
- Replies 9,099
- Likes 95
getting something to run on a computer is kind of a different animal that getting something to run on a closed source commercial gaming device. the progress made in just the last 3 months is incredible. i imagine the kinks will be worked out in less than a year.
Well right now, all of these builds are using libretro to build one core at a time, so it's kinda better than when the whole thing gets dumped and we have to choose cores. But the psx emu is running slow, the Gameboy emu is running slow. The save files are wonky. The retroarch front end doesn't save the settings you set.
Again I'm not trying be ungrateful, I am happy to have full speed SNES in these early days. I know without this libretro switch support it would take each individual emu author to port his emu to the switch, and that would take more time, but I think it would be better.
Nobody of sane mind can say that setting up and using Retroarch is a simple process.
Again I'm not trying be ungrateful, I am happy to have full speed SNES in these early days. I know without this libretro switch support it would take each individual emu author to port his emu to the switch, and that would take more time, but I think it would be better.
Nobody of sane mind can say that setting up and using Retroarch is a simple process.
Last edited by Tomobobo,
D
Deleted User
Guest
most crashes are from the bad libtransistor build , its an early project we have to wait right now endrift is inside irc to talk about mgba lets see what will happen
Honestly I'd be inclined to agree. I have used retroarch on a bunch of devices (mostly Android devices) and it never worked as it should, in most cases being almost unusable. For example, on a older tablet I had the interface text was too tiny and I could barely read, settings wouldn't save, each core used the same controller overlay instead of using the right layout for each console, the physical back button wouldn't work (and retroarch was the only app where it didn't work), due to it not working I also couldn't disable the controller overlay otherwise I wouldn't be able to exit from the game and so on... It was just a big mess... To make everything worse, you also need to enter on 200 different menus (and some of those menus are confusing, for example, a menu called "settings" and other called "configurations" that does different things), change a lot of settings to get things working right. And those settings usually don't save, for whatever reason, so you need to set it up again and again...
Having separate emulators for each system is, IMO, the best approach. Back in 2011 I used to own a chinese gaming device that was capable of running linux and had a launcher called "gmenu2x", it was just amazing. You could just open the ROM file straight from the launcher, and it would run it on the emulator and you could start playing it right away, 0 configuration needed, everything just worked. I think something like that would be ideal for the switch.
Yeah I mean, I know this probably isn't the place to be talking shit about Retroarch but it's how I feel. The UI is terrible and hasn't been updated in ever. Everything is so complex for no reason. I know it's a super emulator or whatever but somebody who isn't just sitting there slamming down C++ code all day should take a look at the thing and make it at least halfway end-user friendly.
If you think loading and playing roms is a complex task you need to go back to kindergarten
I know, I'm a little bit bothersome with my wish for an Dreamcast Emulatior. But I'm extremly undecided if I should buy a Switch, or not. The point is, that there is not even one game I realy want to play and games I want a little bit to play (Xenoblade 2, Yooka Laylee) need a higher Firmware than 3.0.0.
So I want to ask, what you think which probability there is, that RetroArch includes Dreamcast for Switch, or a standalone emulator for it will come? I dont need other Emulators like NES / SNES and so on, because I have this on my Wii U.
So I want to ask, what you think which probability there is, that RetroArch includes Dreamcast for Switch, or a standalone emulator for it will come? I dont need other Emulators like NES / SNES and so on, because I have this on my Wii U.
You have to take into consideration that PC emulators are a dedicated thing.
They work for PC and that's as far as they go, they are worked upon this very same thing and polished for it.
RetroArch on the other hand, focuses on portability and accessibility.
It tries to bring all of those things that were previously only neglected for PC users, and tries to bring it to the most devices possible.
Obviously something is gonna get borked in one of them, you can't expect all of them to work 100%.
Every device has it's own libraries, audio drivers, graphics drives, CPU, GPU, etc. There are way too many variables to count.
RetroArch Switch is on it's bare infancy, baby steps right now, give it time to be polished and it will outshine even the Shield TV I bet.
The issue with retroarch is not what we see here (here we see mostly framework issues) but I tend to agree that it is a sense of ownership that is missing.
You dont have the one guy anymore, that owns the hardware platform, thinks about its implications, and engineers a concept around it. Its always the other way around. The demands of the port set the rules, many people work on it, no one has a unified vision that isnt the same interface from seven years ago, decisions are made from a different perspective. Thats why for me it almost felt typical for the resume (press a three times) screen interaction to be botched, the screenshot button not working, no one having decided to pull any menu item up front that would be Switch centric - its the same port as always, no attention to detail, no optimization (for much of that its too early)...
I've had a comparative experience with "single system" emulators (some of them commercial) on an Android TV Box, and to me the best ones blew retroarch out of the water - every time. The example I usually like to use is PPSSPP, where retroarchs policy for a long time was to recommend a worse core, just because it confirmed to their interface restrictions.
Another issue comes up, when you really find problems with some of the cores. Ownership gets diluted up to the libreto cores tree, where for some of the more obscure cores (scummvm) you find, that no one is willing to do anything anymore, not even visit the bugtracker - even if just doing something small - like redoing(/allowing to change) joystick sensitivity - would make the thing playable on a 100 additional systems with minimal effort - while currently it may not be.
At the same time exactly the claim that "we do everything" makes it daunting for new projects to get a foot on the ground. Also - and I dont want to downplay this, some of their emulation cores really are the best that are out there.
Another issue are milestones and targets. If 80% of Retroarch development is targeting Shield TV performance levels, and the market simply isnt delivering anything along those lines, as far as "supported devices" that people actually buy, you sometimes just make the problem worse, by optimizing your cores away from what people out there would actually need (pcsx rearmed became actually less performant on the Android Box (overpowered for its purposes) I used in the aformentioned comparison).
