How would a trumpet sound without its harmonics past 20khz

I know I am getting on a similar topic that I have touched on before,
but would the trumpet sound different (timbre of a particular note)

I'm talking in real life and not through a HI-FI stereo.
This is an area I am interested to know but has continiuosly puzzled me for a long time.

Now relating it to HI-FI stereo's if it's proven that there is a difference in timbre from using higher banwidths, how could this be since humans can only hear sine wavs up to 20khz max.

I'm thinking maybe the different transients produced by the higher banwidths are producing the changes in timbre due to the difference in sound pressure levels in our ears over a given time.
My creative (and sometimes BS mind) tells me the special hairs that convert air vibration into electrical signal in our ears are resonating out (like a shock wave) by these transients that are produced and therefore thats what makes the timbre change. Anyway thats my seriously BS explanation.  

Anyone know?

[if]

If you're over the age of 25, no different at all, because presbycusis will have disabled the very end of your cochlea. If you are younger than this, then it is possible to detect frequencies slightly above 20k, but they will primarily be detected by bone conduction, because these frequencies are extremely directional, absorbed by just about everything they come into contact with, and don't travel at all well, if at all, up your ear canal.

You said

Now relating it to HI-FI stereo's if it's proven that there is a difference in timbre from using higher banwidths, how could this be since humans can only hear sine wavs up to 20khz max.

The short and sweet answer is that it's not proven. What is proven are the effects of presbycusis, and the reasons are well-known - it's an unavoidable part of the ageing process. What is also proven is that bone conduction effects run at about -55dB ref 'normal' hearing at 1kHz. It works because the cochlea is encased in bone, and can be excited directly when the bone vibrates, but there is no mechanism at any age for hearing significantly above 20kHz, due to the nature of the basilar membrane. It responds to high frequencies where it is narrow and thin (at its base) which happens to be at the front of it, and this is the area most likely to be damaged, just by normal living.

When you also add in the effects of masking, and the properties of the ear-brain relationship when it comes to critical bands (which are documented elsewhere), the chances of frequencies above 20kHz making any difference to the percieved timbre are effectively zero.

(spelling corrected - content unaltered)

I think the real issue has not been addressed, namely that what is happening outside the audible band DOES affect what we can hear.

Trumpet with harmon mute is a good example. LOTS of stuff happening above 20kHz that interacts with itself--remember Helmholtz principle. If you have a 25kHz tone and a 35kHz tone you will get a 10kHz tone-- and a 60kHz tone.

This same idea is found in many pipe organs. Due to the tremendous expense of 32-ft pedal pipes many modest-sized organs have "32-ft resultants".

32-ft low C sounds 16Hz. If you play a commonly-found 16-ft C an octave higher (32Hz) and a G above it (48Hz) you will get a 16Hz resultant tone.

Back to the issue-- if your system is capable of capturing the out-of-band material and reproducing it, then you will also hear what results in the audible band. Of course it will give you the audible info as well-- you must be the judge of how convincing or affordable it is.

In other words, it is what we cannot hear that is sometimes important.

Rich

If the trumpet playing results in a 10kHz signal via interaction between signals the trumpet directly produces at 25kHz and 35kHz, that 10kHz signal should be present in the recording, even though nothing above 20kHz could be captured, no? If so, the recording cutoff still makes no difference to what will be heard, does it?

I think the real issue has not been addressed, namely that what is happening outside the audible band DOES affect what we can hear.

Trumpet with harmon mute is a good example. LOTS of stuff happening above 20kHz that interacts with itself--remember Helmholtz principle. If you have a 25kHz tone and a 35kHz tone you will get a 10kHz tone-- and a 60kHz tone.

If you knew anything at all about the acoustics of waveguides, the diffraction effects involved, and the directional properties of HF sound, you would know that the non-linearities involved in mixing these sounds (where Helmholtz comes into this isn't clear) give rise to harmonics within the audible spectrum that are able to be perceived by the human ear. The resultant sound output from a trumpet (it doesn't need the mute) above 20kHz is a) hightly directional b) subject to air attenuation, which is quite considerable at the frequencies and distances involved, and c) effectively diffracted by interacting paths near the instrument anyway. You can't just add two frequencies together to get sum and difference components if they are radiating even in fractionally different directions - and the HF radiation patterns around a trumpet are by no means consistent. It is only those non-linear mixing effects within the instrument itself that are radiated within the human audible range that affect the perceived timbre. That there are other frequencies present during the generation of the sound is not in dispute; their effect on the timbre will only be present within the human hearing range, though.

