Mastering a wide dynamic range piano track

Hi,
recently I have finished one of my piano compositions, it has wide dynamic range and now I’m trying to increase it’s loudness so the softer parts are more easily heard (so you don’t have to increase volume in your player). And since I’m going to upload it on audiojungle, I want people to hear it clearly when they click play.
Here’s the waveform (peak) to illustrate the situation:
It’s not a real piano, it’s a sampled instrument.

I have tried with a limiter to just squash those peaks, it worked to some extent but I couldn’t add enough volume without destroying the louder parts.
I’ve also tried a compressor with fast attack and release, and automated it’s threshold to make it compress only where I want, but again it gave bad results.

So the only option that seems to work is to automate the volume manually to squash loud peaks (which gives no audible artifacts/distortions), boost softer parts, and slowly and progressively attenuate louder parts where necessary.

Any suggestions would be much appreciated
Thanks!

Have you tried just lowing the velocities of some of the really loud notes? Or does that take away from the dynamics too much..

If it’s just a solo piano, I would automate the volume to balance everything out. It will sound better than trying to squash your track somewhere around 12db to even it out. You could automate volume through the track or use midi cc11 (expression) to smooth things out. CC11 would be the cleanest choice since that will affect the piano sample only and not effects or sends you might be using on the track (like reverb).

But if don’t want to do that and you did ask for suggestions, you could use a compressor and an upwards expander to get gentler dynamic processing. Use the compressor to address the loud sections (maybe around -3 to -4 in gain reduction) and the expander for the soft parts (maybe +3 to +4 in gain makeup). You would then have 6-8db less in dynamic range. However, you might hear those effects working so you’ll need to dial in the settings on them to have them perform well.

Basically both options are doing the same thing so it’s a question of do you want a plugin to do it or do it yourself? I think you’ll make more ‘musical’ adjustments over a plugin.

JonBuice said
Have you tried just lowing the velocities of some of the really loud notes? Or does that take away from the dynamics too much..

Lmz said
If it’s just a solo piano, I would automate the volume to balance everything out. It will sound better than trying to squash your track somewhere around 12db to even it out. You could automate volume through the track or use midi cc11 (expression) to smooth things out. CC11 would be the cleanest choice since that will affect the piano sample only and not effects or sends you might be using on the track (like reverb).

But if don’t want to do that and you did ask for suggestions, you could use a compressor and an upwards expander to get gentler dynamic processing. Use the compressor to address the loud sections (maybe around -3 to -4 in gain reduction) and the expander for the soft parts (maybe +3 to +4 in gain makeup). You would then have 6-8db less in dynamic range. However, you might hear those effects working so you’ll need to dial in the settings on them to have them perform well.

Basically both options are doing the same thing so it’s a question of do you want a plugin to do it or do it yourself? I think you’ll make more ‘musical’ adjustments over a plugin.

expression automation is an excellent idea, definitely better that volume, especially because there is a long hall reverb that will cover things up, thanks

may be some parallel stereo compressor with 3-5 db reduction without pumpink effect, also hardcliping -3db (for peaks -2 db), and limiter for -6db area. dither and bounce it’s all

Try Antress’ 1176LN clone (extremely good):

It’s the plugin titled “The Seventh Sign”. Incredible for piano.

Solo piano sounds best with most of the transient left in. A manual ride on the fader makes the most sense, then, hit it with your color compressor -3db early, slow and and transparent.

A good compressor should do the trick…I wouldn’t select bits and pieces, I would just run the whole track through a good compressor and adjust the volume so that the peaks don’t clip.

I agree that some of the Antress plugins are good and free.

Instead of hitting hard with one compressor, maybe you should cascade several with a bit of softer settings, and if your DAW have a latency compensation, put the compressors on parallel and automate the track, so you’ll have more of the compressed channel in the silent parts and more of the clean one on the loud ones.
I remember we did something like that when we recorded some contemporary opera where the singer was insanely good and had a huge dynamic range.

Only I can propose it’s gentle multiband compressor use in 0-250 hz frequency band it should give some db to raise volume level and normalize it.