Dolphin, the GameCube and Wii emulator - Forums

Quote:Here's a suggestion: just give it up for now. I've been trying to get this game work for ages and nothing helps. You can get rid of the shadow glitch by setting EFP copies to RAM but that also tanks the FPS which thanks to the "new and awesome, super accurate" Dolphin makes the sound stutter.

Trust me, the improvements to sound quality since the New-AX-HLE-AX merger are massive. Prior, you may as well have just turned the sound off and replaced it with your ipod but now it actually sounds like it's supposed to. Stutter is mostly a hardware issue.

(12-23-2013, 12:10 AM)Zaxx Wrote: [ -> ]Dolphin's a bad software for this game, in fact Dolphin's a bad emulator for most Wii games: if you don't have a super overclocked CPU you'll never get good performance because Dolphin sucks at using more than two cores and it fails at fully utitlizing even that two. I'm sorry but if you have a quad core CPU running at 3.5 ghz and someone says you have to overclock more you know you've stepped into badly programmed territory.

By your logic emulators are inherntly bad software.

The reality is that the differences in architecture are huge between a powerPC based Gamecube/Wii and an X86-based PC. The instruction sets have to be processed in a roundabout way and yet you still have to get timings right. What dolphin is doing is basically akin to putting The Undertaker into a UFC bout with GSP. It's inherently a flawed concept and yet it would be amazing if and when it works out.

As for "sucking at using two cores" that is because more cores would make the particular operations MORE difficult not LESS.

(12-23-2013, 10:59 AM)Zaxx Wrote: [ -> ] What you refer to as the strength of the emulator is a weakness in my book: accuracy. Accuracy hurts the performance so focusing on that right now is stupid.

So sound quality, artifacting, crashes, glitches, etc are not part of the performance?

Quote:As far as the required hardware for accurate emulation is not available to consumers

Funny, I bought it a few weeks ago, and haven't even pushed it to its maximum potential.

Quote:Just look at my example: I want to play this game on my PC for 3 years now and I always seem to be just under the hardware requirements with whatever PC I have (and I won't overclock because I don't do that shit to my machine) so for me accuracy doesn't really help. I'm sure I'll buy a Wii dirt cheap before I could actually run this game on PC.

Wait, you don't even have a Wii?

(12-23-2013, 02:05 PM)Zaxx Wrote: [ -> ]I just think of it as a bit of a loss when most people have no hardware for Dolphin and its requirements just keep going higher and higher.

Dolphin is not for a "lot of people". It's for

A) People who are just interested in the concept of emulation itself
B) People who appreciate Gamecube and Wii games enough that they want to see them played with the benefits of emulation IE higher res like me and we will get that hardware for it.

Quote:Just look at my example: I want to play this game on my PC for 3 years now and I always seem to be just under the hardware requirements with whatever PC I have

You know, this game has an actual PC port. You can get it for like 5 bucks off Steam right now. Just sayin'...

Quote: (and I won't overclock because I don't do that shit to my machine)

Huh? First of all overclocking is not "shit" and second, many people who overclock treat their systems with plenty of care - the way you say it makes it sound like abuse and that's far from the truth. Anyways, if you insist on never overclocking, just wait for intel to come out with Haswell-E. We already know how powerful haswell is in dolphin processing, and those things should have more cache, higher stock speeds, etc.

Quote:Saying they are "playable" is a lie, it should be something about how accurate Dolphin is. Anyway isn't actually achieving full speed should be the first criteria in reaching an accurate emulation?

It's nice to get improvements in speed, but "using hacks to get around hardware limitations" has nothing to do with accurate emulation. Using hardware to get around hardware limitations, and software to get around software limitations is good design. Further, the devs DO still make efforts to optimize framerate performance where possible - it might only be an extra 2-3% here or there but it does happen.

