Key Changes

04/02/10 - John Muskett 

I have listened on Spotify,  to lots of versions of Big Butter and Egg Man the un-favourite tune of Peter de Bont who made a number of worthwhile comments in Doom and Gloom a while back. Perhaps there is no such thing as a bad tune (though I have some particular bogey tunes which I'm NOT going to reveal), only poor or tired versions. Versions I enjoyed include those by Armstrong, Spanier, Humph, Braff (with Ralph Sutton) and Wynton Marsalis (with excellent bass work by Reginald Veal and father Ellis on piano). For pedants, keys vary between (apparently) G for original Armstrong and Spanier, C for later Armstrong, F for Marsalis and many British bands, and Ab for Humph and most American outfits. There are also differences of harmonic interpretation between bars 13 and 16. I guess you have to agree before playing or listen carefully and be prepared to defer to someone else.


05/02/10 -  Tony West 

Just looking at a post by John Muskett re the various keys that "Big Butter and Egg Man" has been played in reminds me of a story told to me by the late John Pickering. John was a trombone player, and attended every Australian Jazz Convention since the early 50s. In those days the AJC organisers occasionally had guest international musicians attending, and in my days we had The Turk Murphy Band at a Adelaide Convention, and both Campbell Burnap and Beryl Bryden at the 50th AJC in Melbourne.  Anyhow, "Picko's" story related to an AJC when the international guest was Alton Purnell. In my experience some of the New Orleans fundamentalists always play "Tishomingo Blues" in concert G. "Picko" was playing trombone with Alton in one of the sessions, "Tishomingo Blues" was called, and the key of G was nominated. Alton asked why the hell they wanted to play it in G. Their response was "You recorded it in G with George Lewis". "No!  We always played it in F," responded Alton, "but it would not quite fit on the 78, so the engineers speeded it up and it came out in G.


06/02/10 - Richard Knock

Tony West's anecdote re Alton Purnell rings a bell. I've known a number of bands play in obscure keys due to old or adjusted recordings, and insist that that is the 'right' key. I suppose the truly correct key is the one the composer wrote it in; but it's not unusual for a different key to be chosen to suit a vocalist. I can remember us having to learn 'Misty' in D concert because the range was just out of reach of a singer. Not the most convenient key for a wind player, but we did it. I would like to know why the key of C always seems to brighten a tune though. Do others agree? Is there a scientific answer?-  


08/02/10 - John Muskett

I didn't intend to start a discussion on keys, as my notes on Big Butter And Egg Man were an aside to comments on my loss and recovery of Spotify. For any individual tune the key, in jazz, is surely immaterial. Many singers (Crosby, Sinatra and almost all women) have frequently diverged from the written key, and many instrumentalists have chosen their own key - Coleman Hawkins raised Body And Soul a semitone from the original C (with middle eight in Db and B). It is often a moot point as what might be the "correct" key: indicators are the composer's own key (if he/she recorded it) and the first recorded version (subject to pitch distortion on early waxings). Sheet music is by no means always reliable, often produced by one of the publisher's hack arrangers (did Irving Berlin really write all those flattened fifth chords in a number of his songs?); several of Duke Ellington's best known tunes (Mood Indigo, It Don't Mean A Thing, Solitude, Prelude To A Kiss, Don't Get Around..., Do Nothing Till...) are published in the "wrong" key.

What may be of importance is that musicians should be comfortable in a variety of keys and able to fit in with someone else's preference. It also may be boring for those with short attention spans to be anchored in very few keys. The popular key of Bb, common in multi-themed traditional tunes, is comparatively rare in sheet music of standards; I would guess that at least fifty percent of tunes played in Bb by local bands were published in C.

I would favour G for Tishomingo Blues - a trawl of Spotify versions reveals that almost all American bands used it, and the sheet music supports this, although its title spelled as Tishamingo Blues casts doubt on the veracity of the arrangement! Ellington's 
1928 version appears to have the first chorus in F before going to G. Not only did Bunk Johnson and George Lewis buck the trend by playing it in F, but their chords in bars 12 and 13 differed from those in most versions, both variations copied by many British trad bands, though not Chris Barber's.

Although most tunes are published in keys acceptable to the average male voice, a few are on the high side (Deep Purple in F, Exactly Like You in C and It's Only A Paper Moon in G). It seems reasonable to lower these if they are being sung; if purely instrumental the original key shouldn't present an obstacle. Most of Louis Armstrong's recordings appear to be in the original keys, changing during the last chorus to enable him to finish on a suitably high note. I find that I am largely in agreement with the keys in the French chord books (Les Grilles) and I notice that American bands tend to favour flatter keys (Indiana in Ab).

