This is a post most fascinating as it demonstrates clear video evidence of how to replicate some of this issue. Easy replication is incredibly helpful when diagnosing issues however, I have major problem with title of video and implication of post.
Within the first video I chose to watch you can see clear spike in CPU workload while there is decrease in GPU utilization indicating a definite CPU bottleneck here. I have utilized my fantastic mspaint skills to circle relevant bits of stuffs.
https://i.ibb.co/Mf4hNjV/p1.png - before Ambient is muted (GPU 66%, CPU 49%, FPS 132)
https://ibb.co/chxjNrw - after Ambient is muted (GPU 62%, CPU 72%, FPS 123)
During this moment your GPU was momentarily bottlenecked resulting in decrease of fps.
This is taken from; https://www.youtube.com/watch?v=fX5gt0Dhwd4
Second scene in above video demonstrates both rise in GPU utilization and CPU load. Before can comment further, are you perhaps playing at uncapped frame rate or 144hz? You may have answered this in post, many apologies if I have missed it. It is far more likely that the increase in GPU usage seen here is a result of scene rendering complexity vs the inn room.
Other question one lala has is related to OBS. Are you using cpu encoding or nvenc when recording these?
I feel must reiterate point I made in previous post. You absolutely cannot rely on temperature and hardware utilization alone.
It is significantly more common especially when running at higher frame rates to encounter a GPU bottleneck far before CPU bottleneck. You are much less likely to see total saturation of your CPU for reasons varied. This is one reason why for much time Intel outperformed Ryzen during gaming workloads despite having superior core count.
Anything ranging from architectural constraints to other system bottlenecks can easily prevent the cpu from being fully saturated, yet you will experience slowdowns. Opening task manager and seeing 72% utilization accompanied by slowdowns can be alarming as you should have additional headroom when you in fact do not.
Hmm, this is indeed in line with my theory. I was going to suggest overclocking to those willing. It does appear the sound engine is making far too many calls to the audio api then it should reasonably be. Easy solution for Square would be to temporarily limit the number of audio sources preferably dynamically, this would require them writing code that can detect abnormal audio latency then adjust accordingly. Plenty of open source stuffs on how to write such code in the emulator space.
Mah! Am happy I have been able to help! I am most woefully out of practice but also have many smart friends to help too. Buffering audio before it reaches your ears is incredibly common. It is done everywhere yet done so you will not notice. Increasing this buffer (I am making things simple please forgive) could also be temporary solution. However there is catch as in all things. Increase buffer too far and there will be audio delays. The goal when developing sound engine is to be certain that the game loop lines up with audio presented or delays will not be noticed. The more I look into the more it seems Square would have to increase audio delay to noticeable levels. Think casting spell then hearing effect of sound after spell is cast not as effect is occurring.
Yes, because the Embody team failed to test their spatial audio engine on wide range of hardware or to properly test period. Now Square must work with third party developer back and forth to resolve the issue. This is why in software development you should avoid working with out of house developers on mainline projects. This is idea most terrible. You do not do this. It is preferable to hire these developers to work in house directly, this means you have direct line of communication with them. It is unholy nightmare to work with third parties at times and better to write the code yourself.
Now Square audio engineering team must work with code they likely did not write and cannot fix. It is painfully clear from looking into this that because of this decision there will be major delays in implementing a proper working solution.
I adore this game but seriously, game packs are a terrible terrible idea. Perhaps on small indie project but not something with the scope and size of FF14. Whoever signed off on this goofed. There is no nice way to put it.
Also, cannot seem to replicate on PS4 myself, more testing needed. That is all!
Edit; The more I look into Immerse Gaming™ the more damning evidence becomes. Product on Steam that did not meet deadline, questionable user reviews, and system requirements that simply confuse; Requirements for their HIVE software - Intel Core 2 Duo or AMD Phenom X3
Hardly workhorse CPUs listed right here..
Last edited by MiaShino; 12-27-2021 at 02:40 PM.
Yes, changing from "performance" to "balance" would actually make things worse, not better. I think they did many things at once and only fixated on the last one that brought results. I have noticed it is pretty much in tune with server congestion at this point. No stuttering before prime time, lots of it during.
As far as flooding the devs with bug reports, I'm gonna pass. It won't help. And if anything will slow things down. There is no way they aren't experiencing the same issues on a AMD machine.
As a caveat almost all of Stormblood has been stutter free. Heavensward was horrible.
I would also like to add I disabled all other player effects and sounds, the only ones I see are my own and the enemies.
And while we are talking about "sound" lets also discuss the amount of british invaders infiltrating my ears drums on all characters. Even the samurais are ol' british chaps having a spot o tea. Does no one over there not speak normal english? And the 2 twins... jesus christ... everytime they speak I want to choke them. I've never heard an accent make my blood boil but theirs does. Btw I am Irish.
Last edited by XenophineEX; 12-27-2021 at 03:23 PM.
My takeaway was, that the audio issues are increasing strain on my system - not the other way around - hence the title. But I honestly don't know enough about any of this to make either statement. Am I wrong about this?
I think it's important to note here, that I actually had to go back and reseat my RAM after this video - video made me realise I only had half my RAM showing up in Task manager. If that makes a difference at all.
What would be a correct title - for any future watchers? What would be the correct takeaway from this?
I am aware that I've got a CPU bottleneck with my system. I mostly use this PC to work with adobe programs - gaming is a secondary concern - officially. In fact, I was able to 100% write off all my parts from my yearly tax report. Only reason I've not yet replaced my CPU/Motherboard is because that wouldnt fly. I basically build this pc based on what I could get the government to reimburse me for. I'd actually have to pay money to solve the CPU bottleneck :-PWithin the first video I chose to watch you can see clear spike in CPU workload while there is decrease in GPU utilization indicating a definite CPU bottleneck here. I have utilized my fantastic mspaint skills to circle relevant bits of stuffs.
https://i.ibb.co/Mf4hNjV/p1.png - before Ambient is muted (GPU 66%, CPU 49%, FPS 132)
https://ibb.co/chxjNrw - after Ambient is muted (GPU 62%, CPU 72%, FPS 123)
During this moment your GPU was momentarily bottlenecked resulting in decrease of fps.
Now - how relevant is the bottleneck for the audio issue?
I am playing at uncapped framerate. Yes. My monitor is set to 60hz. Supports up to 75 hz. I did record this using nvenc in OBS. Does this defeat the purpose of the video? Apologies, I do not really stream - I just make recordings here and there for personal reasons. So I dont really think much about my settings.This is taken from; https://www.youtube.com/watch?v=fX5gt0Dhwd4
Second scene in above video demonstrates both rise in GPU utilization and CPU load. Before can comment further, are you perhaps playing at uncapped frame rate or 144hz? You may have answered this in post, many apologies if I have missed it. It is far more likely that the increase in GPU usage seen here is a result of scene rendering complexity vs the inn room.
Other question one lala has is related to OBS. Are you using cpu encoding or nvenc (new) when recording these?
I am aware. I unfortunately made that video a few hours before your enlightening post some 10-15 pages back.I feel must reiterate point I made in previous post. You absolutely cannot rely on temperature and hardware utilization alone. (...)
Last edited by Elliah-Seraheart; 12-27-2021 at 03:42 PM.

