Perceived increased loudness
If I pluck a simple 3 note note guitar chord and then a 2 note version, without changing finger position and plucking at the same strength, then I percieve the 2 chords as sounding at the same volume.
If I play the same combination in MS then the 3 note version sounds louder than the 2 note version. Is this simply my imagination or is something going on with MIDI playback that is different to a real guitar?
Comments
Probably not your imagination, and probably inherent in how synthesizers work. If you play a chord via MIDI at a given velocity / volume, each note is played at that level, so the more notes means a louder chord, probably on a logarithmic scale (?). This tends to also be true of most physical instruments where you can play chords, and of course, situations where the multiple notes on the staff are actually played by different people (e.g., soprano and alto reading from the same staff, or two flute parts sharing a staff in an orchestral setting). So it's pretty appropriate most of the time for more notes to increase loudness. But for guitar, I can easily imagine that each string you play uses up some of the force from your hand, so the overall effect is that each note of the chord is a little quieter than the last, and the overall effect isn't quite so additive as it would be otherwise.
Presumably a good AI-based synthesis engine could anticipate and account for this phenomenon - not just reducing overall velocity/volume to compensate, but doing it per note as per the strum direction. And adding a time offset as well.
In reply to Probably not your… by Marc Sabatella
Thanks. I'll apply a little velocity offset as you suggest but no time offset in this case since I'm plucking with 3, (then 2), separate fingers so there is no strum.
There's a flicker of a distant memory from school days about how sound waves of different frequency combine but I can't recall the detailed maths.
[Update] An offset of -5 was suitable.
So I decided to experiment with this a little bit. As we know, playback is certainly different from live sound. I tried this on my steel string guitar. It was hard for me to tell. Consider that the sound I hear from my position above (to the side) of the guitar is going to be different from the sound directly in front of the guitar. Which is where a recording mic would be placed. Also, which three strings and which two string I play could make a difference. Wound strings might sound louder. The construction of your instrument might make a difference. The wood used and bracing make a difference. And I'm assuming you play a nylon string classical guitar. There's a lot of things that might contribute to certain frequencies sounding louder or softer.
It was hard for me to consistently notice a volume difference playing my guitar. What i noticed more was the tonal difference between two and three strings. In MuseScore there was much less tone difference. But depending on the two strings played, sometimes there seemed to be a volume difference between two and three strings. Playback system also makes a difference.
I also tried several other plucked and even wind instruments. Similar results.
I'm not sure you can, or should try to, do anything about this.
In reply to So I decided to experiment… by bobjp
Yes, I play nylon string classical. These notes were all on the top 3, pure nylon strings, i.e. no wound bass strings involved, and I was plucking 2 or 3 strings whilst holding a fret 7 barre. There is quite a tonal difference with the thicker G string but this is less pronounced by fret 7.
Whichever combination of 3 or 2 that I play, it's hard to discern any volume difference from my playing position. On other hand, the MIDI version always sounds louder with 3 strings, (listening through headphones).
I'm arranging a 'modern-folk' style song for fingerstyle guitar and these particular beats in the song do not have any emphasis so it seems appropriate to apply a little offset in this situation – and it sounds better.
In reply to Yes, I play nylon string… by yonah_ag
Ah. An experiment I can't replicate because my G string is wound. My string length overall is longer than yours, also.
In reply to Ah. An experiment I can't… by bobjp
OTOH, because this is how I learn things, I went a little further. Because my hearing isn't what it used to be, I made a wave file (from MuseScore) of your three notes. D-F#-B. Half notes at 60BPM. All three, D+B, all three, D+F#, all three, F#+B. I listened through Three different headphones (including noise canceling), and two different sets of speakers. To my ears the first two sets sounded the same volume, using headphones. Using speakers, the second set seemed to have softer two stings sound. But in and analyzer it can be seen that every three string chord is a tad louder than the two string.
Playing my guitar, it's hard to judge. Anything without the low string tends to sound softer.
This makes some kind of sense to me because,as Marc pointed out, three notes are bound to sound louder than two.
In reply to OTOH, because this is how I… by bobjp
Interesting.
I can see that 3 notes should sound louder than 2 so maybe I'm compensating on my real guitar, or maybe I'm plucking with less force when there are 3 strings, (fingers i-m-a).
In terms of the score, I need the negative offset velocity to avoid accenting the beat.
Is your wound G string on a classical guitar? (I know that you said you have a steel string but maybe you have a nylon string too).
In reply to Interesting. I can see that… by yonah_ag
No nylon string guitar. What strikes me is the difference in listening experience between my real guitar and listening to MuseScore playback. When I listen to my guitar the sound is clean. I can pick out the tone quality of each string. They are very different. So I notice when I play only two strings. In playback the strings all sound the same.
This it something that most folks probably wouldn't notice. But you and I do.