Software for the OnHandPc : Games, Clocks, and Applications

onHand

The SND-Sound Format

Translation provided by Ko.

What format is the data of the sound file XXX.SND in?

I have analyzed the sound file XXX.SND used in Ruputer (^_^)v.
Is it of any help for the programming of utilities to figure out the formats of the tempo data and the first half of the note data?
The followings are the outcomes of my analysis up to the present.

<I could solve the mistery of the data format of the first half of the tempo data (^_^)v> ('98.7.21)

Finally, I solved the last mistery, the data format of the first half of the tempo (^_^)v
It is only speculative, though..... (^_^;

<I found two different types of formats of the data of the musical notes and rests.> ('98.7.20)

When I was testing the beta program of "RupSound" which Mr. Arataku had made, I found that there are more than one "format type of the musical notes and rests"; thereby I had him make necessary corrections for them in a hurry.
With a feat of strength (^_^;, I examined the relation between the data of notes and the tempo and found the followings:
      ( 1 ) There are two kinds of formats for the first half data of the musical notes and rests.
      ( 2 ) The two formats can be discerned by checking if the initial byte of the second half of the tempo data is "00" or "02."
      ( 3 ) The first half of the data format of the musical notes and rests shows the duration of the physical sound.
These are the hypotheses which I have made.
I will explain on the details of these hypotheses below.

<A part of the format of the tempo data was understood.> ('98.7.17)

I have found the way to understand the speed of the tempo, which is as follows:
The last two bytes of the second half of the data show the speed of the tempo.
As they are handling two-byte long data, the upper byte and the lower byte changes their positions on the memory.

For examples, "XX XX 80 02" -> "0280(H)" -> "640"
"XX XX 78 00" -> "0078(H)" -> "120"

Therefore, these two bytes in the second half explain the speed of the tempo.

<Corrections of my mistakes. m(_._)m> ('98.7.17 added)

In the list of the tempo data,
     3 0       07 00 10 00     00 00 1F 00 should be 07 00 10 00     00 00 1E 00.
2 1 4       07 00 03 00     07 00 03 00 should be 07 00 03 00     02 00 D6 00. I copied them wrong.
     4 8 Data for this were omitted by mistake.
I apologize for any possible troubles caused by these mistakes of mine.

<File size>('98.6.25 added)

When I made a big piece of music, the PC utility "Sound creation" returned an error massage "最後の音符1個が書けませんでした｡" [which means "The last one note could not be written."]
When I checked the size of the file which had been written, the size was "2,046 bytes."
As one musical note takes 4 bytes to 7 bytes of data space, "2,048 bytes" might be the maximum limit.
Therefore, please be careful of this when you make big pieces of music.

<Data sequence>

The data sequence of the sound file is as follows:

07 00 xx xx xx xx xx xx ...... FF FF xx 00 xx xx xx xx xx xx ... FF FF

tempo (first half) pitch note note pitch tempo (second half) pitch note pitch note

The initial 4 bytes show the data on the tempo, and then a pitch (1 byte) + a note (1 byte) combination follows, corresponding to one sound (musical symbol) except for the cases of the 1/2 note and parts of the whole note.
The 1/2 note and parts of the whole note are expressed using a couple of numbers of the sets of a pitch (1 byte) + a note (1 byte).
These data continue to the terminator "FF FF" (2 bytes).
After this, the initial 4 bytes show the data on the tempo, and then a pitch (1 byte) + a note (1 byte) combination follows, corresponding to one sound (musical symbol).

The second half of the data are expressed as a pitch (1 byte) + a note (1 byte) combination correspondind to one sound (musical symbol), irrespective of the kinds of notes.
The terminator "FF FF" (2 bytes) shows the last of the data.

