Sibelius

Music notation software is among the earliest software for desktop computers. SCORE in 1987, Finale came around in 1988, Capella in 1992, and Sibelius in 1993. Many others came and went during this time. Music notation software was so much more than the typical word processing or desktop publishing system. Specialized fonts were needed to display the music notation and there are many other variables for different instruments allowing individuals and others the ability to create complicated compositions in an inexpensive way.

Sibelius [SI] + [BAY] + [LEE] + [UHS] was originally developed for the Acorn system in 1986, then released on Windows and Macintosh in 1998-99. The software became very popular and in 2006 was purchased by the software giant AVID. The software was used enough to get a preservation assessment by the British Library in 2017 and draft status format description by the Library of Congress, written by the amazing Ashley!

Both reviews of the format emphasize the proprietary nature of the file format which has been used since the early versions. Aside from the early Acorn release, the Windows and Macintosh versions used a binary format with the SIB extension. They are actually quite easy to identify.

hexdump -C Sibelius-s01.sib | head
00000000  0f 53 49 42 45 4c 49 55  53 00 00 40 00 02 00 a4  |.SIBELIUS..@....|
00000010  f1 ed 00 00 00 30 00 00  00 02 00 00 00 01 00 00  |.....0..........|
00000020  00 2a 00 00 00 00 00 00  00 00 0f 53 49 42 45 4c  |.*.........SIBEL|
00000030  49 55 53 00 00 40 00 02  00 00 00 00 00 00 00 3a  |IUS..@.........:|
00000040  00 00 00 00 38 a1 28 06  b3 d2 2f 66 03 04 16 4e  |....8.(.../f...N|
00000050  5f 5c 8d f3 95 27 3e f1  2a 1b 68 de 08 81 e8 9a  |_\...'>.*.h.....|
00000060  ea 1c bf dd 54 0e 92 8d  4d be e3 34 ed 42 78 36  |....T...M..4.Bx6|
00000070  d2 e1 67 7b 8d f7 98 6a  3a 70 c4 8b 0b 08 7b 26  |..g{...j:p....{&|
00000080  f9 45 00 00 00 00 48 71  7c 4c 98 df 0b 38 7d 9d  |.E....Hq|L...8}.|
00000090  2a 2d 84 9c a4 39 0f 4d  da a2 cc 97 ad 3d b0 55  |*-...9.M.....=.U|

This is exactly how PRONOM and other identification methods determine they are Sibelius files. PRONOM has assigned the format fmt/696 and is looking for the hexadecimal bytes 0F534942454C495553.

The problem with this identification method is that all the Sibelius files are identified as such, regardless of version. As mentioned by Ashley, version of the software used is highly important as new features were added all the time making backwards compatibility difficult. Add in the fact that there were different releases for each version which would limit these features even more and I can see how a musician could get very frustrated. If you created a score in Sibelius 5 and tried to open in Sibelius 5 Student version, you may find your composition lacking in many ways. The only way to avoid compatibility issues is to always open in the latest “Ultimate” version. Sibelius Ultimate can open all versions of the SIB format back to version 2. The software even has an export feature which allows you to export back to a previous version stripping what is necessary to ensure compatibility.

Sibelius export to previous version

For those with a bunch of SIB files in their archives, how would you know which software version created the file? Well lets take a closer look at the bytes and see if we can find some patterns. Let it be known, I am not reverse engineering the format, just looking for patterns which will allow for proper identification!

I am not the first person to ask this question, many others want to know the versions of their SIB files. Thankfully others have found some clues on which bytes hold the version information. It seems we can determine the version based on 4 bytes shortly after the SIBELIUS string. Specifically bytes 10-13.

hexdump -C Sibelius2-s01.sib | head
00000000  0f 53 49 42 45 4c 49 55  53 00 00 08 00 22 00 47  |.SIBELIUS....".G|
00000010  98 4c 00 00 00 3a 00 00  00 00 4e 81 49 34 41 2c  |.L...:....N.I4A,|
00000020  fa 76 62 f9 71 53 a9 93  0f 54 1e 20 6c 63 61 4d  |.vb.qS...T. lcaM|
00000030  f7 b2 b0 a7 5d bd 82 3a  0d 86 02 8b f2 89 d2 a0  |....]..:........|
00000040  83 1f 8d e0 37 1b ed 1c  6a 8b 82 08 4b 6d 64 60  |....7...j...Kmd`|
00000050  71 59 e8 aa ef b1 3c df  5c 25 0a 9f 66 50 69 de  |qY....<.\%..fPi.|
00000060  2a d3 4e 2a cd 97 88 06  67 5f 50 64 0f 8f 86 2b  |*.N*....g_Pd...+|
00000070  08 0d 3f f7 80 26 e0 63  f6 7d 4e f8 e7 c0 3f fc  |..?..&.c.}N...?.|
00000080  7a 77 ea b3 4a b9 30 59  13 47 6e 09 0a 0b ae 3c  |zw..J.0Y.Gn....<|
00000090  c1 93 85 f6 41 f8 58 22  4b 92 35 3f b2 f5 3f 9d  |....A.X"K.5?..?.|

From what others have gathered and updating it with more recent versions I have come up with a list.

VersionHex 10-13
Sibelius 1.200 00 00 0E
Sibelius 2.x00 08 xx xx 
Sibelius 3.x00 0A xx xx 
Sibelius 4.x00 1B xx xx 
Sibelius 5.000 2D 00 03 
Sibelius 5.100 2D 00 0D 
Sibelius 5.2.x – 5.400 2D 00 10 
Sibelius 6.0.x00 36 00 01 
Sibelius 6.100 36 00 17 
Sibelius 6.200 36 00 1E 
Sibelius 7.000 39 00 0C 
Sibelius 7.0.1 – 7.0.200 39 00 0E 
Sibelius 7.0.300 39 00 13 
Sibelius 7.1.000 39 00 15 
Sibelius 7.1.2 – 7.1.300 39 00 16 
Sibelius 7.5.x00 3D 00 0E 
Sibelius 8.0.0 – 8.0.100 3D 00 10 
Sibelius 8.1.x00 3E 00 00 
Sibelius 8.200 3E 00 01 
Sibelius 8.300 3E 00 02 
Sibelius 8.4.x00 3E 00 06 
Sibelius 8.5.x00 3E 00 07 
Sibelius 8.6.x, 8.7.0, 8.7.100 3F 00 00 
Sibelius 8.7.2, 2018.1, 2018.4.x, 2018.5, 2018.6, 2018.700 3F 00 01 
Sibelius 2018.11, 2018.1200 3F 00 02
Sibelius 2019.100 3F 00 04
Sibelius 2019.4.x, 2019.5, 2019.7, 2019.900 3F 00 06
Sibelius 2019.1200 3F 00 07
Sibelius 8.6-2019.1200 3F 00 0A
Sibelius 2020.100 3F 00 0B
Sibelius 2020.3, 2020.600 40 00 01
Sibelius 2020.900 40 00 02
Sibelius 2022.500 40 00 03
Sibelius 2022.1100 41 00 02
Sibelius 2022.1200 42 00 00
Sibelius 2023.300 42 00 01
Sibelius 2023.800 43 00 07
Sibelius 2024.3.100 44 00 01

That is a lot of versions and I feel there may be some gaps that still need to be identified. It appears that the first two bytes are the major version and the second set of bytes is the minor version. Although it looks like a few major version bytes span across a few software versions. With this chart, one could be very specific in identifying which Sibelius version wrote the file, but for archiving purposes it seems we can group many of these capturing just the major version. The export screenshot above seems to have broken down significant changes and grouped similar formats together, the biggest being 8.6 through 2019.12. A comparison of “student” and “first” formats don’t have any obvious bytes which indicate as such, so for now they are all lumped together.

There is one other similar format which needs to be mentioned. Sibelius Scorch was a product made to share scores online. This has been replaced with Sibelius Cloud Publishing, but for awhile was the best way to share a score with others in a way that protected the original. I have no idea how they were made, but sites like scorestreet.net and sibeliusmusic.com were sites you could upload your score to for sharing. Some SCO files appear to have a PDF embedded within them for proper printing.

hexdump -C smd_h_0000000000097761.sco | head
00000000  0f 43 43 53 43 4f 52 43  48 00 00 36 00 1e 00 c0  |.CCSCORCH..6....|
00000010  d4 55 00 00 00 30 00 00  00 01 00 00 00 01 00 00  |.U...0..........|
00000020  00 22 0f 43 43 53 43 4f  52 43 48 00 00 36 00 1e  |.".CCSCORCH..6..|
00000030  00 00 00 00 00 00 00 3a  00 00 00 00 03 56 11 b9  |.......:.....V..|
00000040  70 dc fe 90 50 48 30 df  eb 39 88 23 8e 88 78 bf  |p...PH0..9.#..x.|
00000050  da ab ab 5b e2 13 98 89  66 eb 94 67 8d 16 00 00  |...[....f..g....|
00000060  00 00 cf 6f 0c 67 85 ec  57 90 e5 c1 ea 8a eb 9f  |...o.g..W.......|
00000070  c8 13 d2 1d 75 bd a5 9f  eb b9 ef 1d 25 79 45 2c  |....u.......%yE,|
00000080  05 bb 74 41 e8 8f 27 6a  01 07 d0 f5 3b 17 ce 87  |..tA..'j....;...|
00000090  7b c2 82 d9 41 6b 82 2f  d8 b8 17 32 fa d3 59 05  |{...Ak./...2..Y.|

I am not sure the best way to handle all the different versions within the PRONOM registry. I went ahead and made a few signatures based on the export dialog of Sibelius 2024. Even with combining a few together, it leaves us with 17 new PUID’s. Maybe further discussion can refine these down a bit more? Regardless, each file can be associated with a specific Sibelius version, making it easier to open and migrate if needed without fear of opening in the wrong version. Take a look at some samples and my signatures on my GitHub page and let me know if there is a better way.

Finale

The amazing Ashley recently did a little writeup on the Sibelius music notation software. I thought I would take the opportunity to talk about another music notation software which needs a little update. Finale was created in 1987 for the Macintosh by a company called Coda Music and became quite popular with musicians and composers. The ability to use a computer to typeset a musical score was a huge advancement. This was all possible by the use of music notation fonts.

Finale was originally written by Coda Music Technology, owned for a time by Net4Music, now currently owned by MakeMusic. Over the years there has been additional products developed along side Finale.

The first version of Finale was developed for the Macintosh and didn’t have an extension. But by version 3.5 there was a comparable Windows version and the use of the extension .MUS. In order to share the files between the different platforms Finale also created an ETF file, which instead of the binary MUS the ETF is a plain text “transportable” file.

Finale 1.0 HyperCard HelpStack

Both formats are based on the Enigma or “Environment for Notation Intuitive Graphic Music Algorithms” format. These formats were last used with Finale 2012 when a new format took over in 2014. Let’s start from the beginning.

hexdump -C Finale1-s01 | head
00000000  46 69 6e 61 6c 65 aa 20  31 2e 30 2e 30 20 45 4e  |Finale. 1.0.0 EN|
00000010  49 47 41 20 53 74 72 75  63 74 75 72 65 73 20 43  |IGA Structures C|
00000020  6f 70 79 72 69 67 68 74  20 31 39 38 37 20 62 79  |opyright 1987 by|
00000030  20 43 6f 64 61 2e 20 41  6c 6c 20 72 69 67 68 74  | Coda. All right|
00000040  73 20 72 65 73 65 72 76  65 64 2e 20 50 61 74 65  |s reserved. Pate|
00000050  6e 74 20 50 65 6e 64 69  6e 67 00 00 00 00 00 00  |nt Pending......|
00000060  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00000080  01 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000090  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|

This is a sample of the very first version of Finale. Currently not identifiable by PRONOM. You may also noticed in this version it was called ENIGA.

