macOS release preparation feedback (FEATURE COMPLETE)

● ARCHIVED · READ-ONLY
Started by ZeroBreak 9 posts View original ↗
  1. Okay, hear me out quick. Just now got into the Mac I decided on this week (nothing too up there because cost of goods) and I have my baseline determined as a result thereof. Only thing now is getting MV to play nice (for reference I settled on a 2008 iMac with El Capitan and set MV accordingly for my plugins upon receipt) and the one annoyance I have is that I have a serious lag problem (it's playable but the later cutscenes I wrote in advance will be entirely off key because of this, nor is it efficient in the least). Obviously I looked at the information regarding the overall issue of lag in another related discussion and unfortunately it didn't help in my case.

    I'm not 100% if using something higher than 1.4.x may be a potential bottleneck (I already have to run 1.5.2 on my Windows environment because plugins) but if someone knows a thing or two here I'd like to have as many ideas as I can as to what's up (as this issue apparently only happens with the macOS builds I'm now capable of self-processing).

    As an aside it's at its worst in full screen mode using a plugin that I have to do so, but which is tested to be unrelated so no feature pulls for the Mac version are expected. Also, I did install OBS for video capture should I need to do so, and which I can do if it helps to understand what's going on.
  2. What are the tech specs of your imac computer?
  3. Duh....

    (cue face palm)

    One thing I'm not 100% is if this is biased by a shell script because reasons, however what I'm able to confirm is Radeon HD2400XT, 4GB RAM and 2.4GHZ Core 2 Duo. Considering I don't see much in terms of overhead I would obviously think this is sufficient for testing at a bare minimum as I doubt MV really needs a whole lot from other discussions I've seen. That being said I did just find an app to create my packages in the event that extra testing may be of use beyond what I already have as my primary goal beyond play testing is actually getting the **** out there for extra feedback and such.
  4. Well, let's take a look.
    You have a 2.4 dual core CPU. Which might not be too bad for MV games, since javascript is single-threaded. However, as plugins pile up, the game turns more and more sluggish. Count in also poor MV optimization (and even worse one with plugins as plugin makers don't tend to optimize their plugins for speed) and you're in for a treat.
    4 gigs of DDR3 RAM are enough to play MV games, but it's not comfortable. I've had a 4 gigs of RAM laptop and if the maker wrapped their game in Enigma virtual box, I experienced crashes every 30 minutes.
    Radeon HD 2400 XT... 256 MB VRAM in your GPU is low, but the main problem is that this card supports OpenGL 2.0. It doesn't support WebGL (which is OpenGL ES - this arrived with OpenGL 4) forcing you to revert to slow Canvas rendering.
    The difference between Canvas rendering and WebGL rendering is huge. Canvas rendering is done by CPU, which hasn't been made for drawing graphics, so it's slow. WebGL operates on the GPU, which renders graphics 4x faster than a CPU.
    So I'd say there's a quite decent bottleneck of the game performance given your hardware specs.

    Btw. I have a Ryzen 5 2600X computer with 3.6 GHz frequency per logical thread, 16 GB RAM and 1050Ti... And still have lag spikes in mv.
  5. And of course I already added the ignore fix.

    I'll keep this going strictly for play testing but will bump up the recommendations accordingly (and possibly retime the affected cutscenes to compensate). For comparison's sake I will of course be mentioning recommended vs. required so there's no confusion. I'll go ahead and mark this as feature complete based on what I have to work with and defer accordingly for any testing that may be of assistance from third parties. Links to my testbed mirror are on the way for comparison's sake (again, I'm primarily adjusting based on low-end/high-end).