This same idea is found in many pipe organs. Due to the tremendous expense of 32-ft pedal pipes many modest-sized organs have "32-ft resultants".

32-ft low C sounds 16Hz. If you play a commonly-found 16-ft C an octave higher (32Hz) and a G above it (48Hz) you will get a 16Hz resultant tone.

The stop that does this is often called a 'quint', and is voiced at 10 2/3ft if it is a pedal stop, and 5 1/3ft if the same effect is desired between 4 and 8ft stops. When I was building organs, I voiced quite a few of them. They only work effectively if you can combine the resultant airstreams, and then are only really effective at a distance, and not very loud.

Back to the issue-- if your system is capable of capturing the out-of-band material and reproducing it, then you will also hear what results in the audible band.

Might be true if it was possible, but in the real world, it isn't, for reasons I already outlined. These frequencies really are extremely directional - if they weren't, how do you think that bats would use them as radar? They get absorbed by just about everything in furnished buildings - which is why bats prefer caves or hard roofs to live in. At ultrasonic frequencies in relatively dry buildings, we are talking about air absorption values of several db/ft, and also about relatively low-level signals in the first place. When you combine this with the effects of presbycusis, it's a total non-starter.

In other words, it is what we cannot hear that is sometimes important.

Yes, inside an instrument as a part of the timbre generation process within the human hearing range, but not the way you are suggesting, I'm afraid.

If the trumpet playing results in a 10kHz signal via interaction between signals the trumpet directly produces at 25kHz and 35kHz, that 10kHz signal should be present in the recording, even though nothing above 20kHz could be captured, no? If so, the recording cutoff still makes no difference to what will be heard, does it?

Quite correct. That 10kHz signal that is radiated within the human hearing range will make a difference to the perceived and recorded timbre, but the originating components definitely won't.

If that isn't convincing enough, remember that timbre is determined by the relative amplitude of component harmonic pitches, and the human ear cannot differentiate pitches above 15kHz.  This means the human ear can't use >15kHz. components to determine timbre, even though they can detect their presence/absence/magnitude in discrete cases.  

Time based wave-shape processing is another possibility, the way an oscilloscope shows a sharper waveform when you pile on the harmonics, but studies show that this is not an ear function but a brain function, that is limited to < about 40 Hz.  
Above 40Hz. the brain processes information far too slow to image a waveform or it's period.  This pretty much rules out the brain imaging high frequency waveforms.

The most likely situation where frequency components above 20kHz. can be detected is with continuous simple pure waveforms during rapid a/b testing where they either exist or don't exist.  Even then timbre is not part of the equation, only a subtle change in highest frequency level.  This is great for showroom demonstratons, but has little to do with music or timbre.

A trumpet is a very complex speaker in it's own right, and so will never sound like a conical speaker in a box or vise-versa.  If you're looking for a way to improve audio reproduction this is a much more important factor than reproducing the ultra-sonic harmonics.

Point taken on HF absorption (although the original post did not mention distance) and several other statements.

The stop that does this is often called a 'quint', and is voiced at 10 2/3ft if it is a pedal stop, and 5 1/3ft if the same effect is desired between 4 and 8ft stops. When I was building organs, I voiced quite a few of them. They only work effectively if you can combine the resultant airstreams, and then are only really effective at a distance, and not very loud.

To paraphrase your less-than-gracious response, if YOU knew anything about organ building on this side of the Atlantic then you'd know that resultants are much more prevalent than separate quint ranks. The size and character of the room has at least as much to do with the success of such an attempt as anything else.

You can't just add two frequencies together to get sum and difference components if they are radiating even in fractionally different directions

Sit in any orchestra pit during the overture to the Nutcracker and you will hear the impossible-- resultant tones from the horn pitches towards the end on the overture. Horns are not very directional and are almost always heard "reflected" and therefore rather diffused.

If it really comes to the crunch, I can point anybody that's interested in the direction of the psychoacoustic research that's been done into the perception of timbre. On finding of interest is that at no stage is any harmonic that is more than the seventh of the fundamental involved in timbre perception at all, and when it comes down to higher notes, this figure is reduced to the fifth harmonic. And if you look at the spectra for a trumpet playing C4, you will find that all of the significant timbre-forming harmonics actually come in below 4kHz! Basically, all the rest of the harmonic series that a trumpet produces does not affect it's timbre - in fact it is rather moot as to what, if any, affect this has on anything, except on bats and dogs who happen to be in the direct line of fire.

