Not to sound dismissive of your workflow 
but having done something similar in the past I find that pasting the ABC into
http://www.concertina.net/tunes_convert.html
and then grabbing the PDF is easier and produces similar output. You can use the PDF screen grab tool to pull out a portion in GIF or similar format.
(MTGuru, I know you know about the converter, I include this partly for people who may not be awareā¦)
Itās my understanding, listening to the recordings, this is the section of MtGuruās transcription that has the āornamentā in question.
Both listening to it, and playing along, it sounds pretty much just as he transcribes it, at least to me.
I just use ABCEdit for almost the whole job - when youāve typed up and edited the ABC to your satisfaction as shown in the standard notation window above, you simply click on ācopy score to clipboardā and then paste that (BMP) into a picture editor, crop away redundant āpaperā and compress it to a jpg and then post that in an image hosting service to link here. No need to convert the ABC separately or mess about with PDF grabbingā¦
Does the first triplet sound different from the other three to you?
I feel like I should be finding a cure for cancer or volunteering to get a politician elected (vote Ron Paul), but I guess that whistle is my medicine and music is my politics.
Iāve recorded Guruās rendering of it with the triplet and slowed it down, and it still seems to clunky, much like my retarded ring finger to begin with. Not the dainty embellishment that Mary plays. You be the judge.
<a href="http://www.jabuffington.com/music/slowlow.wav> Congress w/triplet
<a href="http://www.jabuffington.com/music/slowlow.wav> Slow ātripletā
I know, i know: maybe my medicine and my politics are unorthodox.
Yes, it does. The first E of the first triplet is not ācleanā. It has a weak fundamental and lots of chiffy overtones. The other triplets are cleaner.
Hereās a spectrum of the first triplet E. The fundamental is actually weaker than the next 2 spurious peaks at G and B, and there are strong first overtones at E, G and B. The whistle is effectively playing a Em chord (EGB) on this note! This is what the note looks like when itās āflubbedā.
And hereās a spectrum of the second triplet E. This looks more normal, with a strong fundamental a full 30 dB above the next peak at B.
Sounds like you arenāt tonguing the first not of the triplet, and your arenāt quite snappy enough to match Mary Bergin (not that any of us are). Sounds very pleasant, though; and itās harder on the low D. BTW, I would typically lift both fingers on this sort of triplet:
xxx xxo e
xxx ooo g
xxx xxo e
Rhythmically, the first two notes of the triplet are much shorter than then the last note: e/g/e rather than 3(ege. If you are an ABC-illiterate, thatās two sixteenth notes, followed by an eighth note.
Thanks for the clarification, Bloomfield. Like most things, the key is probably timingā¦and location, location, location.
Well, I like the sound of it. Perhaps intentional āflubbingā is a good skill to learn!
Actually, controlling a whistleās inherent chiff so that it always works for you instead of against you is one of those essential skills I havenāt mastered yet. 
No, not dismissive at all! The Tune-O-Tron thingy works well. In fact the back end seems to be abcm2ps. So the same abcm2ps formatting directives and ABC extensions work in both, and the output is basically identical.
I donāt recommend my approach to people who arenāt comfortable with command lines and batch files or shell scripts, but it gives you a lot of control.
Following Peter Labanās helpful (and now vanished) suggestion to examine the triplets in the Tartini tuner ā¦
Each of these graphs is a note-frequency analysis of the each of the 4 occurrences of the phrase
ag (3ege d
containing the triplet in part B of the posterās clip. The center vertical line is on the 1st E of the triplet. The whistle is tuned/played up to 50 cents sharp throughout, so the pitches tend to be slightly above the lines. But theyāre are clear enough as they glide near and around the note centers. The bottom panel shows relative loudness.
The first graph (red) shows the āflubbedā E, which is analyzed as an A# because of the spurious harmonics. This makes it a bit hard to see whatās happening here.
The second graph (green) is better, with the EGE triplet clearly shown, and the top G played fairly quickly, almost as a cut. There are no magic tricks here, just the same notes as transcribed.
The third and fourth graphs (blue and teal) are the easiest to read, and the notes are well delineated. The tongued A and B before the triplet are clealy separated. The triplet, played e/g/e, is relaxed, and the top note is clearly G, not A. You can see the dip in the first E and the hump in the G as the fingers move too quickly for the pitches to stabilize. You can see the smooth transition between the slurred final E and D. All the notes and articulations match the transcription.
As for timing, I think itās interesting to actually see that, at this speed, the notes occur on the order of ~100 milliseconds, sometimes less. And the transitions are faster still. Itās a reminder that details at this tiny time scale make a real difference for the successful execution of a tune.
Re-reading the posts, I may have an idea of whatās behind the miscommunication.
First, if the original poster is coming from the GHB world, he may be referring to cuts and taps and such as āgracenotesā. In my vocabulary, a gracenote is different and not much used in ITM dance music, in the art music sense of a leading or embellishing tone with a definite pitch and time value, that steals time from the preceding or following note. And though cuts and taps may be represented as small āgracenotesā in transcription (as I, in fact, use them), they are finger articulations, not graces. Maybe this is one source of terminology confusion.
Second, as a piper, he may be referring to fingers according to the āvent holeā convention rather than the āclose holeā convention as most whistler I know do. So when he talks of using the āA fingerā, perhaps he means what I would call the āG fingerā, or T3. And by an āA gracenoteā he may mean āa cut using the T3 fingerā.
In that context, maybe he is really asking if in the figure |z2 ag (3ege dz| there is a descending cut between the g and the start of the triplet, i.e. |z2 ag {a}(3ege dz|. Thatās certainly doable. After the g, you simultaneously lift T3 and drop B1 and B2, then execute the cut with T3. Of course, the answer is still: no, thatās not what is happening in the recording, according to both Tartini and oneās ears. Nor is a similar double-cut as Bloomfield mentioned, |z2 ag ({a}e{g}e) dz|. But at least that question makes more sense.
Anyway ā¦
Absolutely brilliant analysis, Dr. MTGuru. Your spectrograph interpretations border on rocket science! As far my confusion regarding the whistle terminology, thatās very possible as I guess that I use the ācutā and āgracenoteā terms interchangeably and probably not very consistently.
Cheers,
Buffpiper
P.s. I still think itās a tongue trick and an A cut!
J : A belated thank you for your ABC conversion details. Itāll take me some time to rouse the necessary IQ that you apparently take for granted to try this ā¦
Best,
Keith.
PS: And thanks to srt19170 for your helpful response to my question too. K
There is a fast cut on the attack of the first A in the triplet bar that is not notated in the transcription thatās been posted, and it happens all 4 times that bar comes around. Perhaps this is what you are referring to?
MTGuru: Your new image is not as good looking as the original. Is this before or after the Princess kissed you?
Keith.
Iām afraid I had no choice. All hail our benevolent Hypnotoad masters!