Best wishes from key fetishist (a title bestowed by the late John Clack).


08/02/10 - Allan Wilcox

Richard Knock wonders whether there's a scientific reason why the key of C always seems to brighten a tune. As a teenager in the 1960s, I became friendly with a professional guitarist called Arthur Waples, who told me that each key had a different character and then proceeded to identify them all. I'm afraid I took it all with a pinch of salt at that time and didn't commit them to memory, but one I do remember is F, which he claimed was the 'drunken' key, going on to explain, quite seriously, that this was the reason why jazz musicians played in that key most of the time!  


08/02/10 - Tony West

I agree with Richard Knock.  We look forward to tunes in C, "One sweet letter from you" being a favourite. "Elephant Stomp" which starts in C then brilliantly modulates to Ab (another great key) is a real gem.


09/02/10 - George Gilmour

Hi Fred, I read with interest the article on ''key changes'', in particular the story that was attributed to Alton Purnell and ''Tishamingo Blues'' not being recorded in concert G. George Lewis's phrasing most definitely confirms that G was the key of choice that day. 

End of story. 


10/02/10 - David Fox

By observation it seems to me that most of the jazz tunes are played in Bb concert. This puts the Bb instruments (eg trumpet, Clari , Tenor sax ) in the key of C , the easiest one in which to play .Maybe this is why Bb concert is the most popular key. I believe most musicians started life learning in C because there are no sharps or flats. The trumpet player leads a jazz band and he calls the key. I expect non reader Jazz musicians will find a reason to disagree with this viewpoint.  


10/02/10 - John Muskett

Tony, Richard and Allan have postulated the brightness of C. Before the general adoption of Equal Temperament tuning for western music, there was Just (or Pure) Intonation where the determined pitch of notes was perfect for one key (often C) and progressively out of tune for keys with increasing sharps and flats in their signatures. The pitching was based on the harmonic series, but, for example, the interval doh to ray was greater than ray to me. Ray to me in C would become doh to ray in D, but would then be too narrow. Another phenomenon was that if you went right round the cycle of fifths you finished up sharp on the starting note seven octaves higher (the Pythagorean Comma). And stacking up major thirds leaves you ultimately flat. So the pitch values were adjusted (Pythagorean Temperament, Meantone Temperament, several "Well" Temperaments) to make the scale of C slightly less in tune and other scales less out of tune. Only in some of the Well Temperaments did keys with 4, 5 or 6 sharps or flats become more or less acceptable. However each key had its own harmonic characteristics: "joyful", "innocent", "sad" for example. Experts are divided as to whether J. S. Bach intended his Well Tempered Clavier to employ Equal Temperament (where all keys are equally out of tune) to show the equality of keys or in a Well Temperament to display individual key colour.

To me therefore it seems unlikely that C has any intrinsic extra brightness in modern Equal Temperament. Perhaps tunes that our ears have become accustomed to in Bb sound brighter when raised to the published key of C. Here's an experiment: play a tune in C and then in Db - does it sound brighter? On a piano Db always sounds livelier to me than C, though some of this may be attributed to black note strings being less dull than those of the more frequently used white notes.

One Sweet Letter is a favourite of mine; there are several 1927 versions on Red Hot Jazz: I particularly like one by The Indiana Five - beautiful piano work. It may be composer Harry Warren's first hit. Incidentally this version (and the sheet music) are both in Eb - whether super-bright or not I couldn't possibly say! These early versions follow the written chord progression, the subtleties of which seem to have eluded almost all the latter-day recordists I've found on Spotify (Chris Barber included alas, although there are harmonically faithful offerings by Benny Goodman and Michael Holliday). What went wrong - oops - different? 

On The Great Tishomingo Mystery ("... data, Watson, I must have more data...") I have heard a George Lewis recording coming out in F. Many of his phrases sound very like those on Bunk's G recording: I wonder whether George's offering was originally in G and the recording engineer SLOWED IT DOWN TO F? ("These are deep waters, Watson.")


