Patrick, we can help you... forgive us for being a bit of a stickler on the terminology (spent a large portion of my youth working as a recording engineer). When you say “loud enough” that is very different from getting enough record level.

If you setup to record to your DAW, and you turn your speakers OFF... the recording will still take place.
Even though your speakers are Off (not very loud) record level can still be passed from the synth to the DAW.

HOW LOUD your speakers are is very different from the Record Level.

Yes, they are related, but they are also very different. It is this difference you must respect. Yes, if you are not getting enough Record Level, the Volume will likely be low. But turning up your speakers DOES NOT fix the real issue. So when you say it is not loud enough, you are referring to getting enough record level. The technical term is “Gain”.

So first let’s be sure you are measuring things in the right place. We assume you are recording AUDIO (you don’t actually say)?
If this is the case, then we need to establish:
1 - how you are routing audio from the MOXF to your computer? Are you using the MOXF as your sound card (USB) or are you using an external audio interface?
2 - what type of meters are available in your DAW for Audio Recording (what is the measurement scale, does it have a Clip indicator)? You will rely entirely on these meters (not ‘how loud’... ) for setting record levels
3 - what is your MOXF’s Output Gain Setting? [UTILITY] > [F1] GENERAL > [SF2] OUTPUT
4 - what you are recording and the type of metering you’re using will determine how you go about setting the Record Level. Percussive instruments feature transient peaks that can clip the level meter before you get fullness (presence) you require, this is where utility devices like compressor/limiters (leveling amplifiers) come into play. Learning to leave room for accents is also becoming a lost art. So good Record Level can differ depending on what is being recorded and how it is being played.

You don’t want to mix with your eyes alone — in other words, use the meters as a guide, let your ears make the final decision.

Let us know.

My problem is that the recording levels are too low for recording AUDIO.

Answers to your questions:

1) The MOXF8 is my soundcard. I use no external audio interface.
2) In Ableton Live 10 each clip has a meter. It turns green when the recording level is good and red when it is too high. When I record some audio the meter is turned green but too low, indicating that the recording level is too low.
3) The output gain setting is set to +6dB.
4) Things I try to record are sounds generated by VOICE or SONG.

I hope this info will be sufficient?

Well, it helps... the only way anyone can help you with levels is to actually hear the music and see the meters.

You say green is good, but you also say it is green but too low (?) I don’t understand. Do the meters have a scale (you know, numbers - typically these can help you get perspective, precisely) this can help you determine exactly what is “too low”. Saying generally too low, does not mean a problem, necessarily. Knowing when it’s a problem - takes Experience. Watching how erratic some signals are and how others can seem full and present in the mix without moving the meters much at all.

The issue most MIDI musicians have is that they don’t optimize (they actually, pay no attention) to the MIDI Velocity. MIDI Velocity will ultimately affect Audio Output Level.

Say you lay down Drums, Bass, Keys, Guitar, Percussion, and then you add a String Pad, instead of recording the String Pad with the same gusto (downward force on the Keys) that you recorded the other tracks, the Velocity is was too low... you wind up playing the string pad softly, with a light touch... after all you want them in the background.

Well, later when you want to transfer the string pad to an Audio Track, lo and behold, the output level is too low, and you can’t seem to make it loud enough. That’s an example of how it can happen — and DOES happen to a lot of folks.

Once you become aware of this you start to pay more attention to your MIDI data creation.

If you think about setting audio record Levels... in the studio the recording engineer is concerned about your minimum-to-maximum Level Range. Until you do something similar with your MIDI Tracks you may continue to get tracks that transfer with low levels.

For example, let’s look at MIDI Drums...(often it has a problem at the other end of the scale) once you realize that 127 is the maximum MIDI Velocity, you should recognize that this must be the absolute maximum any drum can be hit. You probably should not record all your snare drum hits at 127... because it leaves you no room to create an accent. Definition of an accent is that it is louder than the normal hit.

Get a feel for the dynamic range. Softest-to-loudest. Recognize when playing an acoustic piano that 127 is as hard as you can hit the MIDI controlled acoustic piano. When you play an acoustic, how often in a song do you hit it ‘as hard as you can’? The answer is hardly ever. If you have too many 127s in your Data this is going to be just as bad as not having any Velocities higher than 60.

Velocity on the String Pad might be 40-60 if you are “MIDI Mixing” ... this will not then transfer to Audio with much level.
I had one user who had his studio monitors up so LOUD, that when recording MIDI he never realized his strongest hit never exceed 90. Thinking that how LOUD his speakers were, would translate somehow to record level (of course, it does not).

