Hauptwerk Virtual Pipe Organ

Thanks, Joe.

In Dr. Pykett's test (essentially the same as as I proposed), note that his speakers had a quoted max. IMD figure of 1%, whereas for the Mackie HR824MK the manufacturer's SMPTE IMD figure is *much* less at 0.035%, so I'd hope/expect you to find much lower levels of true IMD products with your speakers.

Martin,

Unless you take the weekend off for this sort of discussion,
perhaps your time would be of greater value attending to HW V 4.0 .
So far, 90 replies in this discussion/response!

I'm completely sympathetic to the goal of HW4 ASAP; however, this thread has been illuminating on several topics for me. It is a mini-education in the issues of speakers and HW and a useful reference. I've been following it carefully, and know I'll refer back to it from time to time as I consider changing my own set-up.

Ken

toplayer2 wrote:One presumes that you have suggested equally tempered thirds for a reason.

Hi Joe and everyone!

I deliberately recorded my experiments using mono so as to record the mixing effect that we experience ordinarily and I hope that you were able to hear it on my recording.

The effects are more proportionate to the loudness of the sounds. This is not quite apparent on my video on account of automatic volume control on the camera, but on diapasons and louder a matter of feet from the speakers, one hears them. Really my Tuba is at a volume which one might only normally experience in the organ case and probably would sound fine 100 feet away. . .

On a pipe organ one hears them too and anyone with console near to diapasons at the front, if you listen carefully, you'll hear them. All of us who tune instruments will be very familiar with this.

?? However, with speakers, we are forcing the sound of the area of the pipes through a much more concentrated point source with many speaker designs so the sound pressures may be concentrated and therefore we might notice high intensity effects more. ??

The reason for specifying equal temperament thirds is detailed earlier in this thread: thirds should be the 4th and 5th harmonics of a fundamental note. This goes to the root of our perception of music and musical intervals.

So taking an A at 55, the 2nd harmonic will be 110, the 3rd at 165, fourth harmonic at 220 and fifth at 275, sixth at 330 etc. So the difference between the 4th and 5th harmonics is 55, the original fundamental note. The beat frequency supplies this and the ear expects it. When you then play the fundamental note in addition, the chord is is wonderful consonance.

In equal temperament, the C# note is 277 instead of 275. This means that the beat note is 57 instead of 55 and sounds sharp and if played with the 55 creates beating. In fact the next note in equal temperament is 329, and if sounded with 277, the beat frequency is 52. The resultant 52 beats with the true note 3 ish times per second and if the triad is played, the 52 beats with the 57 beat note at around 5 per second and the chord is a mess. In the triad, the fifth, 220 to 329 results in a beat note of 119. If the fundamental note is sounded and it contains rich harmonics, then 119 will beat 9 times a second with 110 adding to the mess and a quarter tone sharp. Very nasty sound.

This is why Martin has asked you to use equal temperament intervals.

Organists relying on the Acoustic Bass 32 stop rely on this beat frequency perception to create a false 32ft sound out of two 16ft pitched pipes a fifth apart. Without the stop, sometimes one simply uses both feet for the two notes to create the same effect and sometimes one sees a Quint 10 ft stop to do the same thing.

The phenonoma is real.

In the unequal temperaments and especially Meantone, perfect thirds and minor thirds and some fifths produce chords which are wonderfully concordant where all the notes are sufficiently relating to the harmonics of the fundamental to all solidify the chord. In contrast other chords may not be noticed as particularly out of tune for instance in Kellner temperament but the intervals in the triad are sufficiently wide for none of the beat notes to coincide with any frequency recognisably close to the root fundamental note. I call such a chord "unrooted" and Chopin deliberately used the contrast between rooted chords representing safety to unrooted chords representing uncertainty and thereby setting a disconcerting atmosphere in his funeral march:
http://www.youtube.com/watch?v=X1PQn_1t3kE and on another piano
http://www.youtube.com/watch?v=iw8FjHvHu30 tuned in the same way.

When one is near the piano, with ones head near the strings, the beat frequencies are clear.

This is the importance of exploring especially the 19th century repertoire in unequal temperament and it has the advantage of producing sounds that are nice and clean when the composer intended and when "unpleasant" artifacts are heard, the composer probably intended them!

The moment one starts to hear the nasty artifacts of equal temperament through speakers is the time to start exploring and enjoying the unequal temperaments!