(12-23-2013, 09:35 PM)Zaxx Wrote: [ -> ]The first one is from the middle of 2012, that's ONE AND A HALF YEARS. D I know what "not a priority" means and it certainly isn't "we absolutely don't care about that". The issue is there for sure since 3.0 and it's possible it was always a problem. And that really reflects to my main point: what is the point of putting exteme amount of work into something when not even the controls are fully emulated. I understand it has to be a tough job to do that since the Wii has a unique controller but the PC has a mouse and a keyboard and shaking works almost perfectly so with some work it's possible. I'm not saying to do it right now, I just say that development should switch a bit of its focus to get some of the more practical stuff working, then you can make your awesome accurate emulator.

Emulating a wiimote with a Keyboard/Mouse is not "more practical stuff" - not when anyone can get a regular wiimote working with the emulator effortlessly. But if it's really important to you, I implore you to learn how to code and implement this - that's what open source is all about.

Quote: Trust me, the improvements to sound quality since the New-AX-HLE-AX merger are massive. Prior, you may as well have just turned the sound off and replaced it with your ipod but now it actually sounds like it's supposed to. Stutter is mostly a hardware issue.

Please don't start the "you have a shitty computer" argument, I'm not in the mood for 5 year olds. Anyway I was not really accurate: it's not stuttering, it's that sound halts if you're not running at full speed.

Quote:By your logic emulators are inherntly bad software.

The reality is that the differences in architecture are huge between a powerPC based Gamecube/Wii and an X86-based PC. The instruction sets have to be processed in a roundabout way and yet you still have to get timings right. What dolphin is doing is basically akin to putting The Undertaker into a UFC bout with GSP. It's inherently a flawed concept and yet it would be amazing if and when it works out.

As for "sucking at using two cores" that is because more cores would make the particular operations MORE difficult not LESS.

Okay.

Quote: So sound quality, artifacting, crashes, glitches, etc are not part of the performance?

PCSX2 has no problems with those factors and it runs PS2 games at full speed. Wii games are of course a bit more advanced than that so it's okay if Dolphin needs more horsepower, the point of my argument was that if there is no artifacting, crashing etc. then that's accurate enough to start focusing on optimizing.

Quote:Funny, I bought it a few weeks ago, and haven't even pushed it to its maximum potential.

Most people don't buy new rigs every few months.

Quote:Wait, you don't even have a Wii?

Nope, I don't have a Wii because I'm not interested in the Wii game library and I don't want to buy a console because of one game. I do own the Wii version of Forgotten Sands though so I'm not fully "illegal".

Quote:Dolphin is not for a "lot of people". It's for

A) People who are just interested in the concept of emulation itself
B) People who appreciate Gamecube and Wii games enough that they want to see them played with the benefits of emulation IE higher res like me and we will get that hardware for it.

C) People who are not really interested in emulation and who don't appreciate the Wii but they'd like to play that one game that interests them.

Quote:You know, this game has an actual PC port. You can get it for like 5 bucks off Steam right now. Just sayin'...

The game has no PC port, the Wii version is 100% different than the other versions with its own levels, gameplay mechanics and story.

Quote:Huh? First of all overclocking is not "shit" and second, many people who overclock treat their systems with plenty of care - the way you say it makes it sound like abuse and that's far from the truth. Anyways, if you insist on never overclocking, just wait for intel to come out with Haswell-E. We already know how powerful haswell is in dolphin processing, and those things should have more cache, higher stock speeds, etc.

Overclocking can hurt the lifetime of your hardware. End of argument.

Quote:Emulating a wiimote with a Keyboard/Mouse is not "more practical stuff" - not when anyone can get a regular wiimote working with the emulator effortlessly. But if it's really important to you, I implore you to learn how to code and implement this - that's what open source is all about.

Emulating a wiimote is practical because it lets you play a Wii game if you don't have a wiimote. Viewing practicality from a developer standpoint is very flawed logic.

Anyway: I've figured it out. The feature works but it's not explained anywhere how: if you hook swing emulation up with the mouse it will work. I've spent an hour on it, but I figured out a control scheme that let's me play the game without much of a hassle: I use my 360 pad for everything and do the swings with the mouse. At first it seemed like a horrible idea but got used to it in 30 minutes so it's time for some PoP.

Zaxx Wrote:Please don't start the "you have a shitty computer" argument, I'm not in the mood for 5 year olds.

