Compression, Production Techniques and Ardour Digital Audio Workstation

Well, this started out as a short email and morphed into a semi lesson and I thought I might include a few extra people because the info might be of some use to some of you.  I’m also thinking about writing a blog on the website in the kind of style, so I’d like some feedback on that idea please. Ultimately, I want to document everything I am working on so that we have a sort of library of ideas, techniques and projects.

After I made the first version of the Amcal DBK Apnea Spot, I quickly realised that it wasn’t actually very good and I needed a version 2.0. That’s because in the first version, the way I ran the compression on the VO (voice over) track emphasised Jutas in-breath in a hilariously awful way, making it sound like she was gasping for every word.  I couldn’t leave it like that.  I still don’t know how I missed it, but that first version went to air probably 4 or 5 times before I realised.

Below is some explaination of what I did wrong and how I fixed it.

Simple Version: we use an audio processing technique called compression to level out the difference between the loudest and the softest parts of any given bit of audio by running some maths on it.  We can use this to get rid of unwanted base room noise, to make the track stand out, to level out uneven volume levels, to suppress or highlight aspects of the sound.
If you’d like a detailed lesson in compression, ask Ian (haha) or if you must, you can ask me.  It’s not terribly difficult and I mostly use a cheat sheet of 3.5:1 compression ratio and then fiddle until it sounds good when I’m processing vocals.

Tech Details Version: I was running x42-plugins Dynamic Compression, tiny bit of input gain, 3.5:1 compression ratio (for vocal tracks I find 3.5 to be a pretty good default) short attack, and a medium threshold setting.

This had the effect of picking up the almost silent intake of breath on the track and amplified it, doing exactly what a compressor is supposed to do, but it didn’t sound at all bad in my headphones, I was pretty happy with the mix and I exported the whole thing, loaded onto the playlist and triggered it so I could listen over the air.  


Well.  Now I can’t unhear it, lol! So I loaded up the project and went through and clipped out all of the in-breaths, but accidentally clipped out the tail of several plosives.  It sounded super weird. This was when I discovered that wav editing in Ardour is lossless, so I only needed to extend the segment beyond where I’d clipped and it restored the final part of a couple of P’s and K’s, and so….

I leaned a heck of a lot about the Ardour interface doing this project and I’m really happy with the outcome. Next main project is getting the streaming up and running, so stay tuned!!


Any questions or comments most welcome.