Is there a tendency to sing flat when singing out of tune?
I would not be surprised to hear that singers are more commonly corrected in that direction, but it's certainly not impossible to be "a little sharp." A few reasons this might be perceived:
There is generally a correlation between tension and higher pitch, be it tension in the diaphragm or larynx or whatever; when people sing higher in pitch, they generally need to exert more effort for whatever reason. Usually with proper vocal production this goes away, but until then, when singers get tired or aren't giving as much effort as they need to, the pitch can sag.
The ear can much more easily detect errors of flatness of pitch. If someone is singing or playing a little sharp, it usually manifests itself as an aural brightness rather than a clashing out-of-tune note. High school trumpet players are notorious for tuning sharp without realizing it, because it makes their sound stand out a little from the rest of the section without sounding "bad" compared to the rest of the ensemble.
So, what everyone has said about tendencies of singers to go flat still applies, but keep in mind that a lot of this behavior you notice has to do with the fact that sharp pitches usually sound like a change in tone, making them harder to identify than notes that are flat.
In contrast with many here, I do think that an untrained singer will tend to go flat more often than sharp. This is usually a technique thing; a collapsed soft palate and improper posture will decrease breath support and drop the pitch of the note. It also closes off the airways which will reduce the natural overtones of the human voice, which makes the note sound duller and lower. Lastly, a lot of difficulties are in high-note singing; if a note is at the extreme high end of your range, it requires even more attention to technique to get there and stay there, and if you don't keep the support, the note will trend flat.
However, improper technique can also cause notes to go sharp. A tight throat and "pushing" will generally tighten the vocal chords and force or pinch a note sharp. These are just as common in novice singers, and in combination with the previous you can get a shotgun effect of roughly equal sharp/flat, But, back to my original point, I think that in most group singing, you'll get more notes trending flat than sharp, unless the piece is fast and/or difficult, in which case your singers will tense up and pinch sharp.
I don't believe that people are inherently more likely to be a little flat than a little sharp. There are various situations in which untrained singers may be more likely to go flat or sharp, however, and I've put a few possibilities below as a taster. If you are consistently being told that you are little flat, you may want to examine in particular whether you are keeping your soft palette up (to allow for a more resonant, less dulled sound).
When singing a descending pattern, untrained singers are more likely not to support the sound as they descend and therefore the sound may drift flat.
When singing an ascending pattern, or notes that the singer finds high, untrained singers may be more likely to push the sound from the throat and therefore sound sharp.
If when singing a long note, the singer presses down on the sound rather than letting it spin and be energised, then it is likely to head flat.
When repeating a note a number of times, if the singer does not re-energise the note each time, it is likely to drift flat.
If the singer often practises at home with an old, slightly flat piano, then they may have a tendency to sing flat (it's what the muscles become used to...)
If in general the singer uses darker vowels, or a lowered soft palette (so they are not getting much resonance), they may sound flatter.
And I am sure there are lots more...
I have some experience in this field.* This turns out to be an interesting answer. Listening to good and bad singers in recording studios for years, and singing in choirs as a young man, I'll concur that singers do, almost universally, have a consistent tendency to sing slightly flat, and not sharp, particularly when they cannot hear themselves very well. (Personally, I've never actually heard a singer consistently sing a bit sharp, but I'm sure it happens.) As someone mentioned, there is an effective low-pass filter occurring with a -3dB point usually around 1kHz-2kHz. It's created by the "shadow" of everything in the singer's neck and head between the larynx and ears. This dulls the more recognizable frequency range for human vocals, but roughly equally in both ears. It has no relevant effect on perceived pitch, per se, and it's not the primary factor responsible for the effect in question.
The reason that the tendency to sing slightly flat is often observed is actually due to the higher speed of sound traveling through the denser material of the bones in the singer's own spine, skull, jaw and inner ear, from larynx to ears. For some people, when they hear themselves singing, inside their own head, this faster conduction of sound will actually create a perceived pitch-shift upward. Naturally, these singers will then try to compensate by flattening a bit, the note being sung. The singer will then hear the result as being true to pitch, possibly unaware that the rest of the world is cringing! :)
This effect seems to happen for a pretty consistent number of people in the general population (I'm sure somebody has real numbers on this), and clearly not all people seem to be affected by it. I have found that for many of the affected singers, the problem is nearly hopeless to correct. I have also known of a few people who overcame the problem with steady practice while using headphones and listening to themselves in the studio mix.
