Home of the Squeezebox™ & Transporter® network music players.
Page 1 of 3 123 LastLast
Results 1 to 10 of 23
  1. #1
    Senior Member
    Join Date
    Feb 2007
    Location
    Olde London Town
    Posts
    1,110

    Hi-frequency content of 24/96 files?

    So, I used Adobe Audition 3 to perform a frequency analysis of a 24-bit/96Khz 'Studio Master' file from Linn Records (it's the 'Amen' from the Mozart Requiem). The FLAC was converted to WAV by foobar2000 before loading into Audition BTW.

    I'm wondering why there's a bunch of high frequency material after 30000 then it drops away? Perhaps this isn't a true picture and something in the conversion and analysis isn't right?
    Last edited by amcluesent; 2008-07-01 at 10:59.

  2. #2
    That's interesting... This is what SND said about B.Britten: Simple Symphony, Op. 4, downloaded from http://www.2l.no/hires/index.html (Stereo WAV 24/96):


    Can't make bbcode work. This is the link to the picture: http://i263.photobucket.com/albums/i..._stereowav.png
    Last edited by alekz; 2008-07-01 at 11:59.

  3. #3
    Senior Member Pale Blue Ego's Avatar
    Join Date
    Apr 2005
    Location
    North Florida
    Posts
    1,261
    Probably a limitation of the microphones.

  4. #4
    Senior Member pfarrell's Avatar
    Join Date
    Apr 2005
    Location
    Wayne, PA
    Posts
    4,251
    For example, here is the Neumann site's spec sheet on the U87 mic, a classic "go to mic" for vocals

    http://www.neumann.com/zoom.php?zoom...am&w=878&h=278

    The classic Neumann M50 is spec'd at "Frequency response: 40 - 16 000 cps"
    meaning 40 hZ to 16kHz in modern language

  5. #5
    Founder, Slim Devices seanadams's Avatar
    Join Date
    Apr 2005
    Posts
    2,879
    The energy _must_ go to zero before reaching 48KHz because that is the nyquist frequency. Note the log scale - that is exactly what happens, although it may not look it at first glance.

    But that is a very sharp drop indeed, and it's hard to say why it rises so quickly after 22KHz before hitting that brick wall. I would think that a "good" 24/96 recording should have just a small amount of energy above 22KHz, tapering off gradually before hitting 48. The sharp drop suggests it may not have been LP filtered anywhere in he process.

    I'd say it's probable that this material was not mastered properly, perhaps due to 44.1 processing/equipment being used in the process. For example, older ∆∑ DACs generate quite a bit of noise above 22KHz unless it is properly filtered in the analog domain. Perhaps some such equipment was in the chain. Or perhaps an poorly implemented ADC stage captured some aliasing. I'm just speculating - it's not really my area of expertise, and there are too many possibilities because so much equipment designed to assume a 22KHz filter at playback.

    I think it would be interesting to look at other files, and smaller excerpts.

  6. #6
    Senior Member pfarrell's Avatar
    Join Date
    Apr 2005
    Location
    Wayne, PA
    Posts
    4,251

    Hi-frequency content of 24/96 files?

    seanadams wrote:
    > The energy _must_ go to zero before reaching 48KHz because that is the
    > nyquist frequency. Note the log scale - that is exactly what happens,
    > although it may not look it at first glance.


    True, that doesn't explain the relative peak from 30kHz to 40k or so.
    Note, its down 35+ db, from the peak, so its pretty minor no matter what
    it is.

    > I'd say it's probable that this material was not mastered properly,
    > perhaps due to 44.1 processing/equipment being used in the process.


    Kind of defeats the purpose of 96kHz marketing if they used 44.1k or 48k


    --
    Pat Farrell
    http://www.pfarrell.com/


  7. #7
    Senior Member pfarrell's Avatar
    Join Date
    Apr 2005
    Location
    Wayne, PA
    Posts
    4,251
    Quote Originally Posted by alekz View Post
    This shows all the over 20kHz stuff down 90 to 100 dB.
    That means it is realistically non-existent.

    Typically, "down 70 dB" means "cut out completely" since humans only have about 90dB of range. Plus a quiet living room is typically 35dB, and threshold of pain is about 96dB above that.

  8. #8
    Senior Member
    Join Date
    Feb 2007
    Location
    Olde London Town
    Posts
    1,110
    So, a few more hi-res files analysed by Adobe Edition 3.

    These tracks were taken from (in order) -

    1) Linn Records, 24/88 Studio Master of Mozart Symphonies

    2) The Doors, The Doors 24/96

    3) Philip Glass, Koyaanisqatsi, so-called 24/96!

    No wonder some people are saying they can't hear any difference over 16/44.1!
    Last edited by amcluesent; 2008-07-02 at 01:46.

  9. #9
    Senior Member mudlark's Avatar
    Join Date
    Aug 2006
    Location
    Cumbria, England
    Posts
    736
    I feel that many people forget the masking effect of frequencies at higher levels than these "audiophile" frequencies above 20Khz. In the general course of a piece of music there are few occasions when there is space in the overall sound to hear high frequency material. If the high frequencies are more than 20dB down from the main noise level it is unlikely that you would hear the high frequencies at all.

    I would be interested to see copies of research about high frequency event audiability in the run of a typical music program.
    Transporter or Cyrus streamX>CyrusDACXP>ESPAudio P09B Active filter>Cyrus X x 2>Rhapsody, Avondale and Naim cable, Kubuntu 16.10 server, various boxes for storage.
    SB3 Flycatcher 3A linear power supply. Touch also.
    Using SqueezeBoxServer (LMS) 7.9
    Also piCorePlayer>Rega DAC>B4>Avondale 260>Royd Eden Kubuntu 14.10
    also naim 32.5>hicap>140>Akroyd Coniston

  10. #10
    Quote Originally Posted by pfarrell View Post
    This shows all the over 20kHz stuff down 90 to 100 dB.
    That means it is realistically non-existent.

    Typically, "down 70 dB" means "cut out completely" since humans only have about 90dB of range. Plus a quiet living room is typically 35dB, and threshold of pain is about 96dB above that.
    I agree, my point was that 2L recordings do not have any sharp "jumps" or "drops" in their spectrum. They do look (and sound) like honest 24/96. Not that many are available, though...

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •