Editorial/Mixing Guide

Getting Started

When you accept a card on ClickUp you will receive/download assets from the OneDrive that a filecutter has already cut, renamed and organized. Rarely will you come across any weirdness or corrupted files. Let one of the Audio Department Leads know if you do run into a problem that wasn’t caused by you and ask if it’s intentional. Do not trim the file length and do not add any effects unless explicitly required. The filecutters have instructions they are familiar with, and the game engine handles reverb. Your job is repairing, cleaning and delivering.

All delivered files should have the following specs:

Bit Depth – 16 Bit-Depth
Sample Rate – 44,100Hz (or 44.1kHz)
File Format – .wav files
Channels – 1x (mono)

If you are doing a batch card, can batch process in any DAW of your choice. Just ensure that you can process audio as separate files.

If you don’t have a proper DAW and need one, Reaper is free and pretty intuitive!
Please see the Batch Processing page for more information on that technique.

If you don’t have your own hotkey software solution, you can use AutoHotkey (a free software for Windows) and create your own custom Hotkeys for any processes in your DAW or your Computer that require keystrokes. If you need further guidance, please reach out on Discord.

The general process of mixing/mastering dialogue for Skywind is as follows:

Cleanup: Pops, clicks, plosives, extraneous background noise, hum, etc.
Processing: EQ, de-ess, compression, etc.
Loudness Control: Normalizing the average loudness of all files to fit a specific range
QC: Checking the final content against the script to ensure accuracy

Please utilize our Mix References that you can download from ClickUp. These samples show the standard volume of a normal line is -20 LUFS, while more projected lines can be as high as -16 LUFS and quiet more whispered lines as low as -22 LUFS. There is a range of vocal types in these references, so keep that in mind when you are working on your card.

Final Note: Regardless of the process you use when mixing/mastering voice lines, always check and listen to all the files before delivering.

1. Cleanup

Before any processing such as equalization and compression, you want to clean up the extraneous noises that distract from the performance/quality of the recorded dialogue. Many of the major issues that might come up during the recording process are addressed by the time the filecutters have finished, but any detailed cleanup is up to YOU, our esteemed VA editor/mixer!

Having some kind of audio cleanup software is ESSENTIAL to this process. We recommend working with some version of iZotope Rx. If you have Rx8 or higher, then all the essential tasks/modules recommended for this are included in your version.

Every card will require a different amount of cleanup. There is no ‘one-size-fits-all’ module chain for the cleanup process. This requires your keen eyes and ears! We highly recommend using your DAW using RxConnect to send chunks/batches of audio to Rx to analyze and send back to your DAW when done cleaning. Make sure you send as individual files and not a continuous file. Please ask us on the Discord if you need further clarification.

The listening part is fairly straightforward. You want to listen to some of the voice lines and get a feel for what can be done to repair them, enhance them, bring them together, match the vibe of the already existing voice lines and ultimately just make them sound more pleasing than before. If you feel there is too much background noise, use the de-noise and de-reverb modules incrementally.

Here are some screenshots of the common Rx modules you will likely utilize during the cleanup process. These settings are a good starting point, and you can increase the intensity of each module as needed. For example, on most female voices you can increase the frequency for the de-plosive, as female voices on average have a higher fundamental frequency.

Voice De-noise

De-plosive

Dialogue Isolate

De-click

Mouth De-click

2. Processing

Your goal is to preserve the natural feel of the lines, so avoid going overboard with the processing. Remember you want to hear a voice, not a pop-vocal. Once you’ve got an idea of what you’re dealing with, make sure to set a mark in your project whenever you find a tonal change. Pray that you won’t find any of those in the card you’re working on, because it can quickly become a nightmare of automation.

While listening, use 'clip gain' to gently increase or decrease the volume of specific parts of the voice lines if they seem too loud/quiet and catch your ear. I like to do that while listening, because it’s a time saver. One more important note, for this signal chain to be most effective, make sure the meters of all of the dialogue is roughly -12dB (range of -6dB to -18dB), this is just a suggestion, trust your ears and your meters!

Here is a simplified sample processing chain we recommend you start with. Keep in mind that you always have the freedom to change your processing chain, it doesn’t have to be (nor should it be) the same each and every time. Pick and choose what you need, and how much you need it. We will provide screenshot examples of specific plugins, but feel free to use any version of these plugins you are comfortable with.

Equalizer (pre)

You generally want a hi-pass and lo-pass filter to start. Equalization before compression is corrective EQ, so this is where you want to reduce frequencies. In the screenshot you can see the ‘yellow’ Mid-Frequency is used to reduce a problematic frequency band.

- Hi-Pass/Lo-Cut: Q = 18, Freq = 40Hz
- Lo-Pass/Hi-Cut: Q = 24, Freq = 20kHz
- Low Filter: Q = 1, Freq = 60Hz, Gain = -3dB

2. De-Ess

Make sure your de-ess is BEFORE compression, you don’t want your compressor to make your sibilance (sss) louder, thus this happens before compression.