hexdump -C Finale2.6.3 | head
00000000  46 69 6e 61 6c 65 28 54  4d 29 20 31 2e 38 20 43  |Finale(TM) 1.8 C|
00000010  6f 70 79 72 69 67 68 74  20 31 39 38 37 20 62 79  |opyright 1987 by|
00000020  20 43 6f 64 61 2e 20 41  6c 6c 20 72 69 67 68 74  | Coda. All right|
00000030  73 20 72 65 73 65 72 76  65 64 2e 00 00 00 00 00  |s reserved......|
00000040  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00000080  01 01 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000090  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00000200  00 00 00 09 00 00 02 00  00 00 46 4e 50 65 74 72  |..........FNPetr|

A file from version 2.6.3 shows a different format structure, also not currently identified by PRONOM.

hexdump -C F35-s01.mus | head
00000000  45 4e 49 47 4d 41 20 42  49 4e 41 52 59 20 46 49  |ENIGMA BINARY FI|
00000010  4c 45 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |LE..............|
00000020  46 69 6e 61 6c 65 28 52  29 20 33 2e 35 20 43 6f  |Finale(R) 3.5 Co|
00000030  70 79 72 69 67 68 74 20  28 63 29 20 31 39 39 35  |pyright (c) 1995|
00000040  20 43 6f 64 61 20 4d 75  73 69 63 20 54 65 63 68  | Coda Music Tech|
00000050  6e 6f 6c 6f 67 79 00 00  00 00 00 00 00 00 00 00  |nology..........|
00000060  00 02 00 00 00 00 7c 02  08 00 00 00 03 03 50 03  |......|.......P.|
00000070  46 49 4e 00 57 49 4e 00  02 04 50 03 03 03 50 03  |FIN.WIN...P...P.|
00000080  00 00 00 00 00 00 00 00  00 00 00 00 7c 02 08 00  |............|...|
00000090  00 00 03 03 50 03 46 49  4e 00 57 49 4e 00 02 04  |....P.FIN.WIN...|

By Version 3 we see the format stabilize and this header is used until Finale 2012. There was other various products which also used the format so there is some variation.

hexdump -C Tutorial1a.mus | head
00000000  45 4e 49 47 4d 41 20 42  49 4e 41 52 59 20 46 49  |ENIGMA BINARY FI|
00000010  4c 45 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |LE..............|
00000020  50 72 69 6e 74 4d 75 73  69 63 28 52 29 20 32 30  |PrintMusic(R) 20|
00000030  31 30 20 43 6f 70 79 72  69 67 68 74 20 31 39 39  |10 Copyright 199|
00000040  38 2d 32 30 30 39 20 4d  61 6b 65 4d 75 73 69 63  |8-2009 MakeMusic|
00000050  20 49 6e 63 2e 00 00 00  00 00 00 00 00 00 00 00  | Inc............|
00000060  00 02 0e 01 00 00 6a 02  0e 00 00 00 04 02 02 0b  |......j.........|
00000070  46 49 4e 00 57 49 4e 00  03 04 02 0b 0d 02 00 0b  |FIN.WIN.........|
00000080  00 00 00 00 00 00 00 00  00 00 00 00 6d 08 0d 00  |............m...|
00000090  00 00 31 02 00 0f 4e 54  52 00 4d 41 43 00 10 02  |..1...NTR.MAC...|

The current PRONOM identification for fmt/397 is looking for the “ENIGMA BINARY FILE” bytes but also the string “Finale(R)”, so this PrintMusic variation is not identified correctly.

Another format that is a little more rare to see, but is part of the Finale formats collection. Finale Performance Assessment File (.fpa) is an older format discontinued in 2007, but has a similar format. It was a tool similar to the current SmartMusic tool.