Best wishes

David P

While all this noticing beats as discordant is very interesting, I think that it is important to remember we are in a real world of different temperatures,humidity weird pipe problems, and on and on. In my experience, the only time an instrument is really in tune, no matter what the temperament, is in the 10 or so minutes that it takes for the tuner to pack up his tools and go home. In the 19th century, the problems of tuning must have been extreme. As the concert hall filled up, the temperature would rise, the instruments would go out of tune, the poor organ suffering the most. Fortunately, at that time, the audience made so much noise waving to friends, chatting, and so forth, that it probably did not matter much. While I am sure that most gentle readers of this post will not believe it, there are people today who deliberately detune their sample sets, to more accurately experience this real world effect!!
James

James wrote:While all this noticing beats as discordant is very interesting, I think that it is important to remember we are in a real world of different temperatures,humidity weird pipe problems, and on and on. In my experience, the only time an instrument is really in tune, no matter what the temperament, is in the 10 or so minutes that it takes for the tuner to pack up his tools and go home. In the 19th century, the problems of tuning must have been extreme. As the concert hall filled up, the temperature would rise, the instruments would go out of tune, the poor organ suffering the most. Fortunately, at that time, the audience made so much noise waving to friends, chatting, and so forth, that it probably did not matter much. While I am sure that most gentle readers of this post will not believe it, there are people today who deliberately detune their sample sets, to more accurately experience this real world effect!!

I am with you on this, James. Alternate tunings is definitely not my crusade, although it could be a good discussion on another thread for those who are passionate about this subject. The only relevance here as far as I am concerned is to use Just Intonation for some tests simply to rule out beats as the cause of the perceived artifacts that one hopes can be mitigated with multiple audio channels, which is on topic for this particular thread.

Joe

James wrote:While all this noticing beats as discordant is very interesting, I think that it is important to remember we are in a real world of different temperatures,humidity weird pipe problems, and on and on. In my experience, the only time an instrument is really in tune, no matter what the temperament, is in the 10 or so minutes that it takes for the tuner to pack up his tools and go home.

Hi!

That's why earlier on I made comment about all those six digit "accuracy" figures earlier on and in the analysis of difference frequencies and beating above, stuck to 3 and less significant digits.

Having said that, other than the difference between reeds and flues, an organ should essentially stay in tune with itself very stably, only being altered by accumulating dust in pipes which of course should not be significant in the short term, and slipping sliders. In fact cone tuned organs are particularly stable in their tuning. Unskilled piano tuners will leave pianos to go out of tune rapidly but a skilled tuner working on a well designed piano, and old Broadwoods and Bechsteins come to mind, can leave an instrument to stay in tune with itself for 3 to 6 months at a time, even years, and I expect a concert piano to withstand Prokofiev's 7th and 8th sonatas without going out of tune.

However, you are right . . . in vertically arranged organs in tall basilicas, where Swell might be above Great and even the third manual might be in the stack, they can suffer from the severe temperature differentials inherent in a building, often a problem around the mediterranean. On my instrument I use a form of octave coupling that gives a slight delay and is slightly detuned. The speaker is mounted upstairs, so it sounds as though part of the instrument is further away . . .

For the avoidance of doubt, I have done a recording of beat notes and their effects on a chamber pipe organ and on a piano. Of course, as the chamber organ is much softer than an instrument speaking to a larger acoustic so the voicing of the larger louder instrument will produce more noticeable effects, but the beat notes are still there and they can be discordant.

http://www.youtube.com/watch?v=XmCgZq6Lmm0 I hope this video is useful.

Violinists hear these notes when playing two strings together and call them "Tartini notes", using them as a tool for perfecting the purity of their intervals.

It doesn't matter what tuning (temperament) an instrument is tuned to, the beat note effect is always present. What varies is whether the resulting tones are to what degree in tune or discordant with other notes sounding together. An unequal temperament contrasts what I call "rooted" and "unrooted" chords.

I continued this further examining Chopin's use of temperament and the beating of the beat notes in an unequal temperament which he relied upon to make the opening of his Funeral March sound like bells.

http://www.youtube.com/watch?v=nPvHq8HvTKg
Adolfo Barabino explains its effect http://www.youtube.com/watch?v=YXzSXWaQGmA
and in his examples you can hear a wa-wa-wa-wa sound in many of his chords on account of the slow beating between resultant notes when there are more than one which nearly coincide. At a distance or at lower sound levels beating between the resultant notes is greater than the notes themselves.

The never ceasing, always the same, discordances in Equal Temperament that we hear here possibly mistaken for IMD (it might be argued that the IMD happens in our ears) are one of the reasons why Unequal Temperaments remove the monotony of the dissonences and why exploration of usable temperaments is a really important to the organ.