The new-ax-hle merges eliminated the need for systems to overclock in order to run LLE audio at fullspeed in many games. It's not about whether or not a computer was "shitty". Games like Skies of Arcadia, Xenoblade Chronicles, and Tales of Symphonia needed LLE audio for proper sound emulation, and even a stock i5-3570K needed some amount of OC (or Turbo Boost) to get them running at fullspeed constantly. Now HLE audio allows high-end and mid-range systems to run them at enjoyable speeds. In addition, it fixed a host of audio issues, included a ton of crashes that were related to inaccurate DSP emulation. If you feel that Dolphin should have the old AX ucode HLE implementation, I'd advise reading this article here from the developer that worked on the new-ax-hle branches: http://blog.delroth.net/2013/07/why-dolphin-is-getting-rid-of-asynchronous-audio-processing/

Zaxx Wrote:Anyway I was not really accurate: it's not stuttering, it's that sound halts if you're not running at full speed.

File an issue report on Google Code. Outside of several well-known instances (Super Mario Galaxy 1 & 2, MK: Double Dash iirc) games should not drop, stop, or halt audio (even when not running fullspeed). If this does happen, it's a serious defect and deserves developer attention.

Zaxx Wrote:Nope, I don't have a Wii because I'm not interested in the Wii game library and I don't want to buy a console because of one game. I do own the Wii version of Forgotten Sands though so I'm not fully "illegal".

As long as you dump your own disc, there's nothing illegal about making personal copies in most countries, and in fact such a practice is often protected under law.

Zaxx Wrote:Overclocking can hurt the lifetime of your hardware. End of argument.

Only if you consistently have terribly high temperatures and give your hardware too much voltage. Proper overclocking (which isn't that hard to do these days, just get a cooler, watch your temps and volts) won't noticeably reduce the lifespan of your hardware. That is to say, by the time your hardware begins to fail, you really ought to be in the market for new hardware. I'm not talking 5 or 6 years; modern hardware will easily last 10+ years even when OCed. Too much heat and too much voltage are the main factors that contribute to early hardware failure; avoiding both will give you a stable system that will last about as long as you need it to, before it needs to be replaced.

Zaxx Wrote:Viewing practicality from a developer standpoint is very flawed logic.

That's not to say that ordinary users have better ideas of what is practical and what isn't. I've seen some pretty "out there" suggestions for Dolphin (like compiling specific versions for every known game). It's also not like the developers don't use and play Dolphin themselves for their own enjoyment. Again, the new-ax-hle branch came about because delroth was sick of dealing with crappy audio in some of his favorite games. It was practical for him as both a developer and a gamer, and in turn was quite practical for a number of other users. Simply because someone else's idea of what's practical doesn't match yours, you can't say that their idea is based on flawed logic. It's just having a difference in opinion, only the devs are in a position to make changes to Dolphin more easily than others.

Quote: As far as the required hardware for accurate emulation is not available to consumers

Gotta defend him here. To get the most accurate emulation you'd have to use an Interpreter and a Software Renderer, which simply won't run nicely on any hardware.

But since there's stuff like JIT and DirectX (and OpenGL) we don't have to use those. Anyways, I'm also rocking the i5 4670k @4.5Ghz and it runs most games at fullspeed, even with LLE (Mario Galaxy etc. runs at like crazy 100FPS).

You know, if you pursue more and more accurate emulation far enough, you can bring just about any machine to its knees. That Dolphin's most accurate settings do this (quite easily) doesn't mean much. I still can't run TWINE, Mario Tennis, or any of the Star Wars games fullspeed with the z64 LLE plugins in Mupen64Plus. I haven't OCed past 3.8GHz, but that a Sandy Bridge at that level can't run it is saying quite a bit. I've heard someone working on a PSX emulator basically brought their high-end system down when they developed an accurate implementation of a certain feature (I heard this second-hand on the forums here, so I don't know the exact source). Go accurate enough, and I'm certain you could make an NES emulator that tramples over anything you can buy and OC.

Quote: It's also not like the developers don't use and play Dolphin themselves for their own enjoyment. Again, the new-ax-hle branch came about because delroth was sick of dealing with crappy audio in some of his favorite games. It was practical for him as both a developer and a gamer, and in turn was quite practical for a number of other users.