- Threshold: -6dB

3. Compression

Since you’ve used clip gain to roughly level out the dialogue, you don’t need the compressor to be overly aggressive. Generally don’t push the ratio past 4:1, and adjust the threshold so you have no more than -12dB reduction, average should be closer to -6dB reduction. If your compressor has the Auto-Gain option, try it out! Otherwise consider a 3-6dB Gain makeup.

Ratio: Average 3:1 or 4:1
Threshold: Average -12dB to -18dB

4. Equalizer (post)

You generally want to push the brightness/presence at this point, as EQ after compression is generally considered ‘additive’ and enhancing the sound.

Hi-Shelf: Q = 1, Freq = 6kHz, Gain = +3dB

5. Limiter

Only use as needed, just in case your dialogue is really all over the place. This makes sure nothing is peaking or exceeding excessive thresholds. By setting the threshold and the ceiling to the same dB, the overall volume of the clip won’t change, but only affects what dB that limiting will occur.

Ratio: 30:1 (and higher, pending on the plugin)
Ceiling: -3dB
Threshold: -3dB

If you feel the noise floor is still an issue (the noise between the lines) there are a couple steps you could take.

Another stage of dialogue-de-noiser with a reduction of 3-6dB.
A gentle gate with a low threshold and a low ratio can pull the noise floor down ~6-12dB without affecting the dialogue. In this case, use a longer release time.

3. Loudness

When it comes to loudness, your goal is to make sure your files are all perceived at an even volume in the game, as well as in your session. If you do not have iZotope Rx, then you can use the YouLean Loudness Meter. Not only are you checking to see if the dialogue is similarly loud between lines, but you must measure the volume of lines in LUFS to make sure they are within the correct range. Once all your editing is done, it’s fairly easy to run all your files through batch processing with iZotope Rx! It makes this part of the process go a lot easier/quicker.

As stated in the mixing overview, the loudness requirements are:

Loudness average of -20 LUFS
Overall tolerance of +4/-2 dB
- The Range is between -22 LUFS (for quieter/whispered lines) and -16 LUFS (projected, yelled lines)

This means the majority of your lines should average around -20 LUFS, it doesn’t always have to be exact, but a good number to shoot for could be +/-1dB around that -20 number. Projected lines are often perceived louder than normal lines, so they can go as high as -16 LUFS (do not go higher than that!). On the opposite end, quieter more whispered lines can go as low as -22 LUFS (do not go lower than this!)

Here is a screenshot of the iZotope Rx Loudness Control Module to reference. This is the general settings you want to set for most (if not all) of your audio files. Again if anything is super projected and feels like it was made too quiet, you can bump it up to -16 LUFS. A nice thing about this module is you can simply highlight the line, and it will automatically measure the current levels (shown on the right for example), and will be RED if it means you need to adjust.

The last thing you need to do if you have iZotope Rx 8 or higher, is the Phase Module. This helps prevent the file from peaking and balances asymmetric waveforms. A fancy way of saying it helps make sure the audio file plays well with other audio files :D. It’s very simple, and here’s the settings you need to set it up with, just the ‘Adaptive phase rotation’ setting.

4. Team Feedback

This part is important! You absolutely need to check in with your peers, get their opinions and suggestions and take them into consideration. They often point out things that you might have missed and this is what keeps everything in line and on the same page. The peers have always been a great resource. It can take a while to get feedback at times, in that case you can already start working on your next card and get ready for approval.

5. Refining

After getting those suggestions and opinions on your Card, refine these voice lines accordingly and consider implementing those suggestions into your workflow from the get-go next time. Be very gentle with the changes you do from this point on, because you’ve mixed the entire thing already. Do not undo what you’ve mixed already. Surely, many of them sound good already and all you have to do is zoom into a few and get them fixed.

6. Delivering

Remember, the files should be exported as 16-bit, 44.1kHz, Mono, wav files. This should not have changed or been converted at any point in the editing/mixing process! If for some reason you get files delivered in a different format and you converted right off the bad, the delivery specs never change.

Make sure you:

Export/Bounce/Render your files (each audio file separately)
The name of each file has to represent the one that they originally had in the folder you downloaded.
Put all the exported files into one big folder (there's a guide on how to name the folder on ClickUp)
Zip the folder up (regular settings, nothing fancy)
Upload the folder to OneDrive (ask one of the Leads about it if you don’t have access to it yet and where exactly to put it).

And that’s it!

You’ve now completed your first card. Hooray! It gets easier the second time and EVEN easier the third time so please don’t shy away from a little challenge initially!

It feels very nice to know that your stuff is being used in a huge mod that tens of thousands of people are going to play once it’s released.

This is a guide, if you have a slightly different process that is OK! Trust your ears and trust your meters, you got this! Please reach out to the leads if you have any questions or need clarification on any of this process.

Page updated

Google Sites

Report abuse