Joe - sorry to have taken this tangentially off topic, but it is an inescapable product of examining the beating effect that I hope you'll be able to hear in these videos of real rather than electronic instruments.

Best wishes

David P

Just for the record I hear the same triad sound playing two notes of my real pipe organ, say C and E, at least one octave above middle C. What I hear is the two notes with what I would call a lower almost humming like frequency as the third tone. I hear this with both of the examples Joe did, mixed, separate L and R and my real pipes.

I first noticed this with my 90's MDS Allen and thought it was a product of the digital reproduction. I then went to the church of a good organist friend of mine that has a real pipe organ there, and I heard the same effect.

Again what really bothers me most is are the fast beats due to the thirds etc. not being tuned pure. I still hear this with multiple speakers, but overall the sound is just more at ease sounding, and giving more of an interactive soundstage, more like my real pipes, even with wet samples.

Eric

Eric Sagmuller wrote: I still hear this with multiple speakers, but overall the sound is just more at ease sounding, and giving more of an interactive soundstage, more like my real pipes, even with wet samples.

Hi!

Yes - very good way of putting it, and doing the one-two speaker experiment video last night had that effect on me, of course unapparent through the mono camera recording. At ease . . . more space . . .

Best wishes

David P

Postscript - it was apparent doing the video of the small foot-blown pipe organ, a low pressure soft charming instrument, that whilst the beat-note effects were there, one had to be rather close to the console and the pipes to hear them - and with the "Rompler", as an acquaintance of mine calls the Beast, I had to use the louder stops for the effects to be audible for recording. As one goes away from the instrument, the effects are less objectionable, so the phenonomen is likely to be sound pressure level related. With electronic amplification we tend to reproduce sound at higher levels than we would experience in real life, and in reproducing an organ in a domestic environment we are likely to be closer to speakers than we would be in a large building, thereby experiencing localisation of higher sound pressure levels.

With an inverse square law decay of pressure level related to distance from the source, when we come away from pipes or an electronic speaker, the pressure levels decay from those at which we are conscious of beat notes very rapidly. But in the enclosed volume of a domestic installation, firstly we cannot get distance from the speakers, and secondly the inverse square law is modified as the sound source is not operating in open space and being concentrated into a more enclosed shaft of sound. So we are bound to be more conscious of the beat-notes.

Whilst perhaps some might have thought the piano videos to be irrelevant, it was apparent that in order to hear the beat notes that we are not conscious of as audience and may only be mildly aware of as performers or tuners, I had to put the camera close to the strings and the beat note decayed very quickly. I had to hit the strings very loudly to hear the notes in the first place.

So perhaps if beat notes are troublesome, and troublesome to too great a proportion of organ stops, on an electronic organ perhaps simply turn the volume down!

One last brief but key point, that I forgot to make previously in this thread:

True IMD would apply approximately equally to any music you played through any given speakers (at a given volume), so listening to a good organ/classical music CD through a pair of speakers would be equally subject to the effect.

If speakers had levels of IMD that were really high enough to be noticeable/objectionable, then you probably wouldn't be able to enjoy listening to recorded music through them, so they probably wouldn't be considered acceptably good speakers in the first place.

That's one of the main reasons that I suspect that true IMD probably isn't really so noticeable/objectionable with most good speakers, and that I suspect that what most people think they hear as speaker IMD, and try to minimise with multi-channel audio, is actually (primarily) something different (especially 'subjective tones'/beats).

Hi all,

Upon checking the effect of third beats from my dry sample sets, and
in the home environment, yes they are there,
but when adding reverb I find that the dissonance is not so obvious.

Does anyone hear these derogatory sounds when playing music on the organ, or only during analysis ?

"Music Maestro's Please".....

At least in my case all the beating thirds diminish rapidly as I move away from a piano or organ speaker. I suspect the reason is the additional multi-phase room reflections. That's an interesting point about reverb. Whether from the real acoustic or artificial reverb, the additional phase relationships should mask the beats. I'm sure this is one of the reasons for my strong preference for stereo samples as well. The same effect may explain the practise of turning speakers around to reflect off walls and such. Is it possible that mono sample sets REALLY need multiple speakers and stereo or surround sets beneift a little from extra speakers? Same for dry/wet as has been mentioned previously.

On another topic, I cannot resist the urge to point out that the spec being quoted for the Mackie speakers is not the spec for the speaker, which would be considerably higher, but the spec for the amplifier section of the speaker only. I don't question the quality or well deserved reputation of these speakers, but speakers remain the most non-linear of all the components in the audio chain.

