Counter-Strike: Global Offensive - Page 153

On August 23 2012 09:43 wo1fwood wrote:

I'm basically going to repost what I did on the cs:go forum, basically you have to configure sound to your setup to make it good (optimal setting), which makes directional sound ok, however in order for it to be great a number of tweaks are still needed. Someone mentioned that go uses sound that doesn't penetrate walls, but I don't agree with the statement based on everything that I've heard in game and how I know wave propagation behaves in the real world (sound in go does not behave that way).


CS:GO Sound: Propagation Tweaks Needed

So I've been playing the beta for a couple of weeks now and being somewhat less experienced of a CS player I shouldn't be commenting on balance, however there is one area that I have a lot of experience in that I feel has somewhat missed the mark in CS:GO, sound.

Sound pre-patch around the 10th or so was pretty awful and discerning a lot of locations or enemies, bomb locales, etc was very difficult. That being said with the recent updates sound has been improved quite substantially. I now can hear at least mostly where enemies are or track their location using audio cues, however I was still left with the feeling that a lot more could be done to improve the audio in the game and what I consider one of the larger, though more subtle problems left in CS:GO.


Sound Wave Propagation

Sound wave propagation behaves rather oddly in this game. The main offender is the nature of how each sound propagates when interacting with the environment. For this see Example 1:



In this example you have a T and a CT on Office, and you are trying to discern the location of each other, however its somewhat difficult to do so due to how sound propagates on the map (happens on all maps). The T is moving from side to side in order to check where the CT is going to enter and the CT is trying to figure out where the safest entrance is. Now say the T moves to location B. What happens aurally is that the CT cannot tell through audio cues whether the T has moved from the hostage room to the B location. In actuality the CT hears the T most likely at C and you can see why that could be problematic should he choose to enter through B. Overall what the CT hears is a linear movement that is propagated longitudinally from the T center point outward (see line c1a).

Now the above situation is how sound propagates, but in a very simplified fashion. What we currently do not have is wall or object material, density, or reflectivity affecting how the sound emanates from one place and is received at another. In terms of dB, due to the fact that sound traveling through the B location occupies more ambient space, is longer, and less direct than the A path, if the T moved to the B location from his starting point, to the CT should hear a subtle drop in dB and a more diffuse or muffled footstep cue. For the CT this small change immediately tells him that the T is not moving towards the A hall opening and may in fact be moving towards B or B1, or at the very least is near the far side wall in the hostage room. Now we can pick out movement much better.

Example Two:


The CT has the hostages, and T is trying to figure out where CT is trying to exit. Now because the distance from the CT to the T going down the back corridor is much longer than through the main entrance to the T spawn as well as having a substantially thicker wall (A) between them, footstep indications should be lessened overall if the CT were to move in that direction than if here were to come out of the T spawn main. This situation is slightly different from the last however as there is a large wall between the two players and the wave propagation has a significant travel distance to get from the CT to the T. What makes this different is that the overall audio cues should be greatly diminished, and much more polarized in how the sound gets to the T. For instance, should the CT decide to move down the back corridor, the T should begin to get cues almost exclusively from that direction due to wave travel distance and map/wall obstruction (assuming they are facing each other, the T would only be getting cues from his L channel, or very very minimally from his right).

What we currently get in both these situations however is a much more linear indicator of where movement is occurring, but determining distance, direction, and precise movements is extremely difficult due to the fact that the environment has very little impact on positional indicators. Because of that we hear closer to what those linear lines indicate.


Overall sound has been improved, but waves still do not propagate on maps in a natural fashion. They radiate longitudinally from the origin point (see circles) where only a rudimentary distance and position can be discerned. Environmental cues/considerations are almost nonexistent and is perhaps one of the most important ways to precisely determine movement, distance, and positional locations.