hexdump -C Tuba.FPA | head
00000000  46 49 4e 41 4c 45 20 50  45 52 46 4f 52 4d 41 4e  |FINALE PERFORMAN|
00000010  43 45 20 41 53 53 45 53  53 4d 45 4e 54 00 00 00  |CE ASSESSMENT...|
00000020  46 69 6e 61 6c 65 28 52  29 20 32 30 30 35 20 43  |Finale(R) 2005 C|
00000030  6f 70 79 72 69 67 68 74  20 28 63 29 20 31 39 38  |opyright (c) 198|
00000040  37 2d 32 30 30 34 20 4d  61 6b 65 4d 75 73 69 63  |7-2004 MakeMusic|
00000050  21 20 49 6e 63 2e 00 6f  6c 6f 67 79 00 00 00 00  |! Inc..ology....|
00000060  00 02 06 00 00 00 68 06  09 00 00 00 16 02 00 09  |......h.........|
00000070  46 49 4e 00 57 49 4e 00  01 04 01 09 16 02 00 09  |FIN.WIN.........|
00000080  00 00 00 00 00 00 00 00  00 00 00 00 68 07 0d 00  |............h...|
00000090  00 00 0a 01 00 0a 46 49  4e 00 57 49 4e 00 03 03  |......FIN.WIN...|

As for the Enigma Transportable File, there is a couple variations.

hexdump -C Finale1-s02.etf | head
00000000  45 4e 49 47 4d 41 20 74  72 61 6e 73 70 6f 72 74  |ENIGMA transport|
00000010  61 62 6c 65 20 66 69 6c  65 0d 45 4e 49 47 4d 41  |able file.ENIGMA|
00000020  20 53 74 72 75 63 74 75  72 65 73 20 43 6f 70 79  | Structures Copy|
00000030  72 69 67 68 74 20 31 39  38 37 20 62 79 20 43 6f  |right 1987 by Co|
00000040  64 61 2e 20 41 6c 6c 20  52 69 67 68 74 73 20 52  |da. All Rights R|
00000050  65 73 65 72 76 65 64 2e  20 50 61 74 65 6e 74 20  |eserved. Patent |
00000060  50 65 6e 64 69 6e 67 2e  0d 0d 5e 6f 74 68 65 72  |Pending...^other|
00000070  73 0d 5e 46 4e 28 30 29  20 22 50 65 74 72 75 63  |s.^FN(0) "Petruc|
00000080  63 69 22 0d 5e 49 55 28  30 29 20 31 20 30 20 2d  |ci".^IU(0) 1 0 -|
00000090  38 30 20 32 20 30 20 2d  33 31 36 20 0d 5e 49 55  |80 2 0 -316 .^IU|

hexdump -C Finale37-Sample.etf | head
00000000  45 4e 49 47 4d 41 20 54  52 41 4e 53 50 4f 52 54  |ENIGMA TRANSPORT|
00000010  41 42 4c 45 20 46 49 4c  45 0d 0d 5e 68 65 61 64  |ABLE FILE..^head|
00000020  65 72 0d 5e 30 31 20 22  46 69 6e 61 6c 65 28 52  |er.^01 "Finale(R|
00000030  29 20 33 2e 37 20 43 6f  70 79 72 69 67 68 74 20  |) 3.7 Copyright |
00000040  28 63 29 20 31 39 38 37  2d 31 39 39 36 20 43 6f  |(c) 1987-1996 Co|
00000050  64 61 20 4d 75 73 69 63  20 54 65 63 68 6e 6f 6c  |da Music Technol|
00000060  6f 67 79 22 0d 5e 30 32  20 31 20 30 20 30 20 30  |ogy".^02 1 0 0 0|
00000070  20 0d 5e 30 33 20 31 32  30 20 31 31 20 39 20 0d  | .^03 120 11 9 .|
00000080  5e 30 34 20 22 22 0d 5e  30 35 20 35 37 36 37 32  |^04 "".^05 57672|
00000090  32 30 34 20 0d 5e 30 36  20 22 46 49 4e 22 0d 5e  |204 .^06 "FIN".^|

The current signature of ETF files is only able to correctly identify the later version of the string in all caps. The fmt/398 PRONOM ID could use an alternate signature to ensure all variations are identified correctly. There is a couple versions of the specification out there, but does not add much to what is known.

Starting in 2014 Finale starting using a new file format to store its notations. The native format now uses the MUSX extension. This new format uses a ZIP container to store all the data. Let’s take a look at the inside.

