Overall, I agree with MusikBear. But I can add a little bit...
NOTE: I wrote a lot in this message, but to get the best info, just skip to the bottom of my message and read the Reddit discussion/comments and then come back and read what I wrote. Overall, I messed up a tiny bit because I didn't address average (RMS) levels or even peak (LED) levels, I just talk here about LMMS fader/knob levels:
I've been noticing that LMMS's default fader/knob levels sound good, but are possibly set too high for multitrack projects with a lot of layers, or where the original audio samples are fully normalised.
If you use "pattern-track"s or the "audio file processor" (perhaps for OGG or WAV drums) then you'll notice that that the volume knobs default to 0.00 dBfs which is "full scale" (100%).
That means there's no gain-staged headroom automatically, you have to manually draw down the knob/fader to something more manageable, taking into consideration that you'll be mixing down together several waveforms and files and audio streams before the final output mixture.
Even if you only mix together 2 fully normalised WAV's, each at 100%, you have to somehow balance them out, like -6 dBfs on each one. (-6 dB tends to = 50% down, +6 dB tends to = 50% up).
As soon as you add more than 2 tracks/WAV's (or OGGs, etc) then you've got to attenuate (reduce volume) of each more. Or course, because of different balance and EQ/effect levels, and their corresponding gain (volume) controls, it's easily complicated but workable--there's several places and ways to reduce volume down from 0.00 dBfs "full scalle".
In my own compositions, I've frustrated myself where I started out a project within the "pattern-editor" (step editor/WAV sequencer) and forgot to reduce several tracks' individual volume levels. Another solution is to assign each pattern-editor track to an exclusively unique mixer channel and then lower the mixer channel fader significantly.
Thankfully, it's not so hard to do in LMMS. It might seem tedious at first, but it's the kind of thing that mix engineers of every kind have to wrangle with during every project. Every DAW out there is kind of different about how it implements access to gain staging and volume controls, but it's still adequate in LMMS if we try.
I can understand why the devs chose 0.00 dBfs as the track default. 100% volume is going to sound good at first and doesn't sacrifice any power. But I do remember seeing some studio mix engineers in olde school studios during the 1990's strike their consoles (reset their consoles) to have faders around -15 to -18 dB fader marks instead of all the way up to 100% or at 0.00 volume detente.
Since we DO have an LMMS mixer, we can "strike our console" too(!) to be at -18 dB for example.
It's easy after assigning to a mixer channel, by double-clicking on the fader knob and doing data entry into the "Set value" window. It should say something like "Please enter a new value between -120 and 6.0:"
Nicely, this even hints at what type of internal headroom we have in LMMS per channel: There is (only) 6 dB above full scale before clipping. But this is not bad at all. While it might be less than other environments, it's still plenty to work with because the 32-bit floating point precision is not lost and can be worked with DOWNWARD towards the -120 dB. NOTE: If LMMS won't let you type in more than 2 digits and a mantissa (+|-), use your up|down arrowkeys to speedily zoom to a value.
We are still in good hands even if we attenuate a lot, becaues 32-bit floating point precision is professional, clean, and adequate. I would have expected 18 more decibels of dynamic range than that, but maybe those "missing" 18 dB are part of LMMS's built in headroom.
Admittedly, I am ASSUMING that LMMS using 32-bit floating point precision for it's engine, because that's the highest quality setting in the Export dialog. It's really not bad at all. 32-bit floating point is also known as "IEEE float".
Here's the TL;DR:
I think I'm getting too deep into details, so suffice to say that, because 0.00 dBfs as fader/knob defaults is too high/surprisingly too close to clipping on already normalised audio, we need to manually draw down all of our fader knob levels to "our own defaults" of something like -27 dB (faders).
Similarly, the "sample-track" volume control can be double-clicked for data entry and says something like "Please enter a new value between -96 dBFS and 6.0 dBFS". That implies 16-bit control, but no worries that's still plenty to work with if we work downwards.
Anyways, I'm no expert on this stuff, but hopefully this is helpful one way or another.
Sorry if this is too wordy, but I tried.
P.S. - I found this nice Reddit discussion about fader levels and gain staging:
Also incredibly useful info: https://pointprimerecordings.com/blog/v ... eat-sheet/
