Simpler way to process machine gun fire?

How to make that sound...
32 posts since 4 Aug, 2014

Post Sat Oct 01, 2016 6:42 am

I like to make sound effects, gunfire is most particularly for me. I have a processing chain set up in Adobe Audition CS6's Effects Rack, have it set up where I can process a sound in one click, to make things a lot simpler, because there's a lot of sounds I work with. However, while this works for a single gunshot, this becomes quite problemsome with machine gun fire, because there are multiple shots, instead of one. For most machine guns I can get around that by processing each shot, beat by beat, individually, but some guns like a Minigun shoots too fast where I can't count every individual shot, and thus are unable to process it.

Now I could just create a separate chain and process the machine guns in a more specific way, but the way I have this chain set up is essential to the desire sound I need or want. The way I have the compression, reverb, and limiter set up already destroy the machine gun sound if I process the whole loop in one go, which is why I have to do it individually, but the minigun as I said, shoots too fast. I need to be able to "mark" the audio beat by beat, processing it on repeat so I won't have to zoom in and manually process it, which is nearly impossible for a gun that shoots this fast, especially with a long loop. I know Adobe Audition has a feature where you can mark audio in segments but that is confusing to me, and it doesn't seem to be able to have a repeat feature, so again I'd have to manually mark it each shot by shot, and it's basically the same hassle as processing it shot by shot. If you get what I'm trying to say.

Is there a way to make this much more simpler and possible? Or will I have to make a separate chain and process the machine gun fire altogether, which I didn't want to do, because I didn't want the sound effects to differ in quality. Help would be appreciated thanks.

xalama qo
1114 posts since 7 Jun, 2007

Re: Simpler way to process machine gun fire?

Post Sat Oct 01, 2016 12:53 pm

I'm not too clear what exactly your processing aims are, but here's a shot in the dark...have you tried a transient shaper?

32 posts since 4 Aug, 2014

Re: Simpler way to process machine gun fire?

Post Sun Oct 02, 2016 7:32 am

Yes I have tried a transient shaper. That doesn't help.

"Processing aims"? I don't see where that applies here, I'm simply asking is there a way to process machine gun fire, especially in a loop, in a simpler way? And I thoroughly described what I meant by that, or at least tried.

In case you didn't understand, let me try to explain it this way

Here's an MG42 loop unprocessed ... sed%29.wav
That is the raw recording, now when I process the whole file in one go, it comes out like this (warning processed sounds are loud, and may be unpleasing to some people) ... ole%29.wav
As you see, the whole loop is ruined by reverb, compression, and many other things that are required in the chain to bring structure to the sound.
However, I can get around that, by processing each individual shot in the loop, beat by beat (or shot by shot, whichever is more understandable).
And it comes out like this ... hot%29.wav
Most machine guns don't shoot as fast as an MG42, so you'd think I'd come to a conclusion to my problem, HOWEVER, a Minigun shoots way too fast and I can't count each individual shot in the loop
Two different raw recordings of a Minigun: ... Loop+1.wav ... Loop+2.wav

As you see, I can't process it as a whole (like shown above), and then it shoots too fast to line up each individual shot. So what do I do? I can change the way I process the sound altogether, but then it'll have a different quality from every other sound and then it won't mesh well when added to a video game with other sounds that are processed differently. Is there a simple way to process a looping sound, is what I'm asking.
Here's non-looping if there needs to be a comparison ... sed%29.wav ... sed%29.wav

xalama qo
1114 posts since 7 Jun, 2007

Re: Simpler way to process machine gun fire?

Post Mon Oct 03, 2016 8:15 am

By "processing aims" I mean "how are you wanting to change the sound by means of processing?" (what result do you want to end up with and what are you currently using in your processing chain to try to achieve that result). I did understand your question, but it's really hard to suggest an alternative to what you're doing if you don't describe your current chain in detail. In your own words "...and many other things that are required in the chain to bring structure to the sound." We're talking about a complex process here right?

Anyway, good luck with your quest.