- (fwiw, I'm an electrical engineer, specializing in pro audio, audio digital signal processing, good ol' assembly language DSP chip programming (yuck), the application of head-related transfer functions for audio source placement in 3D space by only using two front loudspeakers in a normal stereo setup, et al. Sadly, though, in most ways, all audio applications had been "all figured out" by the end of the 90's. Very few frontiers were left. The industry shrank.)
I've heard singers (especially choirs) go flat more than I've heard them go sharp.
There's a standard piece of advice that a lot of choir directors give: "Make your ascending intervals big and your descending intervals small." Obviously, this isn't meant to be taken literally (or else everyone would go sharp all the time); it's meant to compensate for a tendency to go flat.
"Is there a tendency to sing flat when singing out of tune?"
My answer based on 22 years of experience with choirs, different singers in bands that it is much more common that people are slightly flat than sharp.
I have only experienced one singer who was noticable sharp. That was a female soprano singing classic pop and rock songs in keys that made most of her notes sit in the chest and lower mid register. She was seldom ever singing notes at her second passaggio. All other leadsingers have struggled to stay on pitch and would be slightly flat at hte end of a gig or a 2 hour rehersal.
I have done a lot of research,, looking at old interviews dealing with pitch and asking professional opera singers about this issue. One singer (tenor Joseph Calleleja) has said that A4=438Hz is the upper limit for heavy dramatic voices. Opera singers have at a conference in Italy around 1990 demonstrated the "early" register shift phenemonen of A4=440-443Hz in relation to Verdis suugested A4=432Hz that limits early register shift at the passaggios. Former European pitch (french diapason normal) or continential pitch at A4=435Hz is much better suited for a large majority of voices.. Less vocal fatigue and less flat singing.
It might just be that the phrase "a little flat" is much better known. My mother, who has no musical training, says this often but never anything else. She has normally picked up some imperfection in the performance but not necessarily flatness. For example, she once described my son's piano playing as flat. If he was playing flat then it would the tuner's fault rather than his. In fact, he had played a couple of wrong notes.
Terminology is often misused by those unfamiliar with the field. As a mathematician, I often cringe at the misuse of "exponential". Similarly, physicists have to tolerate lots of misuse of "quantum". Musicians need to tolerate misuse of "flat" in the same way.
There's a technical reason why even good singers have a natural tendency to go slightly flat; it's a consequence of how common patterns in music interact with our tuning systems.
Some background: The intervals we find most consonant turn out to have simple whole-number frequency ratios; for example, in a pure fifth, the upper note has a frequency of 3/2 that of the lower.
So if you start on a low note, and go up in pure fifths, by the time you return to your starting note, you'll be slightly sharp. That's because (3/2)¹² ≈ 129.746, slightly more than 7 octaves which is 2⁷ = 128. In practice, of course, you'd change octaves, going alternately up a fifth and down a fourth, but the result is the same — you'd end up nearly a quarter of a semitone sharp. (For the technically-minded, that difference is known as the Pythagorean comma.)
Similarly, if you go in pure intervals around the circle of fifths the other way, you end up flat by the same amount.
As it happens, harmonies that change by a fifth (or fourth) are very common in Western music — and falling fifths are more common than rising ones. (I can't find an explicit reference for this, but most of the examples on the Wikipedia circle of fifths page go that way, and I'm sure you can think of many more. I've heard this type of harmony described as ‘centrifugal’, though that term doesn't appear to be standard.)
And since singers and choirs naturally tend towards pure intervals, it follows that on average they will go flat.
Of course, good singers will be aware of this and try to correct for it. (One way is to aim for equal-tempered intervals instead of pure ones, if you can tell the difference.) But it's not always easy, especially when there are other things to concentrate on — and there always are! — and so the tendency remains.
(This isn't the only reason why people tend to go flat; other answers have covered some physiological causes, which I'm sure play a big rôle too. But even if those are countered, the mathematics is against us!)
If you're singing with fixed-pitch instruments/backing (whether equal-tempered or not), that gives you a fixed reference and makes it easier to combat the tendency. But it also makes the fall in pitch more obvious…
My theory would be that if you can't master the song your vochal chords and all that won't achieve the desired note and be a little down, most likely you wont overshoot the note because 1: Your body will try to hit the note on the spot; 2: Your body may not handle the strain very well of a higher note.
So if you aren't in tone, you'll probably be on the one that pushes your body the less...