    The important thing for me is to know that it works so that I at least have something I can reasonably get feedback on. As a side note, the macOS upload will not be immediately available (I just now finalized the package so I need to update the mirror accordingly) and the link is also subject to change as I have effectively shifted away from my working title so please be extra cautious. When it's ready, the mirror for this release is linked below (it's still not the same as the final name as it relates to my working title but otherwise there are no differences to the effect). I'll follow up with patch notes that mention this limitation among other important gotchas.

    https://gamejolt.com/games/DaebreakerMightMind/359362

    EDIT: The builds are now live so I am officially ready to move on from this subject. There may be differences in how the inputs actually function compared to what's actually displayed (on Windows 10 they are intended to be reversed to resemble the input patterns of similar experiences from generations past) and by consequence may not be immediate and will require a one-time preparation on my end, which will hopefully be changed in the next build. Please also give close attention to other portions of the patch notes as they are equally important and will also introduce you to some new features that I've added over the time since the game changed to its final name.
  6. Final update on this matter. I did some research to determine my baseline for recommended specs in light of this and decided that I would jump ahead by just a few model years in terms of settling the matter by standard practice. Aside from the consideration of a resolution switch option I think I have determined what I am going to work with and have set my final performance target so this is officially feature complete and the matter settled.

    Again, I am primarily doing this for playtesting purposes as much as I am for the deployment itself so please bear with me if performance has its limitations on the macOS side for me. This obviously varies from one system to another so I am not concerned about that for my purposes as I primarily need to confirm function. In other words, I am not concerned about performance here so much as "if it works" meaning that raw power is not much of a priority (as my Windows 10 environment already sets the standard in terms of the upper end).

    Therefore, at this point all that's left is for someone to verify that my baseline works for them. Beyond that, my discussion on the subject is complete so I will defer to my posted mirror page for the download if anyone wants to cross check or see what I have to date (or perhaps even both). Additionally, I have noticed a difference between fullscreen mode and window scale on the macOS testbed where the latter option works better than in the other (a possible consequence of how MV does its thing I guess) and therefore it's not enough to ding my testbed unit as outside the minimums simply for providing me with a workaround solution. Thus I will be providing the option in the preferences on both PC-level deployments whether macOS or Windows 10 for customers to decide for themselves on which one works best for them.
  7. I'm officially going to close this subject today (and the mods are clear to make it official). But first, I'd like to announce that I have reworked the offending cutscene from scratch (while still using 99% of the original narration). It now plays in the form of a self-contained slideshow as a means to compensate for the differences in processor capabilities. And ultimately, that's what did the trick in getting the sequence to play as intended (checked even against the baseline in my targeted system requirements). I did have to do some extra work in this department... but only because of the subtitle portion of the transition process, of course.

    In case you'd like to see how it plays out I have uploaded it as a concept clip via YouTube. PLEASE BE AWARE THAT BECAUSE OF THE TENSION OF THE SITUATION AND THE ULTIMATE END RESULT I AM LABELING THE PREVIEW UNDER CLASSIFICATION CODE NSFW. AND WHILE I AM NOT RESTRICTING VIEWERSHIP BASED ON AGE OVER THIS (IT'S A DEPICTION IN A MANGA/COMICS AESTHETIC) THE ULTERIOR TONE OF THE SITUATION IS THE SAME NO MATTER THE FORM OF PRESENTATION. AND WITH THIS BEING CONCEPT MATERIAL... WELL, LET'S JUST SAY THAT MEANS SPOILERS ARE DUE.

    On that note, I would like to add that moving from in-engine to slideshow for the offending cutscene allowed me to expand upon the presentation relative to what happens compared to what I would've done otherwise. The basic idea is that it's the day a young Kyla Larson brings down a crime boss (that you wouldn't know from his accent) and gets away with it on the grounds of defending herself. The reason it goes down like this is because she bore witness to a hit job on her own mother... who just so happened to be a respected member of the community and a leading force in lifestyle innovation. And through an ironic chain of events, this moment plays into the plot later on in her life as is known to the crew by the time her flashback scenario occurs.

    Spoiler
  8. And that's it for this subject. On a semi-related note I came up with another deployment-related trick for my project in the time since to offer an extra touch of information during the unpack/register/install on the Windows side but since that's another subject to itself (and something relative to its own secondary effect) I'll save that part for another discussion.
  9. This thread is being closed, due to being solved. If for some reason you would like this thread re-opened, please report this post and leave a message why. Thank you.