32 posts since 4 Aug, 2014

Re: Simpler way to process machine gun fire?

Post Sat Oct 08, 2016 8:25 am

Okay, if anyone is familiar with Adobe Audition I will describe as is. (long read below)
In Effects Rack I have the chain set up like this very currently (which uses incorrect methods but is necessary to get the sound I want)

(numbered by the slots)
1. Effects Chainer by Acon Digital (To add more slots, because native slots go up to 16, which is not enough for me)
Effects Chainer consists of these plugins
- Monomaker by dfx (to flatten the stereo image into mono, before processing which is required for some other plugins in the chain to work properly)
- MAutoVolume by MeldaProductions (to raise or lower the audio volume evenly, which is required for processing to work properly, I have it set to lowered)
- Envelope by MDA (to add an inverse signal, to help balance uneven bass levels in every part of the sound, and to add more structure to the sound, especially as its further processed)
- MAutoVolume (to raise the volume high enough to compress)
- Dynamics by MDA (to compress the audio of course, you would think I should use a more up to date compressor, but this one works best for me, because they way it compresses completely buffs the sound up evenly working well with the Envelope plugin)
- MAutoVolume (to raise the volume high again, because the compressor lowers the audio to -40db or so, and it needs to be higher in order for the next row of plugins to work)

2. WNS Stereo by Waves (A restoration tool. Most effective way I know how to completely clean up the audio in the high frequency range. Absolutely gets rid of most crackle, and background noise, especially after it's enhanced by the compressor)

3. Maserati GTi Stereo by Waves (Normally you're supposed to use this on instruments, but for whatever odd reason, this adds more body to both the sustain and attack portions of the sound, and even helps shapes the attack portion to add more thud/pow to it, which is awesome for gunshots/explosions)

4. MAutoVolume (to again raise the volume high, because the previous plugin lowers the volume a bit, plus I want the volume high for the next plugin)

5. L3-LL Multi Stereo by Waves (Normally you're told to use a limiter as the very last thing in your chain, but the limiter works best for me at this instance. I prefer this one in particular because I like the way it separates frequencies, helping balance levels, which most limiters I tried don't do, and it adds more thump/pow to the attack portion of the sound)

6. Channel Mixer by Adobe Audition (to flatten the stereo image back to mono. Previously in the chain, creating a inverse signal using the Envelope separates the audio in M/S, and I have to change it back to mono here for the next plugins to work right)

7. Sand Brush by 4Front (This helps add and fill in high frequency detail that was loss by the WNS restoration plugin that was used earlier in the chain. It seems to work best when applied at this instance)

8. OldSkoolVerb by Voxengo (Reverb to help smooth out the clunkiness in the sound. I like this reverb over other reverbs, because this one doesn't sound too chamber-like compared to most other reverbs I tried, plus it has features that dampens the audio to add more oomph to the sound, while also brushing out some of the unnecessary high frequency crackle created by the Sand Brush plugin used previously)

9. Latency Delay by Voxengo (to fix latency caused by the OldSkoolVerb plugin)

10. MORPH VST by Zynaptiq (Why use such a strange plugin? Because it recreates the sound, so any imperfections up to this point are mostly removed, including uneven bass levels, unbalanced frequency levels, leaky reverb, any further background noise that can't be removed by the restoration plugin. I've tried many other different plugins for this very reason, but this one is new and has been a gem for me so far)

11. PT_StableUnstable by Pitch Tech (Removes most of the syntheticness caused by the MORPH plugin, so it doesn't sound strange. Can also help remove unnecessary reverb)

12. MaxxVolume Stereo by Waves (Makes the attack portion of the sound loud in comparison to the sustain portion, so it sounds like a loud BANG with a lower decaying echo like a gunshot should. Can't find any other plugin that accurately does this, so this one works best for me)

13. GTR Amp Stereo by Waves (I use this to add warmth and crisp to reverse some of the muddiness that is created by the MORPH plugin)

14. X-FDBK by Waves (I've tried many different methods, but this one is new and works just the way I want it when tweaked right. Removes any other further unnecessary high and low frequencies without actually damaging the sound)

15. DTS Neural DownMix 7.1 to Stereo 7.1 by Waves (I've tried too many plugins to attempt to accurately recreate a stereo image, and they all have some sort of flaw. This one is a new one by waves, and can completely recreate the stereo image as if the sound was recorded in stereo. So every channel, left, right, center, have no delay, no reverb, nothing, so it's just a complete stereo image. This is a gem.)