Path = Finale26-s01.musx
Type = zip
Physical Size = 98608

   Date      Time    Attr         Size   Compressed  Name
------------------- ----- ------------ ------------  ------------------------
2022-12-19 16:28:36 .....           34           34  mimetype
2022-12-19 16:28:36 .....          252          168  META-INF/container.xml
2022-12-19 16:28:36 .....          347          218  NotationMetadata.xml
2022-12-19 16:28:36 .....         1163          821  presets/10001.preset
2022-12-19 16:28:36 .....          649          544  presets/1.preset
2022-12-19 16:28:36 .....        96140        96155  score.dat
------------------- ----- ------------ ------------  ------------------------
2022-12-19 16:28:36              98585        97940  6 files

The mimetype file appears to be “application/vnd.makemusic.notation”

The NotationMetadata.xml file stores much of the information needed and begins with the root tag.

<metadata version="26.2" xmlns="http://www.makemusic.com/2012/NotationMetadata">

It seems the presence of the NotationMetadata.xml file and the mimetype would be sufficient for identification in a container signature.

The current version of Finale can export to a few different “Music XML” versions. This includes MUSICXML, regular XML, and a compressed MXL file. The only one needs attention is the compressed MXL file and added to PRONOM. It already has a PUID, fmt/897, but no signature. Here is what it looks like inside the ZIP container.

Path = Finale27-s01.mxl
Type = zip
Physical Size = 4737

   Date      Time    Attr         Size   Compressed  Name
------------------- ----- ------------ ------------  ------------------------
2024-02-07 23:55:50 .....           34           34  mimetype
2024-02-07 23:55:50 D....            0            2  META-INF
2024-02-07 23:55:50 .....          202          144  META-INF/container.xml
2024-02-07 23:55:50 .....        18004         1996  Finale27-s01.musicxml
2024-02-07 23:55:52 .....        17554         1953  p1.musicxml
------------------- ----- ------------ ------------  ------------------------
2024-02-07 23:55:52              35794         4129  4 files, 1 folders

Looks like a standard identifiable MUSICXML file within the container with a mimetype of “application/vnd.recordare.musicxml”. The MUSICXML file will be impossible to use for identification because of the variable file name, but the mimetype should do just fine.

Hopefully that covers all the major formats that need identification. I saw on a list that I will soon be working on an old Macintosh which has hundreds of Finale files, I hope these updates cover those needs! Take a look at my GitHub for my signatures and plenty of samples.

PianoSoft DOM-30

I often find myself at a thrift store looking through the well used Compact Discs. Often see the same ones over and over, but occasionally finding a gem. While looking through a set of discs, a few caught my eye. When I pulled one out to look at the cover I noticed it was not your typical CD. Opening the cover I was greeted with a 3.5 floppy inside the jewel case. That was a fun surprise.

The 3.5 inch floppy disk appears to be made specifically for the Yamaha Disklavier piano’s. The disk had the appearance of your typical double density floppy. Unfortunately, when I inserted the disk I was greeted with the error, “no mountable file systems”. I was however able to use ddrescue and make a disk image. Here is what the disk header looks like:

hexdump -C Yamaha_RazzleDazzle.img | head
00000000  e5 e5 e5 e5 e5 e5 e5 e5  e5 e5 e5 e5 e5 e5 e5 e5  |................|
*
00000200  f9 ff ff 03 40 00 05 60  00 07 80 00 47 00 00 0b  |....@..`....G...|
00000210  c0 00 0d e0 00 0f f0 ff  11 20 01 13 40 01 15 f0  |......... ..@...|
00000220  ff 17 80 01 19 a0 01 1b  c0 01 42 00 00 1f 00 02  |..........B.....|
00000230  21 20 02 23 40 02 25 60  02 27 80 02 29 f0 ff 2b  |! .#@.%`.'..)..+|
00000240  c0 02 2d e0 02 2f f0 ff  31 20 03 33 40 03 35 60  |..-../..1 .3@.5`|
00000250  03 37 80 03 39 f0 ff 3b  c0 03 3d e0 03 3f 00 04  |.7..9..;..=..?..|
00000260  41 f0 ff ff 4f 04 45 60  04 48 f0 ff 49 f0 ff 00  |A...O.E`.H..I...|
00000270  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|

The header actually has 512 bytes of the E5 hex values. Not a FAT12 file system for sure. A little farther into the disk image I can see what appears to be file listing.