Hold the horses there, mate, let we get this straight: so a previous solution that worked fine for many games (I know for a fact that the old crappy sound emulation provided continous sound for PoP without desync issues: the music was continous and sound effects were not in desync) was scrapped becuse his favourite games were not running? Oh gosh, someone plase show this game to delroth, maybe he'll like it so it will run better in the next version! Ridiculous.

What is practical? Mainly something that benefits all users, like fixing the key bindings with AND or OR arguments not saving. Objectivity, man, that's all where doing good things start.

Quote: I've seen some pretty "out there" suggestions for Dolphin (like compiling specific versions for every known game).

Absolutely not a stupid idea if you look at the benefits. Every known game in existence is going very far but you must know about some games that are notorious on Dolphin. On how this could be done: listen to community feedback and list the recommended revisions for certain games on their wiki pages. For example Forgotten Sands is running quite well with the revision I wrote about in one of my previous posts but it's slow under every other version and to make make matters worse on the most recent revisions the graphics wouldn't even render so the game is unusable. You should really spare your users from the frustration of having to try out a million revisions to find the right one because realistically speaking most of the unstable revisions are garbage because of the lack of testing and the stable releases are always a bit slower than they should be in my case. The thing that basically got the game working was that I've read a forum post that said "try this revision, it's running great for me on this". You know, some of the "out there" suggestions can give you ideas that are actually good, in some cases they are just badly articulated but that doesn't stop you from getting meaningful stuff out of them.

And yes, focusing on a specific game for a revision is not a crime: do a vote where you list every Wii game and prioritize those that most people want to get running. It would definitely get some games running properly quicker than waiting for the miraculous moment when Dolphin will be perfect and there will be 8 ghz CPUs in your microwave on your flying car. It's a working alternative, most businesses use this to make more money like the store on GOG: they have a wishlist where people can vote on games they want to buy and the GOG team prioritizes accordingly and they make more money in the process instead of just relying on hit or miss.

Zaxx Wrote:Hold the horses there, mate, let we get this straight: so a previous solution that worked fine for many games (I know for a fact that the old crappy sound emulation provided continous sound for PoP without desync issues: the music was continous and sound effects were not in desync) was scrapped becuse his favourite games were not running? Oh gosh, someone plase show this game to delroth, maybe he'll like it so it will run better in the next version! Ridiculous.

Did you even read the article I linked to? It explains the entire reasoning behind the new audio. The old "solution" had more issues than the new one. The new solution works well for a far greater range on games. Again, if you have an issue with the new audio, file an issue report. The devs aren't likely to know anything is broken unless you let them know.

Zaxx Wrote:What is practical? Mainly something that benefits all users, like fixing the key bindings with AND or OR arguments not saving. Objectivity, man, that's all where doing good things start.

I don't know what you're talking about. AND and OR, separately or in any combination, save just fine for me. It's always possible to manually edit the controller profile .ini files to force Dolphin to use your preferred configuration. If the GUI is not saving the arguments correctly, or if the .ini files are being ignored, file an issue. If an issue already exists, you can always add your own feedback about the issue when appropriate. Even just commenting on an old issue lets the devs know that some people are still interested in having an issue fixed.

Zaxx Wrote:Absolutely not a stupid idea if you look at the benefits. Every known game in existence is going very far but you must know about some games that are notorious on Dolphin. On how this could be done: listen to community feedback and list the recommended revisions for certain games on their wiki pages. For example Forgotten Sands is running quite well with the revision I wrote about in one of my previous posts but it's slow under every other version and to make make matters worse on the most recent revisions the graphics wouldn't even render so the game is unusable. You should really spare your users from the frustration of having to try out a million revisions to find the right one because realistically speaking most of the unstable revisions are garbage because of the lack of testing and the stable releases are always a bit slower than they should be in my case. The thing that basically got the game working was that I've read a forum post that said "try this revision, it's running great for me on this". You know, some of the "out there" suggestions can give you ideas that are actually good, in some cases they are just badly articulated but that doesn't stop you from getting meaningful stuff out of them.