For examples, in the case of "tempo 120, c4 (do), 1/4 note, e5 (mi), 1/2 note,"
the data are as follows:

07 00 04 00 1A 38 2A 30 FF FF 00 00 78 00 1A 0A 2A 0E FF FF

tempo (first half) pitch note note pitch tempo (second half) pitch note pitch note

<tempo> ('98.7.21 added)

The tempo data consist of the first 4 bytes and the latter 4 bytes. The contents of their data vary depending on the kind of tempo.

The third byte shows the clock number to define the physical speed of the tempo.
The first byte in the second half defines the number of clocks of the 1/4 note. "00" shows 8 clocks and "02" shows 6 clocks.
From 1 and 2, we get the speed of the tempo by using the following equation.

tempo = 3,840 /("clock number in 1" x "number of clocks of the 1/4 note")

The 3rd and 4th bytes in the second half show the number of the tempo.

"1F 00" -> "00 1F" -> 30
"E0 10" -> "10 E0" -> 480

tempo data physical speed expressed

first second 3rd byte 1/4 note calculated 3&4th bytes

30 07 00 10 00 00 00 1F 00 16 8 30 30

32 07 00 0F 00 00 00 20 00 15 8 32 32

34 07 00 0E 00 00 00 22 00 14 8 34 34

36 07 00 12 00 02 00 24 00 18 6 36 36

38 07 00 0D 00 00 00 26 00 13 8 37 38

40 07 00 0C 00 00 00 28 00 12 8 40 40

44 07 00 0B 00 00 00 2C 00 11 8 44 44

46 07 00 0E 00 02 00 2E 00 14 6 46 46

48 07 00 0A 00 00 00 30 00 10 8 48 48

50 07 00 0D 00 02 00 32 00 13 6 49 50

54 07 00 09 00 00 00 36 00 9 8 53 54

58 07 00 0B 00 02 00 3A 00 11 6 58 58

60 07 00 08 00 00 00 3C 00 8 8 60 60

64 07 00 0A 00 02 00 40 00 10 6 64 64

70 07 00 07 00 00 00 46 00 7 8 69 70

72 07 00 09 00 02 00 48 00 9 6 71 72

80 07 00 06 00 00 00 50 00 6 8 80 80

92 07 00 07 00 02 00 5C 00 7 6 91 92

96 07 00 05 00 00 00 60 00 5 8 96 96

108 07 00 06 00 02 00 6C 00 6 6 107 108

120 07 00 04 00 00 00 78 00 4 8 120 120

128 07 00 05 00 02 00 80 00 5 6 128 128

160 07 00 03 00 00 00 A0 00 3 8 160 160

214 07 00 03 00 02 00 D6 00 3 6 213 214

240 07 00 02 00 00 00 F0 00 2 8 240 240

320 07 00 02 00 02 00 40 01 2 6 320 320

480 07 00 01 00 00 00 E0 01 1 8 480 480

640 07 00 01 00 02 00 80 02 1 6 640 640

<Pitch> from b3 to c8

The pitch is shown by the first byte in the sound data (a combination of a pitch [1 byte] + a note [1 byte]).
There are pitches ranging from the lowest b3 (si) to c8 (do) over 4 octaves.
The musical rests are expressed as 56.

do do# re re# mi fa fa# sol sol# la la# si do
8c

c c# d d# e f f# g g# a a# b

3 19

4 1A 1B 1C 1D 1E 1F 20 21 22 23 24 25

5 26 27 28 29 2A 2B 2C 2D 2E 2F 30 31

6 32 33 34 35 36 37 38 39 3A 3B 3C 3D

7 3E 3F 40 41 42 43 44 45 46 47 48 49 4A

rest

56

<Note & rest> ('98.7.20 added )
The data for notes and rests are divided into the first half and the second half.
Each datum in the second half corresponds with each sound (musical symbol), and the sequence of order of the data is the same as sequence of the symbols in the PC utility "Sound creation."

