Plus custom charges.
Aka & Blue Type-R (Exa Arcade 2018)
-
thegreathopper
- Posts: 212
- Joined: Mon May 16, 2005 6:15 pm
- Location: London England
Re: Aka & Blue Type-R (Exa Arcade 2018)
Yeah, total would be over $4500 if i would be to import it.thegreathopper wrote:Plus custom charges.
Re: Aka & Blue Type-R (Exa Arcade 2018)
I'll wait 5 years for a console/PC version even if it ends up being changed from the arcade one, there's no way I'll be able to play this unless somehow my local gets their hands on a kit...
<Aquas> EDMOND DROPPED OUT OF HIGH SCHOOL TO SMOKE COPIOUS AMOUNTS OF OPIUM
<Zeether> shoe failed college again <croikle> credit feed
<Zeether> shoe failed college again <croikle> credit feed
-
StarCreator
- Posts: 1941
- Joined: Mon Jan 12, 2009 2:44 am
- Location: Maryland, USA
- Contact:
Re: Aka & Blue Type-R (Exa Arcade 2018)
If you're near Baltimore and can make it out to Marley Station, you'll certainly have the machine to yourself all day. No one is playing it, and I'm fully expecting the arcade owner to demand I put the emulator shitbox I had in the cab before as a placeholder back in because it made more money
Re: Aka & Blue Type-R (Exa Arcade 2018)
I did some testing of the games input delay today with nem's help, some results below:
LCD Sanwa T32H075 + I/O Taito JVS K9101262A
Aka & Blue game mode (movement input), I/O test mode:
6 frame delay
exA-Arcadia I/O test mode:
4 frame delay
CRT Nanao MS-2933 (via DAC) + I/O Sega JVS Type 1
Aka & Blue game mode (movement input), I/O test mode:
5 frame delay
Tendak converter + Exibel DVI->HDMI adapter
Reference point: CRT (both direct VGA and DAC) + xbox360 HRAP EX-SE
Mushihimesama Futari xbox360 v-sync enabled (1 delay):
3 frame delay
Radiant Silvergun XBLA
5 frame delay
Dodonpachi Saidaioujou xbox360 480x640 TATE (1x zoom, all extra windows off)
~5 frame delay (4 frames sometimes, but inconsistency isn't a positive)
Hypothesis:
- exA-Arcadia system has base delay of about 3 frames
- Aka & Blue adds 2 frames of delay (reasonable for unity engine game?)
- Particular LCD screen + I/O added about 1 frame of delay
- Overall A&B Type-R input response time seems close to Dodonpachi Saidaioujou on xbox 360.
My method of measurement was very basic so the overall value could be off by max 1 frame I guess. But I did a lot of repetitions for each phase. For reference the HDMI -> VGA Converter was tested not to add lag on Mushihimesama Futari xbox360 ver. On futari measured 3 frames of lag with vsync enabled using either straight VGA or Converter. That value matches with Mark MSX's input lag chart for that game. (I used a HRAP EX-SE)
This amount of lag is unfortunate since the game has some high difficulty patterns that look like Saidaioujou expert levels (was practicing the Stage 4 boss today, that is pretty wild) and fast bullets. But the game appears to lag equal to or more than the SDOJ xbox 360 port. Since the lag is so high seems worth trying to minimize it by all means, difference is more noticable while playing at these higher counts. I'm curious if others are getting different results like trap15 was mentioning.
LCD Sanwa T32H075 + I/O Taito JVS K9101262A
Aka & Blue game mode (movement input), I/O test mode:
6 frame delay
exA-Arcadia I/O test mode:
4 frame delay
CRT Nanao MS-2933 (via DAC) + I/O Sega JVS Type 1
Aka & Blue game mode (movement input), I/O test mode:
5 frame delay
Tendak converter + Exibel DVI->HDMI adapter
Reference point: CRT (both direct VGA and DAC) + xbox360 HRAP EX-SE
Mushihimesama Futari xbox360 v-sync enabled (1 delay):
3 frame delay
Radiant Silvergun XBLA
5 frame delay
Dodonpachi Saidaioujou xbox360 480x640 TATE (1x zoom, all extra windows off)
~5 frame delay (4 frames sometimes, but inconsistency isn't a positive)
Hypothesis:
- exA-Arcadia system has base delay of about 3 frames
- Aka & Blue adds 2 frames of delay (reasonable for unity engine game?)