16. Currently empty

The setup all the way up to the limiter and channel mixer is required for the most part. Everything else afterwards I change back and forth time to time, even sometimes maxing out the 16th slot. This setup I have now is my most current (even more up to date than the chain I used for the samples I provided earlier). It works perfectly for any gunshot or explosion that is used in one instance, however as I attempted to explain above, it becomes complicated with "looping" machine gun fire, especially when the machine gun shoots too fast (like a minigun) and the chain just doesn't work at all. Trying to create a whole new chain from scratch doesn't work for me either so far, because I can't seem to bring the quality up to how it is with this chain I'm using now. So I don't even know what to do. Either I'm gonna have to refrain from processing certain sounds, or if anyone has any solutions, would be greatly appreciated.

1 posts since 11 Oct, 2016

Re: Simpler way to process machine gun fire?

Post Mon Oct 10, 2016 9:36 pm

Ok, first there's way too much going on in your chain to get a good result. It might work ok on single gunshots, but that chain isn't going to lead you where you want to go.

Here's what I do, and I've been doing these kinds of weapons for movies and games a long time. First, get a good base layer. By that I mean a good quality raw recording of a weapon with the rate of fire you're going for and one where the transient peaks of the waveform can be differentiated by looking at that waveform in your DAW. For an MG42 or Mingun, it can be tough to find original material that is suitable because so many of the commercially available recordings aren't that good. This base layer is your "pace car" for building the rest of the sound. If the sound you're starting with is only moderately good, then you might have to bring out the transients a bit so that you can continue. The transient shaper suggestion above is actually a good one. I like the Voxengo Transgainer, but similar results can be had with a number of similar plugs. Start by punching up the peaks (often I need to turn down the overall output to get the peaks to poke out without distorting it too much).

Then, and you're not going to like it because it's not simple and it's not fast, is to start building up layers. For such a high ROF weapon, you're going to want really staccato and tonal layers here. I have a huge lib of processed "elements" for this kind of thing. Low end thuds created by various means, separated gun mechs of all sorts, and just different sweetener elements of all kinds. For the mech layer(s), for this gun I'd go with something very short and metallic. Perhaps a long file of suppressed gunshots from an MP5 or maybe an M4, time stretched a bit if necessary to get them to do the job and be short enough to be audibly single events. Could be metallic impacts, or other things. But have a bunch of whatever you're going to use and have them come from the same recording series (I usually use at least ten, but I don't know how extensive your library is).

So you're going to want to zoom into the waveform and carefully scrub from the playhead so that you can find each and every transient of the base "pace car" layer. Align your new elements exactly to those transients, and build away. Once you take the time to do the first layer of sweeteners, the rest go quickly because you already have the region start points of the first layer to work from. Just make sure everything is cut tight, as a few ms of slop in either direction is very audible and will destroy the effect most of the time.

Once you get a nice punchy set of layers going, add whatever sauce you want to the mix. Be it an electric motor whine carefully chosen like in the movie Predator, or some other thing with unifies the idea in a way that sells it as being a plausible weapon.

In my experience, there really is no shortcut to making huge gun sounds. You really have to take the time choosing good starting material, finding the right sweetener elements, and lots of carefully applied processing (and none that you don't need as that will only leave a noisy mess).

Trying to squeeze all the juice out of one sound with a monster processing chain is typically a recipe for bad results. It's just not that easy unless you have fantastic source material.

Hope that helps,

Return to “Sound Design”