How to make a sound pack and contribute to MB2 or an independent modification.

[k4far]

What is a sound pack? Sound pack is a set of files usually in .mp3 or .wav (I go with first) format that trigger when your character interacts (you are using taunts, jumping).

When a new model becomes available it requires assembling a brand new set of sounds for it. Sounds have to be clear, loud and fit the character that is all there is to it. Do you have a character in mind that could use a better soundpack, do you even know of some possible sounds already? Great!

View attachment 1931
Go download Audacity it is the best totally free program for playing with audio. It is used by all big companies, freelancers. You are not wasting your time that might prove useful to you in your future for your job, project... anything.

The requirements for a sound to work with our game:

Make sure file is mono and sampling must be 44100. These are essentials for our game. It will not play when things go wrong here. In game console will show you an error when you have a file which is not a "proper" one. It will specify what is wrong too so no worries. Game got you covered here.​

THE PROCESS:

I. FINDING APPROPRIATE SOUNDS

Uncle Goo... No, Youtube this time. The character you might be working on is usually known for some TV Series, Movie. You want to find the cutscene with him, compilations, short video you can go over.

When you are downloading the file make sure the bitrate is the highest possible. The higher it is the more detail you can hear and the easier it will be to prepare it.

II. WHAT TO PAY ATTENTION TO

There shouldn't be any noises (as little as possible). Taunts can be short like "Hello" - first one. "The time has come" - the second. "Let's finish this" as last. For it to sound good it can not be too long taunt is usually a quote the character is the most known for. Due to a lot of various quality clips, it is understandable it might not always be possible to achieve. Go with what you believe to fit. I always try to put a comical twist into my sound packs, one line out of context but adding a lot of meaning to different situations. Test in the game, see if the noise is destroying the impression if it is minor it is fine (keep in mind the sound have to be loudened, they are usually pretty silent at first.

III. NOISE

Test it in the game, see if the noise is destroying the impression. If it is minor, hardly noticable it is fine (keep in mind the sound have to be loudened, they are usually pretty silent at first at the end they are 10+ dB we will talk about making them louder in a bit).

a) *Advance technique I. You can try to "amplify" part of the sound down giving it minus value via amplifying function in Audacity. Took me a while to do it properly. You go a few times by lowering using low values like - 2dB and listening to it over and over. You do not want sudden drops in loudness however!

b) *Advance technique II. Credits to @TechShaman. He has told me it is possible to use a copy of song playing in the background (which is extremely rare, usually it is just some shooting and slaughter) set is as noise profile then use the earlier function to remove noise and Audacity will do the magic.

c)*Advance technique III. Let's say we have a a short line "Hello, nice to meet you" if you listen to it very slowly and then try to play each part separately you might notice where the noise is the most noticable. You can use that fragment as noise profile "get the noise profile" and quite succesfully get rid of the noise from whole track.

IV. WHEN IT IS NOT WORTH YOUR TIME

When you believe the noise is too big. When you can not hear clearly because of something happening in the background. If the noise in the background is continuous you are unlikely to get rid of it by modifying the file in Audacity. You do not want any static, deformations. If you like music you know what I am saying - all things unpleasant for ears.

V. LOUDNESS

a) "amplify" function,
Audacity will increase the volume to the maximum as it sees it and it will not cause issues, but that might not always be enough so in rare cases you can turn a blind eye to capping and increase it a little more... just make sure to test it.

b) @Shenghai has reported the "normalize" function can offer lower quality loss from loudening than the "amplify" function.

VI. TESTING

Test it out. Join the game if there are no red text console errors it is a great sign meaning you did first part well.

I usually go for an empty Dotf Server and usually as Rebel since they have the elevator which you can use for testing damage and dying sounds and a generator next to it so you can test the falling sound as well. It is all quite time-consuming, consider if the sounds are loud enough that depends on your settings (I am using 10 percent of my max loudness for testing.)

- basic sounds used by weapons can be found in the GameData\MBAssets.pk3\sound\weapons
- default voices for character can be found in the GameData\zz_MBModels2.pk3\sound\chars

I am quite new to sound editing but I find it extremely enjoyable as my part-time activity. It is very nice to see it in the game later and used by people. I also wanted to help out in some way with the Development of MB... modelling was not for me I tried I tried I am saying (damn it). Coding neither (damn you math too). I am aware there are some mistakes, not everything is explained as yet I am putting this down after a busy day I had a thought to do it since noone bothered before me. May you find some fun in what I could I wish you very much.