- Particular LCD screen + I/O added about 1 frame of delay
- Overall A&B Type-R input response time seems close to Dodonpachi Saidaioujou on xbox 360.
My method of measurement was very basic so the overall value could be off by max 1 frame I guess. But I did a lot of repetitions for each phase. For reference the HDMI -> VGA Converter was tested not to add lag on Mushihimesama Futari xbox360 ver. On futari measured 3 frames of lag with vsync enabled using either straight VGA or Converter. That value matches with Mark MSX's input lag chart for that game. (I used a HRAP EX-SE)
This amount of lag is unfortunate since the game has some high difficulty patterns that look like Saidaioujou expert levels (was practicing the Stage 4 boss today, that is pretty wild) and fast bullets. But the game appears to lag equal to or more than the SDOJ xbox 360 port. Since the lag is so high seems worth trying to minimize it by all means, difference is more noticable while playing at these higher counts. I'm curious if others are getting different results like trap15 was mentioning.
Last edited by Zaarock on Fri Apr 17, 2020 7:39 pm, edited 13 times in total.
-
- Posts: 1754
- Joined: Mon Nov 01, 2010 2:00 pm
- Location: Massachusetts
Re: Aka & Blue Type-R (Exa Arcade 2018)
Speculation time: well, if eXA is basically a windows PC running games, the typical vsync queue is about 3 frames (d3d/ogl), that's the ages-old result.
So if that's what's going on, unless specifically using any method to force a shorter queue, this long a total delay is pretty common on vsynced PC.
What was the eXA's GPU already? these days there's driver options to force a minimal queue of less than a frame, someone mentioned it not long ago, dunno if it was from AMD or nVidia, it's kinda like hard GPU sync in RA, but I doubt eXA and devs have thought about any of that before shipping.
PS: love the bullshit uneducated lesson on lcd lag on the eXA website. whatever, for fuck's sake, if a display like the 32MP58 here has like 2~3ms (which is the real lag I have measured myself) then it's not magically adding more delay than a CRT just because of some untold natural hierarchy, we're not talking of special cases here, a 60Hz LCD measured quasi-lagless with a lag tester is just as responsive input-wise as a 60 CRT period, that's a FACT. Seriously, people and their beliefs *sigh*.
Also on those HD-FHD panels even long ago it was unlikely that scaling would introduce any additional lag, occurences of that are rather few anyway.
It's like they're mixing up together motion and input delay explanations randomly gathered like people who don't know much on the topic typically utter.
Instead of blabbering approximations with a list of cabinets they should stick a tester on the displays models and publish the results.
PPS: a simple DAC can't add lag by itself either of course, the Taito I/O however is suspicious, no matter if they pretend having pushed the limits whatever that means about it.
It should be possible for anyone to shave off about 3 frames by getting rid of the JVS I/O and forcing just 1 frame queue already, if accessing the eXA's AMD/nVidia GPU settings is any possible (reminder: just speculating, I don't own one of course lol)
But if I'm wrong and they have in fact pre-optimized the hardware minding lag already, then maybe most of the lag comes from the game/software, in which case there's probably nothing that can be done, ever.
Strikers1945guy wrote:"Do we....eat chicken balls?!"
-
Bananamatic
- Posts: 3530
- Joined: Fri Jun 25, 2010 12:21 pm
Re: Aka & Blue Type-R (Exa Arcade 2018)
Since it's a pc, does it have vsync on at all times or something? It's weird since they mentioned input lag was one of their biggest things yet it ends up being worse than a normal pc?
Re: Aka & Blue Type-R (Exa Arcade 2018)
That input lag is just strange, it runs perfect on mobile with touch controls. It can't be that hard to make it run perfect with arcade stick and pc hardware, which should have been optimised arcade use in mind.
Those tests pretty much just confirms how it felt when I tried it. Feels a bit (or a lot) like playing some shmups on Switch and it's not s surprise as those results seem to similar to what Mark is getting from some Switch ports. So playable, but far from perfect and not something I expected at all from this.
