Notation Programs' Output Comparison

Have your scores reviewed by other users. Comment on old and new published scores and on publishers.
Knut
Posts: 867
Joined: 05 Oct 2015, 18:07
Location: Oslo, Norway

Re: Notation Programs' Output Comparison

Post by Knut » 18 Apr 2016, 07:13

jrethorst wrote:
John Ruggero wrote:if one selects and clicks any measure in any source, that one measure will be displayed in one window from every source so that one can compare them.
That's amazing.
+1 Wow!

Knut
Posts: 867
Joined: 05 Oct 2015, 18:07
Location: Oslo, Norway

Re: Notation Programs' Output Comparison

Post by Knut » 18 Apr 2016, 07:20

OCTO wrote:Knut, beautifully done! I enjoy very much your example.
One small thing to consider: ties in your example should be more flatten, IMO.
Thanks, OCTO!
There are a few other inaccuracies as well. I'll post the entire piece for comments when I'm done.

User avatar
John Ruggero
Posts: 1256
Joined: 05 Oct 2015, 14:25
Location: Raleigh, NC USA

Re: Notation Programs' Output Comparison

Post by John Ruggero » 18 Apr 2016, 12:16

We haven't had any discussion of scanning, so I thought that it would be interesting to show what SmartScore X Pro 10.2.4 can do with op 119 no. 1. I have not done any editing to these two scans. I was pleasantly surprised to see how accurate they were, given the complexity of the score. I usually use SmartScore for single line parts.

I used the IMSLP scans of the Brahms Complete Works edition for recognition. I converted the PDF to TIFF files (which SmartScore can also do on its own.) and then loaded them into SmartScore. In 5 seconds, I had the result shown.

The areas highlighted in pink are areas that the program thinks have rhythmic errors. In this case the program is fooled by the cross-staffing and by unmarked triplets; it sometimes omits bar lines as well. This is easily fixed with the editing tools, which are quite efficient.

The next step would be to clean it up with the editing tools and load it into Finale as an MusXML file, which SmartScore does with one command. All of this will take only a few minutes.
Brahms op 119 no 1 scan 1.tiff
(189.74 KiB) Downloaded 185 times
Brahms op. 119 no 1 scan 2.tiff
(238.5 KiB) Downloaded 169 times
Mac mini (OS 10.8.5) with dual monitors, Kurzweil Mark 5 with M-Audio Midisport 2 x 2,
Finale 2014d with GPO 4, JW Plug-ins, SmartScore X Pro, Adobe InDesign CS4,
Inkscape .48.5 and .91, FontForge 20150526
http://www.cantilenapress.com

User avatar
John Ruggero
Posts: 1256
Joined: 05 Oct 2015, 14:25
Location: Raleigh, NC USA

Re: Notation Programs' Output Comparison

Post by John Ruggero » 18 Apr 2016, 12:58

This is the slightly edited file loaded into a Finale default file:
Brahms op 119 no 1 into Finale 1.tiff
(125 KiB) Downloaded 164 times
Brahms op 119 no 1 into Finale 2.tiff
(158.71 KiB) Downloaded 161 times
The next step would be to load this file into a piano template with all of my personal settings, which also takes seconds, and start editing. A few measures will be erased and entered normally because it would be faster. However, the majority of the measures in this case are correct or need only a little basic editing. Formatting and fine editing would follow, as with all scores.

This is now my usual approach in entering printed (not hand-written) scores. I only enter normally (Simple Entry style) if the scanning approach does not yield a good enough result.
Mac mini (OS 10.8.5) with dual monitors, Kurzweil Mark 5 with M-Audio Midisport 2 x 2,
Finale 2014d with GPO 4, JW Plug-ins, SmartScore X Pro, Adobe InDesign CS4,
Inkscape .48.5 and .91, FontForge 20150526
http://www.cantilenapress.com

User avatar
John Ruggero
Posts: 1256
Joined: 05 Oct 2015, 14:25
Location: Raleigh, NC USA

Re: Notation Programs' Output Comparison

Post by John Ruggero » 19 Apr 2016, 14:36

Back to the topic at hand.

I am also now working on an engraving of the Brahms op. 119 no. 1 and have a source that is not online: the professional hand copy from which the first edition was engraved. The first edition was done very accurately and there are no substantial differences between this copy and the first edition so far. However, I have found quite a few errors in both the first edition and this copy which I will list when I post my engraving. Apparently, this and the other of these late piano pieces were published after the death of Brahms trusted editor Robert Keller, and they did not received kind of treatment to which Brahms was accustomed. There is an article about this that I am waiting to receive today:

Was Brahms a Reliable Editor? Changes Made in Opuses 116, 117, 118 and 119
Camilla Cai
Acta Musicologica
Vol. 61, Fasc. 1 (Jan. - Apr., 1989), pp. 83-101

i will point out one error: in the 2nd measure of the example below the 1/4-note D in the RH should be tied to the 16th in the next measure so the pattern of a single 16th leading to a dotted 8th matches what follows in the next two measures and a very similar place earlier at measure 24; the tie is clear in the autograph. The slurs are not, and the one between the A and D is most likely an error. A similar erroneous place occurs a few measure earlier in 34-35.
Brahms error 2.jpg
Brahms error 2.jpg (31.52 KiB) Viewed 2696 times
Brahms MS.jpg
Brahms MS.jpg (66.61 KiB) Viewed 2696 times
Mac mini (OS 10.8.5) with dual monitors, Kurzweil Mark 5 with M-Audio Midisport 2 x 2,
Finale 2014d with GPO 4, JW Plug-ins, SmartScore X Pro, Adobe InDesign CS4,
Inkscape .48.5 and .91, FontForge 20150526
http://www.cantilenapress.com

Knut
Posts: 867
Joined: 05 Oct 2015, 18:07
Location: Oslo, Norway

Re: Notation Programs' Output Comparison

Post by Knut » 19 Apr 2016, 16:54

John Ruggero wrote:Back to the topic at hand.