And yes, focusing on a specific game for a revision is not a crime: do a vote where you list every Wii game and prioritize those that most people want to get running. It would definitely get some games running properly quicker than waiting for the miraculous moment when Dolphin will be perfect and there will be 8 ghz CPUs in your microwave on your flying car. It's a working alternative, most businesses use this to make more money like the store on GOG: they have a wishlist where people can vote on games they want to buy and the GOG team prioritizes accordingly and they make more money in the process instead of just relying on hit or miss.

I don't think you really understood what I was saying. The idea wasn't to have a master list of which revision runs which game the best, nor was it to have developers focus on a certain game for a certain revision (the latter actually happens a lot already). The idea was something to the effect of making a new branch in Dolphin for each and every game. The code in each branch would focus on running that game "perfectly". Basically, this user wanted Dolphin to start a mini-project for every known game, thus creating specific versions of Dolphin, each with its own changes made just for each game.

This is an absolutely terrible idea. For one thing, the developers, even collectively, don't have access to every known game. It isn't enough to pass on a revision to users that do have the game when said users don't have the knowledge to properly test if emulation is indeed running as it should under the hood (e.g. if it crashes for some reason, most users have no idea what line of offending code would be responsible). Hence, it's not even realistic to make, let alone debug specific revisions of Dolphin that are tailor made every game. Secondly, code maintenance would be a nightmare. Updating each and every branch is a very time-consuming task. Additionally, making global changes to any of them requires a messy amount of git merges from the master branch, which would mean potentially having to deal with a lot merge conflicts and possibly bringing in code that breaks some of the hundreds of branches but fixes others (which again, makes testing impractical). Not to mention the fact that some branches may get way ahead of others while some continue to use outdated code, that is, even if someone bothers to update each branch properly. Imagine trying to create a new feature in Dolphin that will speed up most games, but to do that you have to create 3 separate implementations to conform with different code bases written at different times, then merge your changes into a couple of thousand branches, then rewrite your code 3 times again and perform all of those merges again when you discover a bug (even a slight one). Few people are willing to go through all of that effort on their free time. There are reasons why other emulators don't make branches like this, namely because most of the devs would prefer to hold onto their sanity.

I understood why it is an unrealistic idea, I merely stated that there is something rational and reasonable to take away from it. If there are revisions focusing on a certain title, make a list of them because all the stuff I've seen next to revisions as descriptions were technical stuff someone who's not qualified won't understand.

As for the article, believe it or not I've read it and yes, I understand why is it better. BUT: understanding why is it better does not help when it doesn't work for you because the game cannot maintain full speed. So yes, from a technical standpoint it's better, from a practical standpoint it's only better if you have the hardware for it but if you don't and the other solution was better for you, you can wipe your a** with all the dev reasoning.

Why do I have to write down the same things all over again? There is good reasoning behind the things you state but in order to get the full picture sometimes you got to get away from the technical standpoint.

Zaxx Wrote:I understood why it is an unrealistic idea, I merely stated that there is something rational and reasonable to take away from it. If there are revisions focusing on a certain title, make a list of them because all the stuff I've seen next to revisions as descriptions were technical stuff someone who's not qualified won't understand.

Anyone's free to make such a list. It's not the responsibility of the devs (or anyone else) unless they choose to undertake that task. The closest things we have are this thread and the hundreds of Testing Entries on the Dolphin Wiki. These things require user participation (and a lot of their time as well, to test which revision worked best for them on a certain game or games), which is why it's not easy to just compile such a list. Most of the revisions that focused on improving certain games carry commit messages (on the git log) like "Implemented floating points, Game XYZ now boots", like 3.0-751, and sometimes the devs do go out of their way to hunt for game-specific bugs.

Zaxx Wrote:As for the article, believe it or not I've read it and yes, I understand why is it better. BUT: understanding why is it better does not help when it doesn't work for you because the game cannot maintain full speed. So yes, from a technical standpoint it's better, from a practical standpoint it's only better if you have the hardware for it but if you don't and the other solution was better for you, you can wipe your a** with all the dev reasoning.