Getting it down to that 5 frames is a start, but I really hope there will be some solution to get it to that 3-4 frames level at least at some point.
Those tests pretty much just confirms how it felt when I tried it. Feels a bit (or a lot) like playing some shmups on Switch and it's not s surprise as those results seem to similar to what Mark is getting from some Switch ports. So playable, but far from perfect and not something I expected at all from this.
Getting it down to that 5 frames is a start, but I really hope there will be some solution to get it to that 3-4 frames level at least at some point.
Re: Aka & Blue Type-R (Exa Arcade 2018)
Had an error in my previous post, the LCD on Vewlix cabinet was Sanwa T32H07, not LG 32MP58HQ.
I'm curious if the unity side on A&B is optimized to reduce lag. On PC unity game executables can take launch flags from the user like forcing exclusive fullscreen mode.
Would help to have more than one game tested like this to see if all of them have a lag difference between the exa I/O test mode and game itself.
These arent super outlandish values or anything, 5 frames is just the treshold for 'high lag' imo. With 3-4 being average and 2 being low. So curious if the exa can be pushed into the 3-4f range on some setup.
I'm curious if the unity side on A&B is optimized to reduce lag. On PC unity game executables can take launch flags from the user like forcing exclusive fullscreen mode.
Would help to have more than one game tested like this to see if all of them have a lag difference between the exa I/O test mode and game itself.
These arent super outlandish values or anything, 5 frames is just the treshold for 'high lag' imo. With 3-4 being average and 2 being low. So curious if the exa can be pushed into the 3-4f range on some setup.
Last edited by Zaarock on Fri Jan 31, 2020 7:40 pm, edited 1 time in total.
Re: Aka & Blue Type-R (Exa Arcade 2018)
It'sZaarock wrote:On PC unity game executables can take launch flags from the user like forcing exclusive fullscreen mode.
Code: Select all
-force-d3d9 -window-mode exclusive
Re: Aka & Blue Type-R (Exa Arcade 2018)
Exclusive fullscreen using the GPU for vsync the default in AMD/nVidia drivers is 3 frames queue anyway, unless different settings/options are specified.
(how much's the default windowed fullscreen on W10 ? I think it changed over time but there's probably stuff you can do about that too, I've personally not looked into it since W10 such shite in that aspect I don't want to bother, 8 was already a pain. 7 for life ftw)
Again; we don't know for sure if it's either the game's engine or vsync that's generating the most lag.
PS: if eXA would publish the full software and settings specs (OS, device and drivers) not just random stuff for customers, that'd be great. ^^
Oh and, I don't really care since I don't own one, I'm just curious. :p
(how much's the default windowed fullscreen on W10 ? I think it changed over time but there's probably stuff you can do about that too, I've personally not looked into it since W10 such shite in that aspect I don't want to bother, 8 was already a pain. 7 for life ftw)
Again; we don't know for sure if it's either the game's engine or vsync that's generating the most lag.
PS: if eXA would publish the full software and settings specs (OS, device and drivers) not just random stuff for customers, that'd be great. ^^
Oh and, I don't really care since I don't own one, I'm just curious. :p
Strikers1945guy wrote:"Do we....eat chicken balls?!"
Re: Aka & Blue Type-R (Exa Arcade 2018)
Anyone else find the Stage 4 boss more difficult than the stage portions of 4 & 5 combined (including midbosses)? Guess the last boss is the same deal. Normal enemies are non-threatening which is strange for a CAVE-style stage 5.
edit: I measured the input delay on Dodonpachi Saidaioujou xbox360 with the same method, being average 5 frames (same as the usual lag chart) which is the same as A&B Type-R.
edit: I measured the input delay on Dodonpachi Saidaioujou xbox360 with the same method, being average 5 frames (same as the usual lag chart) which is the same as A&B Type-R.
-
mikejmoffitt
- Posts: 629
- Joined: Fri Jan 08, 2016 7:26 am
- Location: Tokyo, Japan
Re: Aka & Blue Type-R (Exa Arcade 2018)
For those measuring 4+ frames of lag - which version of the Taito JVS IO is in use? The version string in the settings should be able to tell you; even details like the manufacture date string can represent a significant difference. From my own testing, on an optimal setup it should be 3f from input to visible movement. Everyone says "The Vewlix IO" but I am doubtful that there is only a singular version with no variance between.