I am also now working on an engraving of the Brahms op. 119 no. 1 and have a source that is not online: the professional hand copy from which the first edition was engraved. The first edition was done very accurately and there are no substantial differences between this copy and the first edition so far. However, I have found quite a few errors in both the first edition and this copy which I will list when I post my engraving.

i will point out one error: in the 2nd measure of the example below the 1/4-note D in the RH should be tied to the 16th in the next measure so the pattern of a single 16th leading to a dotted 8th matches what follows in the next two measures and a very similar place earlier at measure 24; the tie is clear in the autograph. The slurs are not, and the one between the A and D is most likely an error.
I've been using the revised manuscript to correct mistakes, seeing as this must be the latest primary source available from IMSLP. I've found a few discrepancies but the error you point out seems to be in accordance with the 1st edition in this source.

User avatar
John Ruggero
Posts: 1256
Joined: 05 Oct 2015, 14:25
Location: Raleigh, NC USA

Re: Notation Programs' Output Comparison

Post by John Ruggero » 19 Apr 2016, 17:51

I now have the article I cited, which shows how Brahms was a selective proofreader who would correct minute matters of hairpin positioning and yet miss glaring errors of the type I just pointed out. He gives an almost exact error in opus 18 no. 6 measures 40—42, where a short tie is misinterpreted as a slur by the copyist who, as the writer says, seem not to understand the meaning of what he copies. These errors were then transcribed with great accuracy by the first edition. There also seems to be a culture of over-respect for the composer, so that obvious discrepancies were maintained since "the master knows best".

Brahms was apparently the typical composer when proofreading his own music, missing the obvious since he heard what wasn't there. He apparently was much better in editing other composer's music, an activity that occupied him more than any other composer of his level.

The copyist's score is nice to have, but since the first edition duplicates so exactly, it doesn't seem necessary for an accurate engraving of this piece. I believe Henle may have published it in facsimile. I have a copy through another source because of my Schenkerian research.
Mac mini (OS 10.8.5) with dual monitors, Kurzweil Mark 5 with M-Audio Midisport 2 x 2,
Finale 2014d with GPO 4, JW Plug-ins, SmartScore X Pro, Adobe InDesign CS4,
Inkscape .48.5 and .91, FontForge 20150526
http://www.cantilenapress.com

Knut
Posts: 867
Joined: 05 Oct 2015, 18:07
Location: Oslo, Norway

Re: Notation Programs' Output Comparison

Post by Knut » 19 Apr 2016, 21:20

I don't know this constitutes an 'obvious discrepancy', but there is an interesting one between the two manuscripts in the fifth measure.

The 1st edition follows the original manuscript:
Skjermbilde 2016-04-19 kl. 23.02.49.png
Skjermbilde 2016-04-19 kl. 23.02.49.png (309.81 KiB) Viewed 2678 times
While the copyist manuscript redistributes the last two notes in the lower right hand voice to the left hand:
Skjermbilde 2016-04-19 kl. 23.03.11.png
Skjermbilde 2016-04-19 kl. 23.03.11.png (331.65 KiB) Viewed 2678 times
The copied manuscript seems more consistent, because when the exact same measure appears again towards the end, these two notes are moved to the lower staff in both manuscripts as well as the 1st edition.

Personally though, I find m. 5 in the original manuscript and 1st edition to be cleaner, but in my mind it should appear identical both times, regardless of the chosen notation.

User avatar
John Ruggero
Posts: 1256
Joined: 05 Oct 2015, 14:25
Location: Raleigh, NC USA

Re: Notation Programs' Output Comparison

Post by John Ruggero » 19 Apr 2016, 23:08

That is definitely the type of inconsistency that I was referring to, Knut. The Complete Works also retains this.

The MS shows that Measures 5, 6, and 7 all started off with the 16ths entirely in the RH, which is probably the way Brahms improvised it. Then, he tried to find an easier hand distribution; measure 7 had to stay in the RH because of the stretch, measure 8 could be entirely redistributed; measure 5 partially.

Whether measure 5 is better split is really a matter of taste; many pianists would prefer the original. Perhaps the editors presented both versions as an uncomplicated way to give players a choice. But considering some of the other cases, it might also have been error or the misguided veneration syndrome. This was the kind of thing that his normal editor Robert Keller would have cleaned up for him.

Ah, the engraver copy is at IMSLP after all! Somehow I missed it. That's great, so all the important sources ARE at IMSLP.

Engraving this piece is definitely a workout when using Finale and it might be for other programs as well! Several notational and engraving issues previously discussed here and at MM are included. I can now see why you were so interested in engraving it.
Mac mini (OS 10.8.5) with dual monitors, Kurzweil Mark 5 with M-Audio Midisport 2 x 2,
Finale 2014d with GPO 4, JW Plug-ins, SmartScore X Pro, Adobe InDesign CS4,
Inkscape .48.5 and .91, FontForge 20150526
http://www.cantilenapress.com

Post Reply