Is there any objective evidence that different keys "feel different" or have different moods?

Question

In an advertisement in YouTube for composer Danny Elfman's master class viewable here the composer says:

When somebody starts talking about “this should be in such and such a key because such a key feels that way”, that’s bulls--t. You can turn any key into anything you want.

This is something that I've long suspected, and hearing someone who's recognized as "good at" intuiting how music sounds to the general public carries at least some weight.

Question: Has anyone put this to the test in some objective way? Is there any objective evidence that different keys "feel different" or can impart different moods or feelings?

I suppose subjects would have to be sorted in to those with and without perfect pitch to check for subconscious bias; if you definitely know what key a passage is in, you may potentially be influenced by that knowledge.

note: A test of this is also challenging because as a passage played on various instruments or voices is transposed through its register the sound will change slowly, so a test for a key-dependent effect might compare nearby keys in order to minimize those effects.

Answer 1

There are other questions here that have dealt with this topic. Basically, there are good reasons why historically certain keys sounded different from others, due to the fact that unequal temperaments were more common, and thus some keys would sound more "in-tune" than others. These aspects of tuning often influenced the choices of genres for writing music in particular keys, so different keys developed associations with emotions or extra-musical characters (e.g., some keys were considered appropriate for "martial" or military type music, often because military instruments were pitched in them, so marches were often written in them).

That last bit is still potentially an issue today: many instruments are still pitched in standard keys, which means they are easier to play (and easier to play in-tune) in some keys than in others. Modern instruments are mostly equipped with various keys and slides and such to help correct for the problems of playing in less-common keys, but there may still be subtle variance in tone, say, for brass instruments playing in a key they are pitched in vs. a key that is a half-step away (which will require much more tuning adjustments, will potentially change the tone slightly by using more valves, etc.). For string instruments, one might consider that sharp keys often take advantage of open strings, which alters the sound. And unless you're talking about truly professional violinist, their ability to play in a key in tune with many flats compared to their ability to play in a key with a few sharps will vary significantly.

So, from a practical standpoint on many instruments, it certainly does "feel different" to play in different keys. Does it therefore also create a different mood? No, not necessarily. But differences in tuning and tone can make a difference, and this was part of the historical origin of the characters of the keys, back when most instruments were quite difficult to play in all possible keys, and various keys would in fact sound quite different due to these practical performance issues.

As for today, the effects are relatively small, and if you're dealing with professional players, asking them to play the same piece in G major or G-flat major or F major isn't going to make much difference. However, if you're dealing with amateur ensembles, there are potential significant differences to asking a group to play in G major vs. G-flat major, and some of those will come out in the sound of various instruments or in the fluidity with which instrumentalists are able to play.

But as for moods and "feelings" associated with specific keys? These were always a bit of a convention, created through associations that built up over decades of composers using certain keys for certain purposes. There's nothing inherent in the sound of different keys that conveys those specific moods.

(The one exception, I suppose, would be for some people with perfect pitch who experience different types of synesthesia. I once knew someone who had a very acute sense of perfect pitch but also had developed very particular associations with keys, including moods, colors, etc.)

Answer 2

Elfman is both right and wrong.

With respect to traditional music, he is very wrong.

Suppose an instrument is tuned to a specific key, not with equal temperaments. A tune played in the instrument's key will sound different than the same tune played in a different key.

For instance, using the diatonic scale, if an instrument is tuned to the key of "C", the interval between "do" and "re" will have a frequency ratio of 1.125 (9/8). If the same tune is played on that same instrument in the key of "D", that same "do-re" interval will have a frequency ratio of 1.111 (10/9), which will sound significantly different.

Classical composers were well aware of this difference and made use of it. That's why their works often include the key in their titles. Bach's Toccata and Fugue in D Minor would sound very wrong if played in any other key.

Bach knew very well what the various keys sound like in meantone temperaments, and in well temperaments. Organs he played were in various versions of meantone, as a standard. (And some still are.) Other wind instruments were also designed to use those specific twelve notes: A-B-C-D-E-F-G, F#-C#-G#, Bb-Eb. The other notes were by design not as well in tune in the temperaments in practical use. Composers understood this. This was practical musical knowledge. Temperament — Bach Cantatas Web Site

But since that time, instruments and music have changed, losing the tonality that defined classical music. In some ways this is a great loss (no more associated mood), but in other ways it is a win (any tune can be played on any instrument, or sung in any key).

Pianos, and many modern instruments, especially synthesizers, are tuned with equal temperament, in which the interval between any two adjacent notes, including semitones, has a ratio of the twelfth root of 2. Using this system the interval between "do" and "re" will be 1.0593781, an approximation of the 1.125 and 1.111 of the diatonic scale. In fact, every whole tone difference will have that same ratio.

A piece written in equal temperament will sound exactly the same regardless of what key it is played in. And today, almost all music is written that way; it has no tonality.

So that is what Danny Elfman was referring to when he said "You can turn any key into anything you want.".

Answer 3

There's no evidence that I'm aware of that in an equal temperament world different keys have intrinsically distinct characters, in the sense that Bb major would relate to a different range of emotions than C major.

I guess I find it a bit surprising that you find Elfman's statement noteworthy - It seems fairly intuitive that with most people possessing only relative pitch, and in an equal temperament world, there wouldn't be any intrinsic correlation between home note and emotion.

In a given body of work, you might nevertheless still find correlations between particular keys and particular 'feels'. One reason might be the tuning of particular instruments - the standard tuning of the guitar could lead to a large number of songs in blues/rock tonality being written in E or A, for example.

I assume that when he says "You can turn any key into anything you want", he doesn't mean that C major has the same range of subjective feelings as C minor - rather that he's just talking about the home note. If he's stating that there's no significance whatever in major/minor, that would be a surprising statement.

Answer 4

If you play a folk song in D flat major instead of D major on a fiddle, it sounds like crap since all the resonances are off and lots of drones don't work. Even if you discount the tremendous added technical difficulty which "should not" make a difference to a perfect soloist but of course does.

Not every instrument is a piano in modern equidistant tuning.

Most wind instruments have compromise tunings combining perfect intervals and valve combinations. Not every wind instrument is a trombone.

Fretted string instruments with fixed frets are commonly based on equidistant tuning, but chord voicing and fingering still depends on the tuning, and the tuning again will often be a compromise different from equidistant tuning, making some chords have better qualities than others.

When you are doing your composition work at the piano or on synthesizers or on MIDI expanders with less than per-note sampling, it is easy to forget that most orchestral instruments are not working from a continuum of semitones and sound samples but from a limited number of co-dependent oscillators employed in various constellations and mechanical vibration modes that may take a variety of skills and effort to coax into different keys.

A brass ensemble will have several chords that sound awful because of the way the different compromises for the tuning of notes of various different instruments interact. Good players can adjust the intonation of single notes, but you don't get to adjust every note in a fast passage.

But if your main listening tool is a MIDI expander, indeed hitting the transpose button will not magically shift moods unless you have come to culturally associate certain moods with certain keys and recognize them reliably even in partly synthesized sounds.

Answer 5

This is the question, that I had been asking myself for quite some time, and though I am by no means a professional musician and do NOT have a perfect pitch I still think that the key is to some extent important.

I would not argue that a half-tone change will make the music sound different for me, but the large differences will sound differently. The reason is that instruments of the orchestra have their ranges (which are much less than that of pianoforte) and they actually sound differently in different regions and changing the key by a fifth up I would lose the sound that I had (e.g. violin would get too squeaky and horn less heroic).

So for me, if I compose something there will be a range of keys, where it sounds right, and the range where it sounds worse.

Answer 6

As someone who has perfect pitch, it's always seemed to me that every key, and I'm assuming you're not talking about different modes, every key has a different inflection or tone to me.

For example, I've always felt songs that are written in D major to be uplifting or more grand, and I couldn't really give you a reason as to why. In contrast, a song that's written in C major, in most cases, will bore me for some reason. Possibly, it could depend on what I'm used to and how many songs I've heard written in that particular key.

Often, a comparison to hearing pitches is seeing colors. Someone could love the pitch Db in the way they love the color red. And maybe, everyone has this happen to them no matter what way they hear pitches. I hope this answer helps to answer your question.

Answer 7

From a biological standpoint, I'd argue that it would be very odd if all keys were exactly identical to the average human brain. If you think about what's happening physiologically, sound waves are triggering certain cascades of neurons in the audio cortex, and the neurons triggered by the key of G major aren't going to be the exact same neurons as those triggered by C major. Many of them will overlap, but some will be unique to each key--otherwise we wouldn't be able to tell a difference between keys. Simply recognizing that there must be a neurological difference implies that how a key is experienced internally would be different as well. However, it may be such a minor difference that it is unnoticeable, and/or the "noise" of perception introduced by individual brains probably overwhelms any average difference that brains in general might experience between the keys. At the very least, we can say that mood/tone differences due to chord progressions compared to differences due to key are close to negligible. Doesn't mean there are no differences between keys, only that they are overwhelmed by noise and other factors. Still, I'd say the burden of proof is on those who claim there is NO difference. To me, the null-hypothesis would imply the opposite.