00000e00  53 4f 4e 47 20 20 20 20  50 30 31 00 00 00 00 00  |SONG    P01.....|
00000e10  00 00 00 00 00 00 00 00  00 00 02 00 00 20 00 00  |............. ..|
00000e20  53 48 4f 52 54 20 20 20  50 30 32 20 00 00 00 00  |SHORT   P02 ....|
00000e30  00 00 00 00 00 00 fa ac  a2 16 0a 00 00 18 00 00  |................|
00000e40  54 4f 59 20 20 20 20 20  50 30 33 20 00 00 00 00  |TOY     P03 ....|
00000e50  00 00 00 00 00 00 05 ad  a2 16 10 00 00 18 00 00  |................|
00000e60  53 57 45 45 54 20 20 20  50 30 34 20 00 00 00 00  |SWEET   P04 ....|
00000e70  00 00 00 00 00 00 12 ad  a2 16 16 00 00 20 00 00  |............. ..|
00000e80  56 49 4f 4c 49 4e 20 20  50 30 35 20 00 00 00 00  |VIOLIN  P05 ....|
00000e90  00 00 00 00 00 00 40 ad  a2 16 1e 00 00 30 00 00  |......@......0..|
00000ea0  51 55 49 45 54 20 20 20  50 30 36 20 00 00 00 00  |QUIET   P06 ....|
00000eb0  00 00 00 00 00 00 50 ad  a2 16 2a 00 00 18 00 00  |......P...*.....|
00000ec0  52 41 5a 5a 4c 45 20 20  50 30 37 20 00 00 00 00  |RAZZLE  P07 ....|
00000ed0  00 00 00 00 00 00 83 ad  a2 16 30 00 00 28 00 00  |..........0..(..|
00000ee0  4c 45 54 53 20 20 20 20  50 30 38 20 00 00 00 00  |LETS    P08 ....|
00000ef0  00 00 00 00 00 00 9a ad  a2 16 3a 00 00 20 00 00  |..........:.. ..|
00000f00  50 49 41 4e 4f 44 49 52  46 49 4c 20 00 00 00 00  |PIANODIRFIL ....|
00000f10  00 00 00 00 00 00 a1 ad  a2 16 43 00 00 18 00 00  |..........C.....|
00000f20  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00001c00  fe 00 00 00 20 00 00 43  4f 4d 2d 45 53 45 51 51  |.... ..COM-ESEQQ|
00001c10  31 31 56 31 2e 30 30 80  00 00 00 d9 01 00 00 00  |11V1.00.........|
00001c20  20 00 00 01 58 00 00 20  20 20 20 20 20 20 20 20  | ...X..         |
00001c30  20 20 20 20 20 20 20 20  20 20 20 20 20 20 20 20  |                |

This gave me a few clues. A few google searches I came across a project on Hackaday about “Hacking Yamaha Disklavier Floppies“. I was eager to test out the python software and see if could export the files on the disk image. To my disappointment, the python script could not read my disk image. So I reached out to the author. After sharing a couple disk images with him, he was able to enable support for this different type of disklavier disk.

python3 disklav.py -t Yamaha_RazzleDazzle.img 
Loading file...OK
Format: PianoSoft DOM-30
Disk: PPC 1919       
Title: RAZZLE DAZZLE
--------------------------------------------------------------------
Track 01 - Song Without Words
Track 02 - Shortenin' Bread Boogie
Track 03 - Toy Bugle
Track 04 - Sweet Tooth
Track 05 - The Mysterious Violin
Track 06 - Quiet Moment
Track 07 - Razzle Dazzle
Track 08 - Let's Have A Party!