At JAEPO 2019, the Vewlixes that had been retrofitted with Blue Diamond monitors felt great. A friend's Vewlix FC with stock monitor has also felt great. And then, another cabinet with ostensibly an identical setup was noticeably worse. Another test, using a Capcom IO in a CRT cabinet, produced results somewhere in the middle.
So many JVS IO units has their own creative way of being a little bit shitty - it could be double-buffered inputs, not supporting the full command set, or deadlocking if polled too frequently. There's also "fun" involved with the fact that JVS uses USB cabling for non-USB signaling. The quality of the cable and subsequent signal integrity can be a factor in JVS's stability and possible need to re-send data or mitigate corruption. That can result in a longer polling interval at times as well.
It's unfortunate that the wide wide world of JVS cabinets have so much variety in their implementations. Even looking at a single model of cabinet, the matrix of possible JVS IO controllers and LCD display models can grow rather large due to revision changes or parts availability. It's very difficult to then say what the expected performance of one setup can be. Testing with a CRT using a DAC more or less eliminates the monitor problem, but there's also the surprising issue of JVS.
At JAEPO 2019, the Vewlixes that had been retrofitted with Blue Diamond monitors felt great. A friend's Vewlix FC with stock monitor has also felt great. And then, another cabinet with ostensibly an identical setup was noticeably worse. Another test, using a Capcom IO in a CRT cabinet, produced results somewhere in the middle.
So many JVS IO units has their own creative way of being a little bit shitty - it could be double-buffered inputs, not supporting the full command set, or deadlocking if polled too frequently. There's also "fun" involved with the fact that JVS uses USB cabling for non-USB signaling. The quality of the cable and subsequent signal integrity can be a factor in JVS's stability and possible need to re-send data or mitigate corruption. That can result in a longer polling interval at times as well.
It's unfortunate that the wide wide world of JVS cabinets have so much variety in their implementations. Even looking at a single model of cabinet, the matrix of possible JVS IO controllers and LCD display models can grow rather large due to revision changes or parts availability. It's very difficult to then say what the expected performance of one setup can be. Testing with a CRT using a DAC more or less eliminates the monitor problem, but there's also the surprising issue of JVS.
Re: Aka & Blue Type-R (Exa Arcade 2018)
There are at least 3.mikejmoffitt wrote: Everyone says "The Vewlix IO" but I am doubtful that there is only a singular version with no variance between.
Re: Aka & Blue Type-R (Exa Arcade 2018)
Thanks for the info. The I/O used is listed in one of my previous posts. nem said the JVS I/O on the CRT cabinet was 'Sega JVS Type 1'
I measured the delay in a way where visual response to input on next possible frame would be counted as 1 not 0, same as how Mark MSX and some others did this. Where 2 is usually the lowest possible value (arcade PCBs like Ketsui, various PC games with v-sync disabled), 3 being low for v-synced console & PC games.
When running the exA-Arcadias own I/O test mode without booting a game we got a result of 4 frames on the laggier LCD. On the CRT and Sega JVS IO that should be 3 frames (that's good). But the values were always 2 frames higher when running A&B or the games own I/O test. Perhaps this is game dependent?
Anyway this isn't an obstacle to playing the game, more of a feel thing for experienced players (maybe it affects how long they keep going after 1CC or similar like SDOJ's console port, but that's speculation). If input delay is partially game dependent and some exA games are down to 3 frames that sounds great and in line with any modern console or PC system.
I measured the delay in a way where visual response to input on next possible frame would be counted as 1 not 0, same as how Mark MSX and some others did this. Where 2 is usually the lowest possible value (arcade PCBs like Ketsui, various PC games with v-sync disabled), 3 being low for v-synced console & PC games.
When running the exA-Arcadias own I/O test mode without booting a game we got a result of 4 frames on the laggier LCD. On the CRT and Sega JVS IO that should be 3 frames (that's good). But the values were always 2 frames higher when running A&B or the games own I/O test. Perhaps this is game dependent?
Anyway this isn't an obstacle to playing the game, more of a feel thing for experienced players (maybe it affects how long they keep going after 1CC or similar like SDOJ's console port, but that's speculation). If input delay is partially game dependent and some exA games are down to 3 frames that sounds great and in line with any modern console or PC system.
