life threatening sound bug
Posts: 2181
Posts: 10665 | Subs: 9
Posts: 273
That one is epic I m sorry for you Romeo but it was really funny to watch
Posts: 3052 | Subs: 15
Posts: 254
Posts: 1740
Posts: 879
A reverb is nothing but a series of echoes that are filtered. To do this digitally you feed the delayed signal back into the original.
We know COH2 has memory management problems. I start lagging at start after 4-5 games. What is probably happening is the bad memory handling is resulting in stuff getting stuck in the buffers for the reverb effect, producing an overload that results in a digital glitch. I think the reason that it's so much louder than the game sound is because the ambient effects subroutine is somehow not tied to the game master volume, it's running on the sound files and those sounds are controlled by the master volume so normally you shouldn't have an issue. But if the reverb/ambient subroutine is bugged then it's perfectly possible for that sound glitch to be louder than the rest of the game.
It's nothing like any glitch I've heard in any effect, and it's not the driver getting stuck (if this happens because of a CPU overload and the entire driver crashes you usually get a fraction of the last sound played infinitely, which sounds like a constant tone). I've never heard anything like this though when I was mixing and experimenting no matter how many reverbs I feed into delays into reverbs and so on...But software for musicians is usually better coded.
If you want to know more (sound design and coding is fascinating) try this article: https://christianfloisand.wordpress.com/2012/09/04/digital-reverberation/
Posts: 122 | Subs: 2
I'm 100% certain it's related to the badly coded reverb in the game because we can hear the massive reverb on that sound. To my ears, COH2's reverb sounds cheap, metallic and poorly implemented (I have used DAWs and done my own mixing for my band and others so I've worked with a lot of software effects and sound design.)
A reverb is nothing but a series of echoes that are filtered. To do this digitally you feed the delayed signal back into the original.
We know COH2 has memory management problems. I start lagging at start after 4-5 games. What is probably happening is the bad memory handling is resulting in stuff getting stuck in the buffers for the reverb effect, producing an overload that results in a digital glitch. I think the reason that it's so much louder than the game sound is because the ambient effects subroutine is somehow not tied to the game master volume, it's running on the sound files and those sounds are controlled by the master volume so normally you shouldn't have an issue. But if the reverb/ambient subroutine is bugged then it's perfectly possible for that sound glitch to be louder than the rest of the game.
It's nothing like any glitch I've heard in any effect, and it's not the driver getting stuck (if this happens because of a CPU overload and the entire driver crashes you usually get a fraction of the last sound played infinitely, which sounds like a constant tone). I've never heard anything like this though when I was mixing and experimenting no matter how many reverbs I feed into delays into reverbs and so on...But software for musicians is usually better coded.
If you want to know more (sound design and coding is fascinating) try this article: https://christianfloisand.wordpress.com/2012/09/04/digital-reverberation/
Good theory. Mine is buffer overflow.
They submit 3 sound effects to the sound buffer. Lets say the data stream has a header which contains Volume V followed by content C.
| sound 1 | sound 2 | sound 3 |
|[V1][ C1 ]|[V2][ C2 ]|[V3][ C3 ]|
pretend data: |[10][sf1fdf]|[15][df3t3d]|[10][7fdg31]|
Although they only submitted 3 pieces of data into the buffer, they request sound 4 played.
| sound 1 | sound 2 | sound 3 | sound 4? |
|[V1][ C1 ]|[V2][ C2 ]|[V3][ C3 ]| [V?][ C? ]|
pretend data: |[10][sf1fdf]|[15][df3t3d]|[10][7fdg31]| [99][gfddd]|
The memory addressed by sound 4 is garbage - partially overwritten data from previous data stream or unallocated memory.
The sound processor goes and:
Play sf1fdf @ volume 10
Play df3t3d @ volume 15
Play 7fdg31 @ volume 10
Play gfddd @ volume 99 <-- romeo dies
Posts: 1194 | Subs: 1
Hopefully relic will pay for my tinnitus.
Posts: 16697 | Subs: 12
Posts: 2885
Good theory. Mine is buffer overflow.
They submit 3 sound effects to the sound buffer. Lets say the data stream has a header which contains Volume V followed by content C.
Code
| sound 1 | sound 2 | sound 3 |
|[V1][ C1 ]|[V2][ C2 ]|[V3][ C3 ]|
pretend data: |[10][sf1fdf]|[15][df3t3d]|[10][7fdg31]|
Although they only submitted 3 pieces of data into the buffer, they request sound 4 played.
Code
| sound 1 | sound 2 | sound 3 | sound 4? |
|[V1][ C1 ]|[V2][ C2 ]|[V3][ C3 ]| [V?][ C? ]|
pretend data: |[10][sf1fdf]|[15][df3t3d]|[10][7fdg31]| [99][gfddd]|
The memory addressed by sound 4 is garbage - partially overwritten data from previous data stream or unallocated memory.
The sound processor goes and:
Play sf1fdf @ volume 10
Play df3t3d @ volume 15
Play 7fdg31 @ volume 10
Play gfddd @ volume 99 <-- romeo dies
Yep that sounds right. It explains why it can be reproduced on replays and it shows how badly coded the game is. That's exactly why in all voip and telephone standards voice volume is coded twice with checksum - to ensure it's always correct. It may be also foult of usage of pointer arithmetic: maybe some other method is writing over sound buffer changing the loudness.
https://www.youtube.com/watch?v=ojrsL2b9Fr8&t=14m15s&t=14m15s
Can you guys hear that a minute later the smoke granade does mortar explosion sound? It may lead to the buffer theory being true: the bad sound from the buffer is played.
Posts: 4951 | Subs: 1
Posts: 1468 | Subs: 4
I guess I'm not sure on the specifics but wouldn't each audio driver/system handle the allocation/processing of data slightly differently?
Although, that could explain why some people never have it, and others do have it at times.
Maybe we should have like 40 brave souls see if we all get it during the same game that Romeo cast?
Posts: 379
Posts: 879
Posts: 2470
Maybe we should have like 40 brave souls see if we all get it during the same game that Romeo cast?
i'm willing to try
Posts: 3103 | Subs: 1
Posts: 1802 | Subs: 1
Somebody PLEASE blow up and GIF Romeo's webcam reaction to the soundbug/gunshot-to-the-head in that video.
PLEASE PLEASE PLEASE PLEASE DO DIS.
Posts: 4630 | Subs: 2
I have around 1400h with CoH2 yet I did not experience this pudding.
Posts: 211
Livestreams
133 | |||||
33 | |||||
26 | |||||
17 | |||||
6 | |||||
85 | |||||
16 | |||||
11 | |||||
8 | |||||
5 |
Ladders Top 10
-
#Steam AliasWL%Streak
- 1.829222.789+35
- 2.34957.860+14
- 3.587233.716+3
- 4.1095612.641+19
- 5.883398.689+5
- 6.280162.633+8
- 7.997646.607+1
- 8.379114.769+1
- 9.300113.726-1
- 10.717439.620+1
Replay highlight
- cblanco ★
- 보드카 중대
- VonManteuffel
- Heartless Jäger
Board Info
3 posts in the last week
23 posts in the last month
Welcome our newest member, service
Most online: 2043 users on 29 Oct 2023, 01:04 AM