Feel free to ask for help.​

 

[Shenghai]

Just to add on here, make sure the .wav is saved as 16 bit, 32 bit won't work.

Also, instead of the amplify function, I would suggest the normalize function. It doesn't hurt the quality as bad in my experience.

I can also add onto this: How I gave my mandalorians a helmet filter.

Once you have a sound you're ready to work with, make it a stereo track. Then, vocal remover while retaining frequency bands (I'd recommend not tinkering with this). Then apply a high pass filter (You can play around with this a bit) and normalize. Make it mono again, and export.

EDIT: Also, you can go into MBassets3/, find the folder for the character you're replacing, and in modeldata.cfg, you can add a specific soundpack for a specific character (like mandalorians).

 

[Noob]

You should never edit existing pk3s. instead create a pk3 using winrar or whatever archiving program you wish. Have the folder path to the sound then arhive it as a .zip (important not to archive as .rar) then rename the file to zzz_placeholder.pk3. Remember that quake loads pk3s in alphabetical order so when you have the replacement be the lowest possible name (with z's) you override the previous pk3

 

[eezstreet]

Should add some corrections to this...

anger sounds are played 50% of the time instead of a taunt
if the anger sound check fails (ie, 50% of the time), it will play a taunt1-5 instead
if there is no taunt4 or taunt5 it will play a random one from taunt1-3 instead
taunt.wav is only used for gloat/flourish emotes

bit rate of the audio file does not matter
it can be in wav OR mp3
correct sample rate is 44.1 khz / 44100 hz

 

[k4far]

I have in plans to write a bit about helmet and robotification effects. Also, how to use generic sounds and make something brand new out of them so advance pitching & playing with tempo/speed.

You should never edit existing pk3s...

I am currently using my own .pk3 file that overrides all other files but I have never experienced any issues by editing an existing archive. We can not edit a file that is being used by game in real time. Replacing sounds is a simple action and it should not cause any issues.

Bitrate of the audio file does not matter.

Have to disagree. Listen to a sound in 320 kbps and 64 kbps, you will tell the difference right a way - the detail and the clarity.
Have you ever listened to a .flac file on good headphones? Oh, you should!

 

[MaceMadunusus]

Have to disagree. Listen to a sound in 320 kbps and 64 kbps, you will tell the difference right a way - the detail and the clarity.

He isn't saying in terms of audio quality, hes saying the engine doesn't care what bitrate it is at.

Also, if you care about audio quality, you should never use a youtube video as your source. Youtube limits the audio bitrate to 128kbps or 192kbps depending on the source file.

 

[k4far]

He isn't saying in terms of audio quality, hes saying the engine doesn't care what bitrate it is at.

I did not say that it is making a difference for the engine. Well, the engine. I am only saying you can hear more thus clean it up better if you get a better quality file. It is quite important in the process of working on the sound to have it at best.

Youtube limits the audio bitrate to 128kbps or 192kbps depending on the source file.

As for the bitrate: Recommended upload encoding settings - YouTube Help
It can be up to 512 kbps actually with five point one surround sound. Stereo is a thing already and it supports up to 384 kbps on YouTube.

 

[MaceMadunusus]

It can be up to 512 kbps actually with five point one surround sound. Stereo is a thing already and it supports up to 384 kbps on YouTube.

Yes, but no one uploads in that and that change was fairly recent. So like 98% of what you find for movie quotes, is gonna be in horrible quality.

 

[Puppytine]

I have never experienced any issues by editing an existing archive. We can not edit a file that is being used by game in real time. Replacing sounds is a simple action and it should not cause any issues.

It will cause issues with repairing and updating -- Launcher will replace files you've modified with official ones, so you have to repack your sounds after each repair or updating.
Also it's hard to partially rollback changes if you forget which files you've modified; the only options you'll have are leave it as is or perform a repair, eliminating all changes at once.
Also to redistribute your custom sounds you'll have to place them to a separate package, since I hardly doubt that people will like an idea of replacing original assets.

Short version -- editing existing pk3s fucks up versioning. Versioning is important.