22/02/10 - Richard Knock

A little more on the subject - following John Muskett's excellent little treatise on the tempered scale, when I was a bored seafarer on long voyages I amused myself investigating the mathematics of tempered scales - and I remember realising that if you multiply the frequency of each note by 1.059463094...... you end up doubling it after 12 intervals achieving a perfect octave, and each frequency is a tempered semitone, meaning that you accomplish the feat of equal spacing giving you scales that can start on any note. Keyboard instruments that are required to be played in all the keys have to be tuned in that way and we have all got used to hearing the compromise and don't really notice it. 

Brass instruments are a little different in that they play a series of harmonics determined by fractions of the tube length ( nearly! the partially concal bore also distorts this a bit!) and the tube length of course is variable using valves or a slide. But trombonists have an infinite means of adjustment, valved brass also have 'triggers' and the bigger ones have compensatory tubing which comes into play when valves are used in combination. 

I won't attempt to go into reed instruments with their overblowing 12ths and octaves conical and cylindrical bores etc. it all gets too much for me! Just for interest though using the the above multiplier - the harmonics of the major chord (+fah) come out like this: doh 1.0000 (Harmonic 1.0) me 1.2599 ( " 1.25) tempered note is sharp fah 1.3348 ( " 1.3333) tempered note is a teeny bit sharp soh 1.4983 ( " 1.5) tempered note is a teeny bit flat doh 2.0000 ( " 2.0) Spot on! similar errors occur whichever key you are in. 

I have a horrible feeling now that some super theorist is now going to correct me ................... But HEY WHO CARES if it sounds good! All us jazzers bend notes anyway, and it's the ear that matters. (I still like bright tunes in C though).


22/02/10 -  Allan Wilcox

I enjoyed John's 10/02 contribution and learnt a lot from it, but I don't think it's quite true to say that I 'postulated the brightness of C'. I was simply picking up on the fact that Richard had asked for a scientific explanation for something HE had observed. It's a fascinating question though. I've grown up with the idea that the brightest key was E, based upon the fact that the slow third movement of Mahler's Symphony no. 4 in G modulates magnificently into E towards the end in preparation for the final movement, a song expressing a child's view of heaven, which is sung in E and really does seem appropriately bright in that context. If someone were to re-orchestrate it in F, say, so that the modulation brought us to a heavenly song in the key of D, would the effect be the same? I find it difficult to believe that the great composers chose (or choose) their keys quite that arbitrarily.  

That said, even if E is the brightest key, it may well be that C is the brightest we'll ever get to play jazz in. I haven't played anything in the key of E since I hung up my bow and threw away my music stand in 1978.


23/02/10 - Jeff Lewis

Hi Fred,

May I chuck my two penneth into the Strange Keys Debate....... Some people have got close, but the answer is even simpler....... and well known to anyone who is into transfer and restoration of early recordings. It is as follows:

The early days of acoustic recording coincided with the time when the good old US of A was still largely (thanks to the intransigence of a bloke called Edison) on DC electricity, locally generated with voltages varying wildly from place to place, with the result that it was impossible to guarantee the rotational speed of an electric motor. Synchronous (constant speed motors) motors requiring AC, were unusable because of this situation. To get around this problem, disc lathes (the cutting machines used for recording), were driven by falling weights in an attempt to produce even torque (turning power). This was OK as far as it went, but it did nothing to counter the fact that the drag on the cutting stylus is greater at the periphery than it is at the centre.

As the cutting stylus moved across the wax blank, the drag progressively reduced, allowing the lathe to speed up. 78rpm at the periphery may be 78.5 or 6 at the label!!. This has the opposite effect on playback, causing the pitch to fall progressively across the disc.

When you transfer an early record, you have to allow for this, but it is not possible to produce a standard pitch resample curve useable for all records, as the wax blanks used for recording at the time were not all of uniform density, and also had to be warmed before use (cooling and stiffening during the recording process), so the speed could be all over the place. When lacquers replaced wax at the end of the 20's the problem became even worse, and engineers often tried to split the difference by having 78 rpm occur somewhere in the middle of the disc!!!!!

As most of the record lathes were made in the USA, the problem spread all over the world, regardless of each country's power supply, and persisted until AC became the norm and lathes were either replaced or converted during the 1930's.

You will never find the problem on a cylinder due to constant drag at all times, and just to put the icing on the cake, Edison, who caused the problem in the first place, powered his lathes with DC current from lead acid batteries.

Sadly, all the musicians out there who have dedicated their lives to accurately reproducing the "keys" and phrasing on these records, must be applauded for surmounting some almost impossible ergonomic problems with instrument fingering etc.

