-
Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.
Cuphead [Switch] - Docked/Handheld & Co-op Frame Rate Analysis
- Thread starter Equanimity
- Start date
OP
OP
It comes out tomorrow. I don't know when the embargo ends, CN might have broken it.
Embargo has been very loose since March allowing you to provide basically anything apart from a review. That said, this channel has a long history of breaking embargo.
Niceeeeee. Easy purchase.
A broken embargo for them is one more pre-order from me haha
He's just showing footage from the first world, which sites and YouTubers have been showing for weeks.
looks great, pretty much perfect 60. will get it as soon as i'm done with yoshi.
and native res 720p/1080p.
Isn't that on purpose because of the parries? Just asking, i'm curious.
There's a few small, brief dips into the 50s actually, but I doubt many will notice when playing.
OP
OP
Yes
But this one was just a jump
I honestly wouldn't put much stock in this, the frame rate tool requires some optimisation for accurate results and I'm not convinced Contra Network has done that. The DOOM test for instance basically reports a locked 30 when that's very unlikely the case.
Curious about load times. Remember playing through the game on first edition Xbox and it had some pretty long load times.
Giant bomb had a quick look up a few weeks ago and the load times seemed pretty comparable to the Xbox version
OP
OP
They use TrDrop (Tear Drop) to measure frame performance, but the tool requires a ton of optimization to get accurate results. The tool thinks if you stand still that the software is running at 0 fps.
Yeah, he's notorious for breaking embargoes and spamming websites to get his videos out everywhere. He's even done so on this very site without actually getting into discussions, which makes him annoying to see
OP
OP
Oh, thank you for sharing this info. I'll keep it in mind for future posts.
When I wanna know about this stuff, I will always go to Digital Foundry regardless of when the video is posted. I know they're legit and you can take their word to the bank.
Just to be clear, trdrop can be reliable. We currently use a modified version of it. However, you need to know what you're doing in order to take advantage of it, and make sure to work within its limitations to get accurate results.
Oh, I've heard about this. I might have to mess around with it to see how it compares.
Right. It can be reliable once you commit the time needed to tinker and optimize the codex. It just requires a lot of testing and know-how. In the right hands it can be a valuable tool.
At least with you guys, the software will work more optimally since you all use uncompressed files for analysis, which is not something that everyone does, unfortunately.
Agreed.
Ah, yes. That does help.
What worries me about this tool is that some people seem to use it incorrectly which creates invalid data. That means there is now variability between different people/channels and it's especially bad when the data is presented without commentary. People see it and can draw the wrong conclusion.
This is true. Though it should be noted that, if the public had access to DF's tools, it would be possible to misuse them as well without the expertise to properly prime, analyze and potentially correct the data. No frame analysis tool is automated flawlessly since these types of algorithms are prone to error for a variety of reasons (especially when new games present frames in unanticipated ways, making it difficult to distinguish between duplicate and unique frames), and identifying/correcting such issues requires the oversight of someone who understands how the software works, what variables can affect its accuracy, and what rigorous measures must be taken to resolve any errors and verify the veracity of the results.
Essentially what I'm saying is that, due to the nature of this kind of work, it is inevitable that public access to programmatic approaches to frame analyses is going to lead to inaccurate reporting, so this isn't really a 'trdrop' issue, but a credibility issue.
At GameXplain, we will do our best to make sure that our frame data reports are as accurate as possible, but there is nothing that can be done about people misusing public/open source software like this. Will that cause people to become generally more distrustful of frame graphs? Possibly, but there are more objective ways to demonstrate the accuracy of the results of a frame rate analysis beyond a front-end GUI, and should we be challenged to do that I'd be more than happy to.
Yes, you're right, the tool isn't the issue here at all - just the usage of it. It's true that our tool is also not a simple to use thing at all (though we have some huge new changes that are almost ready - which is exciting). I trust people like you, VG Tech, NX Gamer and the like to get it right as you guys understand what you're doing!
Not a locked 60 at all. There are small drops throughout the video with it dipping to high 40's during super moves. Still fantastic overall performance though.
He is very useful for literally being one of the only sources of Switch version footage of Japanese games and without commentary, and at 60fps.
Samurai Warriors 4 DX on Switch for example only has a couple uploaders at 60fps and all have commentary.
Thanks for the work you put into the video! Can't wait to pick up the game myself.