I'm not quite sure where this myth about ultrasonic output from some instruments affecting the human perception of their timbre started, although it has been around for a while. But even if you ignore the psychoacoustic research, and human presbycusis, the simple acoustics of it don't add up. Or all the post-grad teaching and research into this is wrong... (I think not!)

To paraphrase your less-than-gracious response, if YOU knew anything about organ building on this side of the Atlantic then you'd know that resultants are much more prevalent than separate quint ranks. The size and character of the room has at least as much to do with the success of such an attempt as anything else.

I can show you plenty of examples of both methods in the US. Utilising the same rank is the cheap-skate method of doing it - the advantage of using a separate rank is that it can be combined with more than one stop to give more than one effect, and I've seen plenty of US organ specs that do this. Also there are some that combine both methods, like this one. Strictly speaking, the use of the 'fifth above' method on the same rank will put the resultant out of tune, but since the ear has trouble recognising pitch at these frequencies, it's not normally a problem.

You can't just add two frequencies together to get sum and difference components if they are radiating even in fractionally different directions

Sit in any orchestra pit during the overture to the Nutcracker and you will hear the impossible-- resultant tones from the horn pitches towards the end on the overture. Horns are not very directional and are almost always heard "reflected" and therefore rather diffused.

Since you've quoted me out of context (this was referring to HF, not lower frequencies), it's hardly worth replying to. But if it was truly impossible to hear this, then of course I wouldn't, would I? But since the frequencies involved are actually quite low, then it's not impossible at all.

This thread seems to have much in common with the recent discussion we had on the similar thread 'what are harmonics' - and the same queries remain unresolved!

It would seem that the pure  physics of sound become very muddied whenever pyschoacoustics become involved.

At a simple level -- domestic room, normal conversation - Uncle Fred's voice is recorded, and then played back to him - his reaction 'doesn't sound like me' - everyone else 'oh yes it does' ---- so far so good, easy to explain.

But the one characteristic that has changed is the perceived ' reverb' of the room, in the recorded version. In the live 'case' you don't notice any 'echo' effect when listening to Uncle Fred - but the recorded version has a distinct 'reverb' content, even though the microphone was at the same distance from him as you were. - clearly the brain has filtered out this content in the live situation, but cannot when listening to the recording - the physics have remained largely the same ( dependent on the mic quality!) - so the perceived change must be pyshcoacoustic  -

And from that simple example, the minefield in all live perceptions of sound are correspondingly more complex.

If it really comes to the crunch, I can point anybody that's interested in the direction of the psychoacoustic research that's been done into the perception of timbre. On finding of interest is that at no stage is any harmonic that is more than the seventh of the fundamental involved in timbre perception at all, and when it comes down to higher notes, this figure is reduced to the fifth harmonic.

I would like to know under what test conditions this is the case.

From my experience the first 30 harmonics at least are involved in perception of timbre for an 80 Hz. fundimental waveform.  I will have to repeat this experiment again, but i tested my own perception of the removal of one harmonic at a time from a signal, and the difference was easy to detect, all the way up to and including the 30th harmonic.  I didn't bother testing higher harmonics.  I could detect removal or inclusion of the top harmonic almost as easy as a middle harmonic.  
[With an 80Hz. fundimental this means i only included harmonics up to 2.5kHz.]

This may be because i removed the harmonic completely from the set in each case, and this generally never happens in natural instruments.  Usually all the harmonics are present only in different levels.
I was also testing a/b with continuous tones, presumably making it much easier to spot a missing harmonic than with a dynamic waveform like from an instrument.  The choice of a rather low fundimental also helped.

I really shoud repeat this test since this is all from memory of an experiment done over a year ago.  If/when i do i would be happy to post the session and files for others to take the same test.

Steve what about Transients?

I read somewhere on the internet that it plays a role in timbre perception?
Is this true? And if it does, Doesn't a higher bandwidth allow for more solid transients?

While you are waiting for a response, you might read this:

http://audiomastersforum.net/synforum/viewtopic.php?t=12976

If you click on "Forum Archives" at the top of this page, and then search on "transients", you will get quite a few hits.

THanks for the link bonnder.
It brings back memory's.  

Getting off the topic temporarily
Is it true to say that LP's produced more pleasant transients (for me at least)due to there extended frequency response, or did post 78 have a valid point then (phase distortion caused it)  

By the way does phase error and phase distortion mean the same thing?
Anyway back to the topic.