Even more sadly, when the pitch on these records is corrected, you find that, often as not, the musicians are playing in the most "convenient" keys, and often the same as we do today.

Hope I have helped clarify matters.


24/02/10 - Phil Yates

I used to have a simple solution to the problem of practising along with vintage recordings in weird keys. As we all know, when the recordings from the 20s were transferred on to LPs, they usually retained the same pitch of the original 78s. Do you remember those fiddly things known as "dust bugs" which we attached to turntable decks on the opposite side to the pick-up arm, and were supposed to clean the record as it played? I used to stick 2p coins onto mine, which slowed the turntable down the the required pitch. 2 coins dropped it a semitone, 4 to a full tone and so on. 

Edison might be dead, but Heath-Robinson is alive and well! 


25/02/10 - Jeff Lewis

I thought there might have been a Dust Bugger out there....... or was it the cockroach in the coilspring?

2P for your thoughts Phil.


25/02/10 - Harry Roberts

Within the debate a fact that should not be overlooked, is that up until the mid 1930s all Brass Instruments had been made in High Pitch then the change was made to Low Pitch, [A=440 Vibs; per sec;] The difference in musical terms being High Pitch sounded nearly a semi tone higher than Low Pitch so the point is that anyone transcribing from early recordings should take this fact into account. Its worth pointing out that in the Brass Band world the change over was slow to be adopted and took place in the late 50s early 60s.


07/03/10 - John Muskett

1 David is probably correct about Bb, but from my own dabbling with a cornet it seems that all keys on a trumpet are of equal technical ease (or difficulty), although........

2 It may be that more tempering for keys remote from Bb is needed, as implied by Richard. However I rarely see trumpeters manipulating their valve slide hooks/rings. My own cornet doesn't have valve slides tuneable on the hoof and I suppose that sometimes I temper with the lips and sometimes play out of tune. On double bass I experience more intonation problems playing with a piano than without, but this may reflect my own incompetence rather than tempering problems.

3 A semitone discrepancy at 78 rpm is about plus or minus 4.5 rpm and a tone would equate to 9rpm. These seem large amounts, although I can see that the cumulative effects of pre A=440 pitch (as Harry says not standard), variations of studio piano tuning, erratic lathe speeds and several editings/remasterings could all add up significantly. Nevertheless the nearest semitone to the reproduction often gives a sensible key, and may agree with the sheet music, although the latter, for popular tunes, was sometimes published in several keys (high and low male and female voices). The most frequent discrepancy that I have encountered is of published G tunes sounding in Ab - and I suspect that they actually were played in Ab. Also is it possible that the person remastering has often adjusted the speed/pitch to agree with his/her supposition as to the original key in modern pitch?

Difficult to judge are recordings that are perhaps a semitone adrift and in an unusual key. Jimmie Noone's Apex Blues comes out in B on Red Hot Jazz. Should it be Bb (as picked by many British bands) or C (favoured, I understand, in the USA)? The jazz-on-line site has it in C and I would plump for this. 

1930s recordings surely suffered from pitch distortion. I have Ellington's Sophisticated Lady, In a Sentimental Mood and Caravan all apparently a semitone flat. And for Miles Davis's recording of Kind of Blue in 1959 one of the tape machines was running slow and on the issued LP three tunes were sharp, not being corrected for many years.

Now here's one for trombonists: the apparent choice of G for the Hot Five's Savoy Blues. The trombone upward glissando into the penultimate chorus appears to go from D above middle C to G, which I think, on the Bb tenor trombone, could only be achieved by blowing the (seldom if ever used) seventh harmonic (flat Ab in the first position) and with non-standard slide positions. The Armstrong Orchestra's 1939 recording comes out in Gb (!) but other tunes from the same session are Save It Pretty Mama (in E), Hear Me Talkin' To Ya (in A) and West End Blues (in D). Although they could have been played a semitone lower, it seems to me that a semitone higher for all is more likely (no trombone gliss on Savoy Blues). Of course Ory's own Creole Band recording sounds in F, where the gliss from (middle) C to F is possible and the gliss between bars 8 and 9 in the penultimate chorus may start on E (above middle C) going up to G.

On the same date as recording Savoy Blues the Hot Five also laid down Hotter Than That which appears in Eb. So.. was Savoy Blues really in F (and Hotter Than That really in Db?), or did Kid Ory have mastery of the seventh harmonic, or was there a C (tenor) trombone available, or am I mishearing things?


Main Menu

Please visit my Home Page