FFT say sample should be 84 but polyphone wants it to be 48

• NC 5 0

  消息来自 Nikolai Christensen，发表于 …
  So for example this sample below, it gets recommended to be assigned to 48 (which is properly correct).
  But then where are all the 48 peak(s)? Most of them seem much higher than 48.
  I am confused..
  
  "Sample","Peak","Intensity","Frequency","Key","Correction"
  "bagpipe_36",1,1,1045.74,84,1
  "bagpipe_36",2,0.883675,1308.18,88,14
  "bagpipe_36",3,0.845851,915.189,82,32
  "bagpipe_36",4,0.795557,261.098,60,3
  "bagpipe_36",5,0.658877,522.196,72,3
  "bagpipe_36",6,0.620683,391.647,67,2
  "bagpipe_36",7,0.54827,784.64,79,-1
  "bagpipe_36",8,0.309523,130.549,48,3
  "bagpipe_36",9,0.298829,1569.28,91,-1
  "bagpipe_36",10,0.297959,1176.29,86,-2
  "bagpipe_36",11,0.248368,1699.83,92,-40
  "bagpipe_36",12,0.241345,1830.38,94,32
  "bagpipe_36",13,0.228553,1438.73,90,49
  "bagpipe_36",14,0.180459,1049.78,84,-5
  "bagpipe_36",15,0.163673,654.091,76,14
  "bagpipe_36",16,0.161908,1041.7,84,8
  "bagpipe_36",17,0.156557,2092.82,96,0
  "bagpipe_36",18,0.14217,257.06,60,30
  "bagpipe_36",19,0.138127,2223.37,97,-5
  "bagpipe_36",20,0.131071,265.136,60,-23
  "bagpipe_36",21,0.113723,1301.45,88,23
  "bagpipe_36",22,0.11266,1313.57,88,7
  "bagpipe_36",23,0.0875812,1053.81,84,-12
  "bagpipe_36",24,0.0786726,907.114,82,47
  "bagpipe_36",25,0.0756991,528.925,72,-19
  "bagpipe_36",26,0.0731286,1037.66,84,15
  "bagpipe_36",27,0.0724841,790.023,79,-13
  "bagpipe_36",28,0.0684046,253.023,59,-42
  "bagpipe_36",29,0.0682555,1297.42,88,28
  "bagpipe_36",30,0.0656064,1317.6,88,1
  "bagpipe_36",31,0.0641041,1960.93,95,13
  "bagpipe_36",32,0.06288,269.173,60,-49
  "bagpipe_36",33,0.0577617,777.911,79,13
  "bagpipe_36",34,0.0575596,1057.85,84,-19
  "bagpipe_36",35,0.056317,923.264,82,17
  "bagpipe_36",36,0.0563127,3008.01,102,-28
  "bagpipe_36",37,0.0555374,1565.24,91,3
  "bagpipe_36",38,0.0549318,2746.91,101,29
  "bagpipe_36",39,0.0519686,903.076,81,-45
  "bagpipe_36",40,0.0506109,1293.38,88,33
• 消息来自 Nikolai Christensen，发表于 …
  Attached the soundfont
  tuningError.sf2
• 422 17

  消息来自 bottrop，发表于 …
  the soundfont protocol transposes samples by changeing the frequency. (like playing a 33rpm vinyl disk at 45rpm). thats why you need one sample per three keys to obtain a decent acoustic virtual instrument.
  regards bottrop
• NC 5 0

  消息来自 Nikolai Christensen，发表于 …
  bottrop 写于 … - 显示被引用的消息

  the soundfont protocol transposes samples by changeing the frequency. (like playing a 33rpm vinyl disk at 45rpm). thats why you need one sample per three keys to obtain a decent acoustic virtual instrument.
  regards bottrop
  I have not tuned it at all. So its not transposed.
  Polyphone's suggestion of 48 is correct. But the 1.0 peak at 84 is wrong, it should be 48 (strangely the 2 digit seems reversed, not sure if that's a coincidence.)
• 429 1

  消息来自 Davy，发表于 … 1
  The different peaks show the harmonic content of the sound, and this is not the highest peak that gives the frequency of the sound.
  This graph and the peak list (which are based on a Fourier transform ) are here to describe the harmonic content, and are internally not used for finding the root key.
  The root key, 48, is determined by using the autocorrelation  of the signal.
  
  The combination of both tools gives us the following: the root key of this sound is 48 (around 130Hz) and the harmonics are:
  

  • fondamental: 130x1 = 130Hz (line 8 : "bagpipe_36",8,0.309523,130.549,48,3)
  • octave (2nd harmonic): 130x2 = 260Hz
  • fifth (3rd harmonic): 130x3 = 390Hz
  • ...
  • octave of the octave of the octave (8th harmonic): 130*8 = 1040Hz (the strong 84 harmonic)
  • ...

  
  

登录或注册后即可参与讨论。