-
mikejmoffitt
- Posts: 629
- Joined: Fri Jan 08, 2016 7:26 am
- Location: Tokyo, Japan
Re: Aka & Blue Type-R (Exa Arcade 2018)
Gotcha, thanks for the clarification. I don't have enough Sega units myself to give any concrete data, but my impression of the Sega JVS IO is quite poor. I've tested with the JAMMA-style one for Gimmick development, and it's been worse than the other IOs I've tested with. A friend's New Net City was disappointing with both Naomi games and a JAMMAfier.Zaarock wrote:Thanks for the info. The I/O used is listed in one of my previous posts. nem said the JVS I/O on the CRT cabinet was 'Sega JVS Type 1'
I measured the delay in a way where visual response to input on next possible frame would be counted as 1 not 0, same as how Mark MSX and some others did this. Where 2 is usually the lowest possible value (arcade PCBs like Ketsui, various PC games with v-sync disabled), 3 being low for v-synced console & PC games.
When running the exA-Arcadias own I/O test mode without booting a game we got a result of 4 frames on the laggier LCD. On the CRT and Sega JVS IO that should be 3 frames (that's good). But the values were always 2 frames higher when running A&B or the games own I/O test. Perhaps this is game dependent?
Anyway this isn't an obstacle to playing the game, more of a feel thing for experienced players (maybe it affects how long they keep going after 1CC or similar like SDOJ's console port, but that's speculation). If input delay is partially game dependent and some exA games are down to 3 frames that sounds great and in line with any modern console or PC system.
There will always be a bit of game dependency too, but I know A2B can do better than those test values. Since the CRT cabinet has the Sega JVS IO, that gives a little more sense to your figures.
Even though it's a bit of a bother, I encourage you to jury rig the Vewlix cabinet next to the CRT cabinet, with the video output going to the CRT cabinet while the Vewlix handles IO. If it's with a stock Sega IO, testing in that cabinet does tilt the playfield in a bit of an unfair way against the unit under test. Generally the Vewlix JVS IO fares better than those Sega ones.
Last time I played it, G.Rev's Strania EX had fantastic results, with 2f lag between stick press and visible character movement.
Last edited by mikejmoffitt on Mon Feb 03, 2020 8:29 pm, edited 1 time in total.
Re: Aka & Blue Type-R (Exa Arcade 2018)
Alright thanks, I'll ask if we can try hooking it up to a CRT with the Taito JVS K9101262A (or similar) I/O and see how it goes.
That Strania EX bit is very impressive if true.
That Strania EX bit is very impressive if true.
Re: Aka & Blue Type-R (Exa Arcade 2018)
I'm on a Diamond Black with a Sega Version 3 I/O and can't tell any noticeable lag. I have some Taito I/O's on the way eventually. Curious if that will make any changes or not. Right now I'm pretty happy with the Sega Version 3 though.
Re: Aka & Blue Type-R (Exa Arcade 2018)
Man, remember when playing an arcade game was just a question of plugging in a jamma connector, and you're go?
I'm starting to believe Cave made the correct choice.
I'm starting to believe Cave made the correct choice.
-
Bananamatic
- Posts: 3530
- Joined: Fri Jun 25, 2010 12:21 pm
Re: Aka & Blue Type-R (Exa Arcade 2018)
If only there was a way to play games where you get to decide what kind of hardware you use...
Re: Aka & Blue Type-R (Exa Arcade 2018)
Anyway, I found this post by Shou on Arcade-Otaku.
(one frame of lag from Sega JAMMA to JVS IO )In our tests:
3.5ms - Noir
4.0ms - Sega JVS Type 3
7-8ms - Taito JVS IO (2 variations)
15-16ms - Sega JAMMA JVS IO, IONA-JS
So Sega JVS Type 3 + CRT should be pretty much the best setup possible?
Re: Aka & Blue Type-R (Exa Arcade 2018)
Interesting, the Taito being half-a-frame wouldn't be much under the condition you don't have a monitor that adds another half.
A shame for that Sega which is as good as trash.
