Audio Feedback Thread

Ah well that would spoil the fun with the peeps trying to guess what engine config it may have...

 

Well, we are working on a toolset right now which will allow us to create our own loops from RPM sweep recordings. Then we won't have to rely on a 3rd party supplier any more. That's our 'early/quick' fix to a much bigger and complex solution for the vehicle engine sounds.

 

We need a lot of things

 

This first pass will allow us to record and process those recordings in to loops. So no, there won't be any resynthesis or anything like that. What the next stage should be is still to be decided. We still need to find a way to create the "unmuffled" exhaust sounds to complement a recorded sound.

But at least this way we will be able to add new content instead of using the same sounds again, and again, and again.......

 

Thanks for the feedback!

The windscreen and larger crash lengths in general were considered too long - that was from internal team feedback. I also looked at making the mid crash audio more "crunch like" rather than harder impacts - everything was starting to sound too 'samey'. I also then added a random amount of 'panel' impact into the impact sounds which is more prominent with the larger hits. The pitch of the crashes has not changed, but perhaps it's the panel part which you're hearing as it does change the overall timbre of the sounds.

The race transmissions have been tweaked to be more violent sounding. Once we have the cabin filtering sorted out, these will bed in to the mix a lot better I feel. I will check that one as you've highlighted it.

We now have the ability to alter the EQ of exhausts on the exhaust jbeam as well as the engine jbeam. So what I've done is the remove a lot of the exhaust bass at the manifold/header, and then add it back in on the actual exhaust. This is now flexible so you can have different amounts of EQ on a per exhaust basis. A stock exhaust will now sound more muffled as I'm reducing more high frequencies than before, but as you progress through different exhaust types this changes. This is an ongoing process, as with much of the audio in BeamNG, and it's all subject to change again as I play and compare, I get more feedback both internally and externally.

Why the changes? I prefer to call them improvements.

Will the audio be more precisely adjusted in a future hotfix? - I'm not sure what you're asking here, but if you're asking to revert back to 0.24 then the answer is certainly no. This is an early access game, things change and sometimes you simply have to get used to those changes. Changing the UI sounds is a classic example. We had quite a lot of negative comments on that when released, now nothing at all. We all get used to things which are not specifically better, or worse, and the beauty of early access is that we have the opportunity to change things and see how the community receive those changes.

 

It's quite satisfying now, isn't it

 

I've had another play with the transmissions, one reason is that I want to hear the gearbox output whine at lower pitch... so I'm going to redo them all again, hopefully before the next hotfix. Thanks for picking that up...

 

https://www.beamng.com/threads/audio-feedback-thread.44084/page-45#post-1457695

 

Hey - thanks for the well worded message!

Yes, it is the same V10. We're still not in a position to use our own source, or licenced, sweep recordings. But we're not so far away now with the tools. In terms of improvements, nothing was done to the actual samples, just the EQ settings in the engine and exhaust jbeam files.

Funnily enough, I replied to a thread on Reddit last night talking about exactly the same thing, so I[ll copy and paste a little bit:

--------
It's all about CPU usage. If a sound has a larger overall distance that you can hear, imagine a circle with the 'thing' making the sound in the middle of that circle.... The larger the circle then more sounds can be playing wherever the camera/listener position is.

Or another way to try to explain. You have 10 AI cars on the map. 5 of them are 1km away and 5 of them are 1.5km away. If the max distance is 1.2km then the sound system only needs to cope with 5 vehicles. If the max distance is 2km then it uses twice the amount of CPU to play the sounds.
--------

The max distances of the exhausts are now automatically adjusted based on the amount of muffling. 0 muffling has a range of 1.2km whilst full muffling has a range of 400m - of course if it's a quiet exhaust to start with, you'll likely not hear it 150m away as it'll be drowned out by asphalt sounds.

I will review this again, it's always a bit of a balancing act!

 

I don't think there will ever be a time where CPU for the physics would be compromised for audio.... What you say about the slider, that's exactly the same thing the person I was commenting to on Reddit said.... and frankly, it's a blooming good idea and I'm a little embarrassed that I didn't think of this myself. Just goes to show that it's always good to interact with the community.

The excitement of the potential of what BeamNG is, is what drives me too!

 

Yes, but that is different. The voice count is the complete number of samples being played. A single engine band uses two channels to be able to crossfade between then, but FMOD has some form of lag in releasing the channels of around 64ms it seems, so this adds another channel.

Elsewhere, we prioritise types of sound, highest, high, medium and low.... and you can decide what the priority type is, whether that's quietest, furthest etc.

The engine/exhaust naturally has the highest priority.

This distance we're talking about, is the distance from the listener position that you can hear the sound - as they are the highest priority shouldn't be cut by voice count because they are the highest priority. So a max distance slider for the eng/exh is a good idea for people with more powerful systems to increase the distance that those sounds can be heard.

We have another issue here currently with the AI cars as they are spawning in to the map at a distance far close than their minimum audio distance, so the sound of them jumps in - and jumps out if they spawn out too quickly. These sorts of things happen to limit the amount of CPU being used so we don't compromise the CPU the physics system requires.