Compression/Editing Advice for my Recording

[MrJoshie333]

Hey everyone, Josh here again

 

So I recently bought a Scarlett 2i2 Audio interface and have attached it to my keyboard, and I recorded a piece of music that I have attached to this post.

I used Audacity to edit the music - the only thing I actually used was the compressor.

Specifically, I did the following:

Threshold at -29 db

Ratio at 2.1:1

I left noise floor, attack time, and release time the same, and the "make-up gain" and "compress based on peaks" options I left off.

 

However, for some reason, when I exported the song, it came out very quiet - you can still hear it, but I had to turn up my volume very high to hear it. It didn't sound like that in Audacity. 

Also, I still had clipping, even though the sound is very soft!! (You may be able to hear it if you don't use headphones and turn the volume high.)

 

I'm not sure how professional people create music that is loud and full but doesn't clip. Mine is quiet and clips. For some reason.

 

Am I not using the compressor correctly? Do I need to use a limiter (and if so, may I have some advice on how to)? Do I need to adjust the gain when recording the keyboard? Or do I need to use an EQ to prevent my clipping while making it sound good?

 

Any advice would be very welcome!

 

Thanks,

MrJoshie333 

 

 

P.S. I am using the latest version of Audacity and a Windows laptop

Chopin Op. 9 No. 2 - Nocturne in E flat major.mp3

[john]

You missed off the "makeup-gain" option. When you apply compression it reduces the volume by the ratio amount above the threshold. So, above -29db, the volume would be reduced by a ratio of 2.1:1, or roughly 29 - 13.8 db = 15.2 db reduction. Your peak volume is now only -15.2 db.

 

If you apply make up gain, the compressor automatically applies gain to bring up the volume of the track. It makes it sound a lot louder overall, reducing the dynamic range (the range between the quietest sound and the loudest.)

 

Solution is, assuming your threshold is correct, apply make up gain.

 

You can of course apply gain manually post compressor.

 

Note, by applying make up gain you will of course bring up the noise floor by 13.8 db. So any noise on the track will be much more noticeable. You are likely to need a noise gate to remove noise between sounds and/or other noise conditioning.

 

See, no need to go to the audacity forums

[john]

31 minutes ago, MrJoshie333 said:

How do people record loud songs without clipping?

 

Nowadays, for pop mixes, they don't care too much about dynamic range. They often apply a brick wall filter (inf:1) after a more gentle compression.

• Minimise all environmental noise where you are recording
• Minimise fan noise from computers (try setting them up in a different room and using another room as the control room, and keep the live room as quiet as possible.
• Make sure you keep audio cables apart from mains cables, digital audio cables and midi cables to avoid crosstalk noise (badly designed computers and computer interfaces can still have some crosstalk.

Record at a lower volume to preserve dynamic range. Often broader dynamic range instruments need that extra headroom.

 

For a louder mix but preserving dynamic range:

• Scan your audio to find the peak volume.
• Scrubbing your audio will also give you an idea about how often you get close to that peak.
• Apply gain to take the peak to about -3 db
• Now apply a gentle compression, say 1.5:1. If you set your threshold at the volume of your quietest played note, the compression will be applied evenly across your dynamic range (keeps the relative loudness of your notes as you played them)
• Apply make up gain

The audio will be a bit louder without it completely trashing the feeling you put into the song.

 

With the peak being -3 db you still have some headroom for mixing etc and for different final mix treatments for different platforms.

 

If you don't care about about dynamic range (which would be unusual for a piano player) then you can apply a more severe compression. Brickwalls tend to only be useful for mastering engineers and even then only with certain styles, but perhaps 8:1 with a threshold of -3 db might be applied after doing the compression above. It will color the sound, and abnormally affect the louder notes, but that is the price of having a louder mix. If you used 8:1 with the threshold at the volume of the lowest note, with make up gain, relative dynamic range is preserved but the overall difference between the loudest and quietest will be pretty small, sacrificed to get overall volume.