Again, no one's forcing people to use the latest revisions, or revisions after the new-ax-hle merges, or even to upgrade their hardware. If they want to have the old audio, they can use old revisions; the only stipulation is that they can't ask for support on any issues they encounter. That's the compromise.

Zaxx Wrote:Why do I have to write down the same things all over again? There is good reasoning behind the things you state but in order to get the full picture sometimes you got to get away from the technical standpoint.

Like I said, it's not like the devs are totally focused on the technical aspects and only the technical aspects of Dolphin. Some of them don't just work on Dolphin because they enjoy programming or like the challenge of making an emulator; they honestly want to play GC and Wii games on their PCs (and soon Android devices). It's simply not the case that they can't see things from an ordinary user's perspective. Many decisions are made because the devs feel they are good from a user's standpoint, e.g. the inclusion of all the necessary .dlls in the latest development revisions. It makes it easy to click and run without having to install other dependencies on Windows, which was a headache for many users (see the SDL.dll, VCOMP100.dll, XINPUT1_3.dll, and 0xc000007b errors on the FAQ). They also frequently ask for input and feedback in the Development Forum on major changes (the new-ax-hle branch was one example, as was the Global User Directory addition for Windows).

(12-25-2013, 03:04 AM)Shonumi Wrote: [ -> ]

' Wrote:I understood why it is an unrealistic idea, I merely stated that there is something rational and reasonable to take away from it. If there are revisions focusing on a certain title, make a list of them because all the stuff I've seen next to revisions as descriptions were technical stuff someone who's not qualified won't understand.

Anyone's free to make such a list. It's not the responsibility of the devs (or anyone else) unless they choose to undertake that task. The closest things we have are this thread and the hundreds of Testing Entries on the Dolphin Wiki. These things require user participation (and a lot of their time as well, to test which revision worked best for them on a certain game or games), which is why it's not easy to just compile such a list. Most of the revisions that focused on improving certain games carry commit messages (on the git log) like "Implemented floating points, Game XYZ now boots", like 3.0-751, and sometimes the devs do go out of their way to hunt for game-specific bugs.

' Wrote:As for the article, believe it or not I've read it and yes, I understand why is it better. BUT: understanding why is it better does not help when it doesn't work for you because the game cannot maintain full speed. So yes, from a technical standpoint it's better, from a practical standpoint it's only better if you have the hardware for it but if you don't and the other solution was better for you, you can wipe your a** with all the dev reasoning.

Again, no one's forcing people to use the latest revisions, or revisions after the new-ax-hle merges, or even to upgrade their hardware. If they want to have the old audio, they can use old revisions; the only stipulation is that they can't ask for support on any issues they encounter. That's the compromise.

' Wrote:Why do I have to write down the same things all over again? There is good reasoning behind the things you state but in order to get the full picture sometimes you got to get away from the technical standpoint.

Like I said, it's not like the devs are totally focused on the technical aspects and only the technical aspects of Dolphin. Some of them don't just work on Dolphin because they enjoy programming or like the challenge of making an emulator; they honestly want to play GC and Wii games on their PCs (and soon Android devices). It's simply not the case that they can't see things from an ordinary user's perspective. Many decisions are made because the devs feel they are good from a user's standpoint, e.g. the inclusion of all the necessary .dlls in the latest development revisions. It makes it easy to click and run without having to install other dependencies on Windows, which was a headache for many users (see the SDL.dll, VCOMP100.dll, XINPUT1_3.dll, and 0xc000007b errors on the FAQ). They also frequently ask for input and feedback in the Development Forum on major changes (the new-ax-hle branch was one example, as was the Global User Directory addition for Windows).

Look, I won't write everything down I've already said because my opinion won't change. "Everyone is free to do this" is not a solution because no one is doing it, simple as that. Writing a list based on feedback needs no testing, you don't have to overcomplicate everything: you just write down a revision's name others said was working fine. It might not work for the user at the end, but at least it's a starting point. As far is it's not working releasing 3 revisions a day is pointless whil expecting that the user will try out a hundred of them: that is stupid.