The data in the first half probably show the physical duration of the sound.
I assume that there are two separate tables of the durations of the notes for the two separate series of the tempo speeds, because we can not use a single number and get successful division--the divisible numbers differ among different speeds of the tempos.
As for the discrimination of the two tables, we use the first table if the first byte in the second half of the tempo data is "00," or we use the second table if it is "02."

(1) For the tempos "00 00 XX XX" (30, 32, 34, 38, 40, 44, 48, 54, 60, 70, 80, 96, 120, 160, 240, 480)

1/32 note 1/16 note 1/8 note 1/4 note 1/2 note whole note

first second first second first second first second first second first second

31 00 32 02 34 06 38 0A 30 0E 40 30 11

dotted 33 03 36 07 3C 0B 40 38 0F 40 40 30 12

double-dotted 33 04 37 08 3E 0C 40 3C 10 40 40 40 38 13

triplet 31 01 31 05 33 09 35 0D

1/32 rest 1/16 rest 1/8 rest 1/4 rest 1/2 rest whole rest

first second first second first second first second first second first second

01 64 02 66 04 6A 08 6E 00 72 00 00 75

dotted 03 67 06 6B 0C 6F 00 08 73 00 00 00 76

double-dotted 03 68 07 6C 0E 70 00 0C 74 00 00 00 08 77

triplet 01 65 01 69 03 6D 05 71

(2) For the tempos "02 00 XX XX" (36, 46, 50, 58, 64, 72, 92, 108, 128, 214, 320, 640)

1/32 note 1/16 note 1/8 note 1/4 note 1/2 note whole note

first second first second first second first second first second first second

31 00 31 02 33 06 36 0A 3C 0E 40 38 11

dotted 32 03 34 07 39 0B 40 32 0F 40 40 34 12

double-dotted 33 04 35 08 3A 0C 40 35 10 40 40 3A 13

triplet 31 01 31 05 32 09 34 0D

1/32 rest 1/16 rest 1/8 rest 1/4 rest 1/2 rest whole rest

first second first second first second first second first second first second

01 64 01 66 03 6A 06 6E 0C 72 00 08 75

dotted 02 67 04 6B 09 6F 00 02 73 00 00 04 76

double-dotted 03 68 05 6C 0A 70 00 05 74 00 00 0A 77

triplet 01 65 01 69 02 6D 04 71

Speculating from these, the data in the first half seems to show the rough physical duration of the sound.
The followings are the data in the order of the duration of the sound starting from the shortest duration.

notation tempo (1) tempo (2)

note (1) rest (1) duration (1) note (2) rest (2) duration (2)

triplet 1/32 31 01 1 31 01 1

1/32 31 01 1 31 01 1

triplet 1/16 31 01 1 31 01 1

1/16 32 02 2 31 01 1

triplet 1/8 33 03 3 32 02 2

dotted 1/16 33 03 3 32 02 2

double dotted 1/16 33 03 3 33 03 3

1/8 34 04 4 33 03 3

triplet 1/4 35 05 5 34 04 4

dotted 1/8 36 06 6 34 04 4

double-dotted 1/8 37 07 7 35 05 5

1/4 38 08 8 36 06 6

dotted 1/4 3C 0C 12 39 09 9

double-dotted 1/4 3E 0E 14 3A 0A 10

1/2 30 00 16 3C 0C 12

dotted 1/2 40 38 00 08 22 40 32 00 02 18

double-dotted 1/2 40 3C 00 0C 28 40 35 00 05 21

whole 40 30 00 00 32 40 38 00 08 24

dotted whole 40 40 30 00 00 00 48 40 40 34 00 00 04 36

double-dotted whole 40 40 40 38 00 00 00 08 56 40 40 3A 00 00 0A 42

The data in the lowest 4 bits are meaningful.
"0" is regarded to be 16.
Preceding "X0"s mean 16 to be added. ( X0 X0 X0 08 = 16+16+16+8=56)

If my hypotheses are correct, the shorter notes can not be discriminated against each other; am I right?