I was lazy to do it at first, but I did it in the end. Please, everyone do this. The more bug reports there are for this problem, the higher it will climb on the priority list of bugs to be fixed ASAP. Not to mention that it helps with the investigation, as fixing bugs the devs can't re-create is a challenge and a half. Even with my measly insignificant programming experience, I know that.
Edit: I just checked the in-game bug reports page, and it seems many of these bug reports have been moved to "duplicate bugs". Now I'm not very familiar with these forums, I only made an account to post here about this issue, but a sticky thread on that page says the bug reports in "duplicate bugs" won't be investigated. I have a bad feeling about this.
Last edited by MediArgentum; 12-27-2021 at 03:32 PM.
Ohhh, only issue was in suggesting it may be gpu issue as I am fairly certain the audio engine does not perform any calculations on the gpu. Uncertain what proper title could be though.![]()
By the twelve no, not with this current awful awful market. I am still running one i7 4790K. These FX cpus are more then sufficient for FF14 and should they switch away from Dx11 to low level api (Vulkan) in the future you may even see decreased cpu load.
I believe exceptionally, though bottleneck has less to do with the age of cpu and more with the architecture its self.
Ahhhh yup! I thought so because unless you were using some form of adaptive sync you would never see refresh rate sync to 123 like that. Also nope nope, nvenc was what I hoped you were using. If you were cpu encoding that might muddy results with it increasing cpu load quite substantially.
Tis okay! I am only hoping to assist others in figuring this out and helping make certain they look in the correct places when diagnosing. Mah! English is driving me bonkers!
The video was rather enlightening though, a 49% to 72% jump in cpu load simply by muting one audio channel is.. alarming. I believe no one here needs me to say that this should not be occurring. Something else I caught when re-watching, there was brief moment where the entire game process stalled when you muted ambient sound. Ummm roughly 10 frames I thinks. That is also well, strange.
Speaking of this, I have one curiosity door that requires closing. Would anyone be willing to run FF14 while placing their cpu under artificial load, then attempt to re-create the audio stuttering issue?
Last edited by MiaShino; 12-27-2021 at 03:33 PM.

What kind of load do you mean? I routinely ran FF with other stuff running in the background. Firefox, Discord, Bannerlord or Rimworld. Usually have no issues with it. Since the sound issues started up, I've more-or-less stopped playing FF except to get in, do a few roulettes on lower jobs, and then log off. Easy, non-story stuff that I can do without sound. Since the sound issues cropped up, it doesn't matter what I'm running - absolutely nothing or my usual batch of whatever. None of it has any effect on the sound problems, positive or negative.




If Bug Report A and Bug Report B are the same bug (duplicate bugs) Bug Report B will not be investigated, but Bug Report A will. Bug Report B will be moved to duplicate bugs, and Bug Report A will be the one focused on. That's what that kind of language typically means.
|
|
![]() |
![]() |
![]() |
|
|