Thats why I strongly hoped for the Switch to be a potential change in direction, as it represents a very popular plattform, that opens previously unseen usage scenarios - and could have very well been a potential focal point to rethink some of their previous approaches...
But then you look at the "choosing not the best library" issues, and at HBL crashing 50% of the time - and it just doesnt seem to be a change in direction.
Hopefully once a CFW gets released the experience will get better, so will stability of the RA cores, obviously, and eventually (there hasnt even been a first official release), but its still the same old retroarch.
And looking at those points, yes most of them have an organisational reason behind them, that might just as well point at the project having become too big, to be "nimble" in any sort or form.
There are other approaches to this as well (glancing at cpasjustes work), and to be honest - I like them better. But then again, even though one dedicated guy overlooking all builds for a system is great, its not scalable "for all systems" and all emulators.
And looking at todays audience... They are more happy - when they can put up another youtube video with one more supported system - than to actually talk details about the emulation performance, or a fitting ui design.
You dont have the one guy anymore, that owns the hardware platform, thinks about its implications, and engineers a concept around it. Its always the other way around. The demands of the port set the rules, many people work on it, no one has a unified vision that isnt the same interface from seven years ago, decisions are made from a different perspective. Thats why for me it almost felt typical for the resume (press a three times) screen interaction to be botched, the screenshot button not working, no one having decided to pull any menu item up front that would be Switch centric - its the same port as always, no attention to detail, no optimization (for much of that its too early)...
I've had a comparative experience with "single system" emulators (some of them commercial) on an Android TV Box, and to me the best ones blew retroarch out of the water - every time. The example I usually like to use is PPSSPP, where retroarchs policy for a long time was to recommend a worse core, just because it confirmed to their interface restrictions.
Another issue comes up, when you really find problems with some of the cores. Ownership gets diluted up to the libreto cores tree, where for some of the more obscure cores (scummvm) you find, that no one is willing to do anything anymore, not even visit the bugtracker - even if just doing something small - like redoing(/allowing to change) joystick sensitivity - would make the thing playable on a 100 additional systems with minimal effort - while currently it may not be.
At the same time exactly the claim that "we do everything" makes it daunting for new projects to get a foot on the ground. Also - and I dont want to downplay this, some of their emulation cores really are the best that are out there.
Another issue are milestones and targets. If 80% of Retroarch development is targeting Shield TV performance levels, and the market simply isnt delivering anything along those lines, as far as "supported devices" that people actually buy, you sometimes just make the problem worse, by optimizing your cores away from what people out there would actually need (pcsx rearmed became actually less performant on the Android Box (overpowered for its purposes) I used in the aformentioned comparison).
Thats why I strongly hoped for the Switch to be a potential change in direction, as it represents a very popular plattform, that opens previously unseen usage scenarios - and could have very well been a potential focal point to rethink some of their previous approaches...
But then you look at the "choosing not the best library" issues, and at HBL crashing 50% of the time - and it just doesnt seem to be a change in direction.
Hopefully once a CFW gets released the experience will get better, so will stability of the RA cores, obviously, and eventually (there hasnt even been a first official release), but its still the same old retroarch.
And looking at those points, yes most of them have an organisational reason behind them, that might just as well point at the project having become too big, to be "nimble" in any sort or form.
There are other approaches to this as well (glancing at cpasjustes work), and to be honest - I like them better. But then again, even though one dedicated guy overlooking all builds for a system is great, its not scalable "for all systems" and all emulators.
And looking at todays audience... They are more happy - when they can put up another youtube video with one more supported system - than to actually talk details about the emulation performance, or a fitting ui design.
Last edited by notimp,
Yeah I saw one of the official libretro devs give a similar response to Retroarch criticism on reddit. It's probably why the emulator is still in the state it is in today.
It's not that it doesn't work, because sometimes it does. It's about whether it's worth the trouble to fiddle with a seemingly endless batch of settings using a text only barely functional menu full of submenus and names of cores that you have to be knowledgeable of to run any given game. On windows, standalone emulators are up and running seconds after install. On android, a standalone NES emulator boots up, choose a rom from a familiarly styled interface, and you're playing NES. On both of those platforms, setting up Retroarch can take forever, you must manually download and insert cores, reset controller schemes which are forced on you, set pathways, and god knows what else I've never gotten through to get anything working I've always given up and gone with something much more useable.
You can call me stupid and not capable of running emulators but it's not that I can't, just that I wish for something better. I know it's what we have now and I am thankful, just not hopeful.
I understand this emulator is trying to do many tasks on many platforms, but I feel like that is it's major weakness. For years I have watched this emulator grow, and think it could use a little bit of user friendliness tossed in. It seems the developers are against such a notion, and it is reflected in this software.
Here you go, fixed and tested. Succesfully back to HBL after quit.
For me the latest snes release does allow a return to HBL, but attempting to go past the first screen's list of roms causes a lock up. Exiting and returning to HBL and launching again gives an error of No Directory Found.
I dunno. I have a list of roms, so that all of them aren't displayed on one "screen" and you would have to scroll down to get to more roms. Upon reaching the bottom of the list, the cursor dips into a line that has nothing on it. And the controls seem locked. You can hit X or Y or something and it will bring up a search option and maybe with enough button mashing you can get back to the part of the list you can see, but I can't load roms that aren't on this screen, and ran into shut down errors eventually every time.
Oh, I've had that issue before as well. Not tested the latest build, but trying to load a game, RA crashing - reloading RA, and it then not being able to see content on the sdcard until you reboot the device and start over - happened to me before as well.
Great!!! load everything right, save correct and you can go to hbl and return without problems, now you just need to be able to modify the size of the screen and it will be perfect !!!