Concerning historic temperments, it's one of the strengths of HW that it becomes so easy to experiment with these temperments. Temperment "fixes" always come with a price and making a set of preferred intervals sound better results in other intervals sounding much worse. So it's a tradeoff that can be seen as more musical, and sometimes it's a tradeoff that happens more by accident such as when tuning using perfect fifths. But always a perfect beatless C to E, as an example, makes some other interval sound worse.

jkinkennon wrote:At least in my case all the beating thirds diminish rapidly as I move away from a piano or organ speaker. I suspect the reason is the additional multi-phase room reflections. That's an interesting point about reverb. Whether from the real acoustic or artificial reverb, the additional phase relationships should mask the beats. I'm sure this is one of the reasons for my strong preference for stereo samples as well. The same effect may explain the practise of turning speakers around to reflect off walls and such. Is it possible that mono sample sets REALLY need multiple speakers and stereo or surround sets beneift a little from extra speakers? Same for dry/wet as has been mentioned previously.

On another topic, I cannot resist the urge to point out that the spec being quoted for the Mackie speakers is not the spec for the speaker, which would be considerably higher, but the spec for the amplifier section of the speaker only. I don't question the quality or well deserved reputation of these speakers, but speakers remain the most non-linear of all the components in the audio chain.

Concerning historic temperments, it's one of the strengths of HW that it becomes so easy to experiment with these temperments. Temperment "fixes" always come with a price and making a set of preferred intervals sound better results in other intervals sounding much worse. So it's a tradeoff that can be seen as more musical, and sometimes it's a tradeoff that happens more by accident such as when tuning using perfect fifths. But always a perfect beatless C to E, as an example, makes some other interval sound worse.

Hello John,

I find myself in complete agreement with your observations. Your point about distance from the speakers is interesting. Presumably, the room reflections take on a larger role and create some comb filtering. An interesting thing (to me) about comb filtering is it's effect looks terrible on a frequency plot, but seems to make instrument ensembles more listenable. Take for example an orchestral string section. Few sounds are more glorious. It seems likely that at least some of the reasons for the richness include comb filtering, the multiple slight tuning errors that are unavoidable, and the non-synchronicity of the vibrato/tremulant. (Oh, oh, I may have just opened another can of worms ).

Stereo samples seem more, hum, not sure what term to use, "tangible"? Unfortunately, they are a luxury item for sample set users who want multiple audio channels because one needs twice as many.

Thanks for your comments.

Joe

I think I have figured out a simple mathematical explanation for the mystery tones that I perceive, although this is very unlikely to be a new discovery of any kind... I was a music major, not a math major.

When I play the C above middle C (523.251 Hz ) along with the major third E (659.3 Hz), I hear a tone that is roughly the G above middle C. The difference between the frequencies of C and E is 136.0 Hz. If this is subtracted from C's frequency, the result is 387.25 Hz, pretty close to G (392.0).

When I play the C above middle C (523.3 Hz ) along with the minor third Eb (622.25 Hz), I hear a tone that is roughly the G# above middle C. The difference between the frequencies of C and Eb is 99.0 Hz. If this is subtracted from C's frequency, the result is 424.25 Hz, pretty close to G# (415.3 Hz).

Roughly the same with Just tuned notes with no beats. Coincidence perhaps, but it seems that these could well be the phantom pitches (387 and 415 Hz) that I hear.

FWIW.

Joe

toplayer2 wrote:When I play the C above middle C (523.251 Hz ) along with the major third E (659.3 Hz), I hear a tone that is roughly the G above middle C. The difference between the frequencies of C and E is 136.0 Hz. If this is subtracted from C's frequency, the result is 387.25 Hz, pretty close to G (392.0).

Dear Joe

I'm wondering what stops you are using to hear this? If the tones that are producing the notes you're observing are flutes with fairly strong sine fundamentals and no upper harmonics, then your explanation is interesting but appears to be rather a rare phenonomen. The fact that your frequencies you predict are near to what you hear but not exact suggest that either that your observation is loose or that a different process is going on. 424 is distinguishable from 415 etc.

The notes are really generated by an interference process:
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | with
|..|. |..|. |..|. |..|. |..|. |..|. |..|. |..|. |..|. |..|. |..|. |..|. |..|. |..|. |..|. |..|. results in maxima which might not show correctly on different screens with different font widths
|.. . .|.. ..|.. . .|.. ..|.. . .|.. ..|.. . .|.. ..|.. . .|.. ..|.. . .|.. ..|.. . .|.. ..|.. . .|.. ..|.. . .|.. ..| and these maxima aren't distortions, simply where your ear is picking up the rhythm of constructive interference of the addition of the two periodic maximae of the two frequencies and heard therefore as an additional note.