What I am suggesting is not always a simple matter of increasing the MIDI Velocity... because when Velocity swapping is going on in the sound, simply increasing or decreasing velocity to adjust record levels does not always work... if you increase the Velocity on a snare drum it may change the snare sample that sounds. If you change the velocity on a slap bass you could change the articulation. So you have to be careful when adjusting Velocities, after-the-fact.

What I am suggesting is to become more aware of *how* you are playing when generating the MIDI data.
Take a look at your overall Velocities — after you’ve recorded something.... test what kind of Audio Output Level they are generating. You can definitely catch your mistakes early in the process, if you think through to Audio transfer.

Some times as a recording engineer you would adjust a musician’s headphone mix to change how they were playing... not to harm the recording, always to help the performer play better. Here’s how that worked... if you’re listening to yourself in a headphone mix, if you cannot hear yourself well, you naturally would starting playing harder. If the engineer detected a musician banging on the the keybed, turning them UP in the headphones would have a corresponding change (reduction) in the energy with which they were hitting the keys.

And there was quite a wide range you could adjust them before they even became aware of any change. You’ve experienced a similar thing when a friend is asked to put on headphones to listen to some music, if they attempt to talk to you, they wind up shouting at you — even though they are standing right next to you. They are listening only to their voice internally... and they make the volume UP adjustment so that it feels right, unaware that the balance between internal and external hearing is off. They are trying to raise the level of their voice to go over the music that (only they) are hearing... they never realize they are almost shouting! Same thing...

_Anyway look at your Velocities.
_Percussive hits move meters more than sustaining sounds.
_Recognize that the distances on the Meters only LOOK large... only a few people out of 100 can detect a change of 1dB, the majority of people can hear a change of 2dB, and if you can’t hear a change of 3dB you should do something other than music (kidding). But everyone hears a change of 3dB.
_Turn the volume down on your monitor mix, when recording the pad type sounds... so you play with the same physical energy you do your other parts. This will help generate decent Velocities on sounds you would “MIDI Mix” soft.
_Remember it is ‘human nature’ to play soft when adding soft parts... and MIDI makes this an issue... when later you try to get more level from a too softly played part. (If you were going to record yourself playing Direct as Audio, you would make the adjustments so that ‘too low’ levels are not an issue).

“Too low” on a meter, is never defined by a difference of 1dB.... why? It is hardly detectable. It may look big on a meter but it ain’t hardly nothing...

I will take some take to read your answer.
Thank you very much for helping me!

You can find two JPG-files in attachment showing my problem.
Hope this can help in some way.

As I mentioned the meters are showing you the amount of signal coming in... it is either a voltage type which is helpful in registering peaks. Drum/percussion and instruments like flutes and some plucked instruments have sharp peaks the are far stronger than the average output level of the sound.

Knowing whether your meters are reading peaks or the more general VU (Volume Unit’s). One is helpful in identifying/avoiding transient peaks and the other slower moving registers the average overall level from the source.

You’d need to tell us what instrument and if you can look at the MIDI VELOCITY that is generating this level, that would be helpful. The level showing could be fine since setting levels is capturing a range, from soft to loud... from a picture no one could tell much...

Otherwise it’s just a picture of a meter, from that I can’t tell if that the softest it plays or the loudest it plays. You don’t get any sense of range from a static picture.

I do notice that the level on that Input has peaked at -13.1 far more than half its range...
If ‘nearly maximum velocity’ moves this meter to -13.1 we cannot conclude if that is typical or not.

I think you know to check...
What is the Voice Volume?
What is the Part Volume?
and...
What sound is it?

If you are isolating one instrument sound, you can edit the sound for more output level but this would be for extreme cases only... for example if it is a pad sound, without much percussive energy it may not move your meters a whole lot. A performer can easily eat up the 13.1dB of room between the current maximum reached by that channel with a single accented hit...

Getting a feel for normal output versus an accent gesture on a particular sound is something that takes a lot of listening and watching sounds interact with meters. Assuming your music is utilizing dynamics, try to get feel for the accent... you want to always leave room for accents.

I receive many SoundCloud files from end users... if the wave shape looks like a “stick of butter”, I know what I’m in for... typically an onslaught of sound without much dynamic range. Peaks an valleys are more interesting to listen to than everything hovering near or at maximum. You do not win anything for having all track near or at maximum all the time.

Just FYI every Element has a +6, a +12, and a +18dB Boost... use only in extremely low levels (your levels are not extremely low necessarily (again it would depend on what sound it is).