Anyway still no clue about the amount of vsync lag ? there is some, that's compulsory, and depending on how it is managed it can be either rather minimal like 1 frame, or significantly higher like 3.
Speculation again of course but I have a hard time imagining Unity is an engine that absolutely has to lag this much, I wouldn't be surprised it calls for an unnecessary long queue from the driver, and would it indeed, then it would probably be possible to mitigate the issue with an update that'd kill a frame or two.
EDIT: Apparently QualitySettings vSyncCount setting goes 0~4 (ugh imagine 4 frames)
A shame for that Sega which is as good as trash.
Anyway still no clue about the amount of vsync lag ? there is some, that's compulsory, and depending on how it is managed it can be either rather minimal like 1 frame, or significantly higher like 3.
Speculation again of course but I have a hard time imagining Unity is an engine that absolutely has to lag this much, I wouldn't be surprised it calls for an unnecessary long queue from the driver, and would it indeed, then it would probably be possible to mitigate the issue with an update that'd kill a frame or two.
EDIT: Apparently QualitySettings vSyncCount setting goes 0~4 (ugh imagine 4 frames)
Strikers1945guy wrote:"Do we....eat chicken balls?!"
Re: Aka & Blue Type-R (Exa Arcade 2018)
Some links from this post by MrJBRPG on Arcade-Otaku https://forum.arcadeotaku.com/viewtopic ... 50#p524750
https://docs.unity3d.com/ScriptReferenc ... rames.html
There is a value called maxQueuedFrames which is another variable that Unity devs can use to tweak latency.
https://docs.unity3d.com/ScriptReferenc ... rames.html
There is a value called maxQueuedFrames which is another variable that Unity devs can use to tweak latency.
And another thread with some discussion on various ways to reduce input latency. https://forum.unity.com/threads/underst ... ty.762161/Use QualitySettings.maxQueuedFrames to limit maximum number of frames that are queued. On PC, the default value is 2, which strikes the best balance between frame latency and framerate.
Note that you can reduce input latency by using a smaller maxQueuedFrames because the CPU will be waiting until the graphics card finishes rendering previous frames. The downside of this however, is that it can result in a lower framerate.
Re: Aka & Blue Type-R (Exa Arcade 2018)
Remains to find out how high it was set by Tanoshimasu and if it's indeed tweakable/update-able, either by the user or by the manufacturer.
In any case my money on: they've set it too high just to be safe, just like most game devs do, assuming they even pay attention to that.
In any case my money on: they've set it too high just to be safe, just like most game devs do, assuming they even pay attention to that.
Strikers1945guy wrote:"Do we....eat chicken balls?!"
Re: Aka & Blue Type-R (Exa Arcade 2018)
Don't know what options there are for pushing updates (online? patches?) but that's really the only viable option, seeing as it's a locked-down system that isn't meant to be modified by the end user.
Re: Aka & Blue Type-R (Exa Arcade 2018)
Been playing the game hooked up to a CRT for a couple hours now and it feels a lot better. Still got 5 frame response but it's better than 6 and some CRT properties likely help too.
Still 5 frames on CRT with both Sega JVS Type 3 (837-14572) and Sega JVS Type 1.5 (837-14505) *shrug*
Starting to enjoy the game a lot more now & stage 5 is a blast. Scoring and stage flow reminds me a lot of Akai Katana Shin, except with infinite katanas (legitimately fun)
Still 5 frames on CRT with both Sega JVS Type 3 (837-14572) and Sega JVS Type 1.5 (837-14505) *shrug*
Starting to enjoy the game a lot more now & stage 5 is a blast. Scoring and stage flow reminds me a lot of Akai Katana Shin, except with infinite katanas (legitimately fun)
Re: Aka & Blue Type-R (Exa Arcade 2018)
You should e-mail them.parodius wrote:Out of stock
If there's any input lag in the pre-ordering system then they might have 5-6 kits left. Or is that backward, maybe they sold out before they shut it off?
Breaking news: Dodonpachi Developer Cave Releases Hello Kitty Game
Re: Aka & Blue Type-R (Exa Arcade 2018)
Did you know if the sun disappeared, it would take us 8 1/2 minutes here on Earth before we noticed? I just thought that was interesting.
Damn sun lag.
Damn sun lag.