This evening I was chatting about these phenonomae to a violinist who says that in the course of normal playing he uses these spurious notes to check and adjust his tuning to get pure intervals rather than the equal tempered intervals to which musicians are now entrained. (http://www.guardian.co.uk/books/2008/nov/22/how-equal-temperamnt-ruined-harmony is a good introduction) He was describing invervals of a tenth, clearly probably produced between interferences at the second harmonic level.

A couple of posts back or in one of my videos I tried to explain the construction of sounds in terms of the harmonic series and I illustrated the example with a low A of 55 Hz. Playing Treble C and the E above that, you should be getting Tenor C as the beat note. But if the stop that you are using has strong 2nd and 4th harmonics, you'll get those producing beat notes in higher octaves and filling in the missing note of the triads that you are experiencing. (Anyone with a Hammond can test this by setting up a sound with high values of 2nd and or 4th Harmonics - a drawbar registration of perhaps (0)(0)2508000)

So . . . I have half a feeling that you're hearing beats of the harmonics rather than the fundamental notes and one often has experience of this when tuning pianos.

Finally, these phenonomae are only heard at high volume levels - the violinist hears the tones privately - thank goodness we in the audience aren't troubled by his tuning cues. Similarly turning the speakers around to face walls or corners might be enough to reduce the volume below the critical level, or, confuse the pattern of standing waves inherent in a smaller space with walls noted in one post as a comb effect, so that the pattern is broken up to smooth over the points of maximae which were above the sound pressure level necessary to trigger perception of the interference frequencies.

One of my sons borrowed my sound pressure meter - when this turns up I will measure the critical point at which the beat notes start to be observed. I'd guess somewhere in excess of 80-90dB . . . .

My violinist friend said that he had been told that the frequencies were on account of non-linearity of the ear. Certainly this is possible: http://www.merck.com/mmhe/sec19/ch217/ch217b.html

The stapedius muscle is attached to the stirrup and oval window; it contracts in response to a loud noise, making the chain of ossicles more rigid so that less sound is transmitted. This response, called the acoustic reflex, helps protect the delicate inner ear from sound damage.

So really the answer is to turn the volume down to below the point at which the acoustic reflex kicks in . . . !

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Raised although tangential to topic, it's commonly thought that temperaments were used to create greater harmonicity. Yes and no. . . . They also created discord which was intended and special effects. In my last video relating to the piano I showed how Chopin used these in the Funeral March and the following piece by Couperin is really interesting. In equal tuning, it's bland although a discord or two shows up. It's easy to play so worth playing, even if classical music isn't your thing - many people here are into cinema instruments, I know - but it's worth giving a go.

Try it in Meantone if you can, and use a basic Bourdon at 8ft and 16ft pitches, soft Diapasons at 8ft and 16ft (leave them out if you only have hard Diaps), a principal 4ft a flutish 2ft and as many high sparkling mixtures as your organ has, all coupled together. (St Max - Resonnance Petit Fourniture IV, Cymbale IV Grande Fourniture II Prestant 4' Montre 8' and 16' and Bourdons to thicken it if you want, Positif Cymbale III Doublette 2 Montre 8 Prestant 4' and Fourniture III, both Positif and Resonnance coupled to Grand Orgue http://www.youtube.com/watch?v=yFTStijofmw for actual sound). There's certainly enough in here to generate plenty of beat frequencies - so when they all accord, the sound is BRILLIANT and when they discord you wince, and are intended to, and it's hideously unforgiving of mistakes: http://www.organmatters.co.uk/couperinkyrie.jpg



As you go through this in Meantone, some chords are soothingly pure and express beautiful love - perhaps linger a shade on those chords which are wonderfully harmonious, and then in the 4th stave, second bar, first chord is excruciating crisis - Couperin goes from harmony to crisis and then resolves it. When playing this in Meantone, beyond playing the notes, I prefer to linger on those chords which have more meaning and don't simply rush through them. http://www.youtube.com/watch?v=oj6mqEiSwr8 is a nice interpretation.

Your beat notes will do interesting things in this and I hope that you might enjoy playing and exploring it.

Best wishes

David P

Another go at showing the interference process if the example above was confounded by the font width:
|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|added to
|.......|...... |.......|......|.......|...... |.......|......|.......|...... |.......|......|.......|...... |
produces
|...............|...............|...............|..............|...............|...............|...............|
an audible note heard an octave below the lower of the upper notes