Now that Darius Cozmic Revelation is out, it's a good time to do a comparison of all of the different home releases of G.Darius, to gauge their accuracy when compared to the original arcade version. While this kind of test has been done before, I wanted to do one that also offers up some quantitive data for comparison's sake, to provide better and more objective comparisons compared to the "magic and feelings" type of comparisons of the past.
We're running the both the UK PAL and the Japanese version of the PS1 game, the Japanese PS2 Taito Memories port, the original PC version, and both the PS4 and Switch versions of Cozmic Revelation for our comparisons, all on original hardware - UK PS1 Slim, fat Japanese PS1, Japanese PS2, UK PS3, UK PS4, UK Switch, and two PCs, an old one running Windows XP, and a modern one running Windows 10. We're also timing each clip from a defined start and end point to be as accurate as possible.
Check out the video for some interesting insights! You can find it here: https://www.youtube.com/watch?v=YLkNsLaajkI
Many thanks to チャーリーの7チャンネル, who has posted lots of high quality footage from the arcade PCB, of which we've used some in the making of this comparison video. You can find his channel here: https://www.youtube.com/channel/UCPheCd ... HcZ6tDth7A
G.Darius - Home Versions Comparisons & Data Analysis [video]
Re: G.Darius - Home Versions Comparisons & Data Analysis [vi
Thanks for sharing this.
After messing with the underclocking feature of mame for months, I still think playing the game faster than the arcade improves the experience. I would recommend that over all those pc and console ports (excluding the m2 releases of course).
Most of their 3D hardwares were a mess back then (ddg games have the shittiest framerates you can imagine, and their racing games have huge slowdowns), and G seems to struggles a lot during some stages, bosses, when loads assets, etc.
After messing with the underclocking feature of mame for months, I still think playing the game faster than the arcade improves the experience. I would recommend that over all those pc and console ports (excluding the m2 releases of course).
Most of their 3D hardwares were a mess back then (ddg games have the shittiest framerates you can imagine, and their racing games have huge slowdowns), and G seems to struggles a lot during some stages, bosses, when loads assets, etc.
Re: G.Darius - Home Versions Comparisons & Data Analysis [vi
Wow, nicely done. Thanks for your hard work. So is it safe to assume Cozmic runs the same in either HD or classic mode?
Re: G.Darius - Home Versions Comparisons & Data Analysis [vi
Good stuff. Not surprised though. M2 is just the gold standard when it comes to these ports. I can’t throw my money at them fast enough.
Re: G.Darius - Home Versions Comparisons & Data Analysis [vi
Aside from the ship model and general texture quality, there's no appreciable difference in speed.Rastan78 wrote:So is it safe to assume Cozmic runs the same in either HD or classic mode?
Re: G.Darius - Home Versions Comparisons & Data Analysis [vi
Great work, man. Thanks for doing all of this and sharing with the community.
Re: G.Darius - Home Versions Comparisons & Data Analysis [vi
Great comparison, thank you for doing this!
Results for the PC version were especially surprising - based on some scattered reports and it being based on the PS1 version I had assumed its overall speed to be less faithful to the arcade version than you've shown it to be (based on Zone α).
As you note, because you only measured the time passed between three reference frames, any more subtle characteristics of the differences remain largely hidden, e.g the frequency and severity of any individually discernible points of slowdown or "fastup". So reading your finishing notes, as an idea, maybe one could calculate motion vectors for the individual videos, e.g. by using some flavor of algorithm that calculates dense optical flow (two quick pointers). One could add up the motion vectors to obtain an overall amount of motion which could then be used to draw a nice graph that visualizes gameplay speed over time. I guess it'd remain to be seen how well it can work considering the subtlety of the differences, and the slight differences in gameplay will also add some noise (perhaps one could let the algorithm look only at a specific portion of the background to try and filter this out a bit), but maybe it could be worth a try with regards to any improved reissue.
Results for the PC version were especially surprising - based on some scattered reports and it being based on the PS1 version I had assumed its overall speed to be less faithful to the arcade version than you've shown it to be (based on Zone α).
As you note, because you only measured the time passed between three reference frames, any more subtle characteristics of the differences remain largely hidden, e.g the frequency and severity of any individually discernible points of slowdown or "fastup". So reading your finishing notes, as an idea, maybe one could calculate motion vectors for the individual videos, e.g. by using some flavor of algorithm that calculates dense optical flow (two quick pointers). One could add up the motion vectors to obtain an overall amount of motion which could then be used to draw a nice graph that visualizes gameplay speed over time. I guess it'd remain to be seen how well it can work considering the subtlety of the differences, and the slight differences in gameplay will also add some noise (perhaps one could let the algorithm look only at a specific portion of the background to try and filter this out a bit), but maybe it could be worth a try with regards to any improved reissue.
Re: G.Darius - Home Versions Comparisons & Data Analysis [vi
Interesting video. I think it´s great that you used footage from the original PCB as reference as IMO that should be the gold standard. Way too many comparison videos on YT claim to use original hardware, only to use MAME as a source in lieu of the original version. Not to hate on MAME but I believe if one doesn´t have access to the PCBs or direct captures, one doesn´t have any business doing these kind of videos.
-
MintyTheCat
- Posts: 2023
- Joined: Sat Jun 19, 2010 3:46 am
- Location: Germany, Berlin
Re: G.Darius - Home Versions Comparisons & Data Analysis [vi
That's crazy that there's that much lag involved. Really quite surprising given that they had access to the original source. In a way, I'd feel a bit cheated. I won't be playing the PS1 port any time soon
Good work, Icarus.
Good work, Icarus.
More Bromances = safer people
-
Bydobasher
- Posts: 261
- Joined: Wed Jan 26, 2005 4:16 pm
- Location: The Bydo Empire
Re: G.Darius - Home Versions Comparisons & Data Analysis [vi
Haha, I love how my preferred version of the game (PSX) is referred to as a "Tool-Assisted" version! It's a great video, thanks Icarus . Should get the Switch version sometime for sure....