When I used the extract option with the software I was rewarded with eight files with the .FIL extention.

sf Yamaha_RazzleDazzle-track01.fil           
---
siegfried   : 1.10.1
scandate    : 2023-12-04T22:38:04-07:00
signature   : default.sig
created     : 2023-12-04T22:37:35-07:00
identifiers : 
  - name    : 'pronom'
    details : 'DROID_SignatureFile_V116.xml; container-signature-20231127.xml'
---
filename : 'Yamaha_RazzleDazzle-track01.fil'
filesize : 6409
modified : 2023-12-04T22:28:34-07:00
errors   : 
matches  :
  - ns      : 'pronom'
    id      : 'UNKNOWN'
    format  : 
    version : 
    mime    : 
    class   : 
    basis   : 
    warning : 'no match'

No surprise the file was not known to PRONOM. I doubt these files have made a big appearance in many archives.

hexdump -C Yamaha_RazzleDazzle-track01.fil | head
00000000  fe 00 00 00 20 00 00 43  4f 4d 2d 45 53 45 51 51  |.... ..COM-ESEQQ|
00000010  31 31 56 31 2e 30 30 80  00 00 00 d9 01 00 00 00  |11V1.00.........|
00000020  20 00 00 01 58 00 00 20  20 20 20 20 20 20 20 20  | ...X..         |
00000030  20 20 20 20 20 20 20 20  20 20 20 20 20 20 20 20  |                |
*
00000120  20 20 20 20 20 20 20 00  76 04 02 00 1e ff 00 00  |       .v.......|
00000130  ff ff ff ff ff 00 ff 00  00 00 00 ff ff 00 00 00  |................|
00000140  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000150  00 00 00 00 00 02 02 02  02 02 02 02 02 01 01 01  |................|
00000160  01 01 01 01 01 01 01 01  01 01 01 01 01 00 00 00  |................|

The header of a .FIL file has an ascii string “COM-ESEQ”. A little investigation shed a little light on the format. Turns out there is a proprietary format Yamaha developed called “E-SEQ”. The E-SEQ format is compatible with all Disklaviers, Clavinova digital pianos, and a few other Yamaha products. I was curious if the format was something similar to a MIDI file, which was commonly used with early keyboard systems, but I was unable to find anything to suggest they are similar in any way. Yamaha does mention there are tools out there to convert an E-SEQ file to a SMF (Standard Midi File) which was used on other systems.

There is another tool called PPFBU which can be used to extract a disk image from a Disklavier floppy and the E-SEQ files. Along with a companion tool called MID2PianoCD claims to be able to convert a E-SEQ to a WAV or MP3, although I haven’t had much luck.

Another set of tools are available here, they allow for copying of a disk and converting back and forth from E-SEQ and Midi. A text file in DVUtils from the link has the following background about the disk format and the FIL files:

DISKLAVIER FILES AND DISCS


Yamaha Disklavier discs are always on Double Density (2DD) media, High Density (HD)discs, which are more common nowadays,  will not work. Furthermore, they are formatted to 720 Kbytes not the default of 1.2 Mbytes.  The original discs are copy protected.  This has been achieved by placing invalid data on the first sector.  As DOS and Windows always refer to this sector to check out a floppy, they will report that the discs are bad.  The Yamaha machinery ignores the first sector so it reads them normally.

The music files on a Disklavier disc have the extension .FIL . They are frequently identified with titles like PIANO001.FIL but sometimes they have names similar to DOS like MUSIC1.FIL.  In addition to the music files, there is an index file on the disc.  This contains a list of the active music files on the disc, their titles, and pointers to their position on the disc.  The index file is always called PIANODIR.FIL and always has a size of 6 Kbytes. In order to set up a Disklavier disc to function on a Disklavier, you must first copy the music files onto it in Disklavier format (ESEQ) and then run the ESEQ EXPLORER program to build the index file.

Although there are many Disklavier Piano’s still out there and quite expensive if you want to pick one up for yourself, the websites dedicated to the format as slowly disappearing. One archived website has plenty of sample files to download and write to a floppy for use if you happen to have a Disklavier.

You can check out the signature I put together for the E-SEQ on my Github. Might be good to explore the disk image format more and add that as well to PRONOM for identification as well.