8 kHz drop-off?? Who's to blame?

Discussion in 'Archived Threads 2001-2004' started by Dennis B, Mar 30, 2002.

  1. Dennis B

    Dennis B Stunt Coordinator

    Joined:
    Nov 1, 2001
    Messages:
    189
    Likes Received:
    0
    I was wondering if anybody knows the actual frequency response of the Marantz SR7200 which I'm using as my pre/pro or how I could measure it without the influence of my room.

    I've done extensive ETF testing on my speakers and room, primarily at low frequencies, and am starting to worry about the mid to high freq band now.

    It's really weird that for all conditions that I've logged so far, the full bandwidth frequency response of the room + speakers starts to roll off at about 8 kHz all the way to 20 kHz in an almost straight line, with an amazing 25 dB difference between these two points.

    I'm using an RMB-1075 as my amp and B&W 602s as my main speakers, and for both the specs are very good all the way up to 20 kHz. For the 7200 I have no such spec, there's nothing in the manual or the web, so that's why I'm asking about it.

    Could this possibly be right?

    Thanks,

    Dennis
     
  2. Pete Mazz

    Pete Mazz Supporting Actor

    Joined:
    May 17, 2000
    Messages:
    761
    Likes Received:
    0
    Are you using a Radio Shack SPL meter as the mic? If so, they are notoriously innaccurate at higher freqs.

    Pete
     
  3. Dennis B

    Dennis B Stunt Coordinator

    Joined:
    Nov 1, 2001
    Messages:
    189
    Likes Received:
    0
    Pete,

    Yes, but I though the ETF calibration file would compensate for this...
     
  4. Rick Guynn

    Rick Guynn Second Unit

    Joined:
    Mar 23, 1999
    Messages:
    473
    Likes Received:
    0
    I'm not sure on the range of the ETF cal file.. you might want to open it up and look at it.. Also do a forum search... There was a thread a few weeks back discussing the cal file and additions to it.

    Rick
     
  5. Pete Mazz

    Pete Mazz Supporting Actor

    Joined:
    May 17, 2000
    Messages:
    761
    Likes Received:
    0
    When I compared the ETF mic and pre amp against the RS SPL meter it wasn't even close, and that was done with and without the cal file. One of these days I'll do a cal file for the DIY mic that I made. It's here if you want to check it out.
    Speakerbuilder.net
    Pete
     
  6. Dennis B

    Dennis B Stunt Coordinator

    Joined:
    Nov 1, 2001
    Messages:
    189
    Likes Received:
    0
  7. Patrick Sun

    Patrick Sun Moderator
    Moderator

    Joined:
    Jun 30, 1999
    Messages:
    38,717
    Likes Received:
    463
    The correction for SPL reading from the RS meter for frequencies over 8KHz are very wrong.

    I compared my RS SPL readings (recorded them) to the values I got with a friend's calibrated mic setup, and for 1/3 octave values, these were the correction values above 8KHz:

    10KHz +8dB

    12.5KHz +16dB

    16KHz +21dB

    20KHz +27dB

    that I came up with that gave be "consistent" RS SPL that agree with the calibrated mic reading.

    If you used what I listed, I doubt you'd be off by more than 3dB of a real measurement setup.

    But as always, YMMV.

    Bear in mind anything over 10KHz will be cymbals and stuff like that. The place to really focus your efforts is in the midrange (500-5000Hz), that'll really make-or-break a speaker's audio quality.
     
  8. Dennis B

    Dennis B Stunt Coordinator

    Joined:
    Nov 1, 2001
    Messages:
    189
    Likes Received:
    0
    Patrick,
    Thanks a lot for this info - so Pete was right, after all [​IMG]. It amazes me that I've done some research on this and you along with Pete are the first ones that I've noticed to mention this huge difference in the higher freqs.
    Would it be asking too much if you could please post more values of that curve for everybody's benefit?
    BTW I've got all freqs between 200 and 8k within 6 dB referenced to what I think is the corrected SPL curve...
    Thanks again.
     
  9. Patrick Sun

    Patrick Sun Moderator
    Moderator

    Joined:
    Jun 30, 1999
    Messages:
    38,717
    Likes Received:
    463
    For quick and dirty measurements, I'd just interpolate the correction values for SPL readings in between each 1/3 octave. I only have an 1/3 octave test tone CD, so I can't offer much more than those correction numbers.
     
  10. Pete Mazz

    Pete Mazz Supporting Actor

    Joined:
    May 17, 2000
    Messages:
    761
    Likes Received:
    0
    If you're handy with a soldering iron, check this mod out for the RS meter.
    Eric Wallen's Site
    I did the low Freq mod, but not the seperate mic mod. It makes the LF readings very close.
    Pete
     
  11. Dennis B

    Dennis B Stunt Coordinator

    Joined:
    Nov 1, 2001
    Messages:
    189
    Likes Received:
    0
    Pete,
    Thanks, this is great info.
    Just a last question, if I understood correcly, the capacitor changes (including C12) correct for both LF and HF, right?
     
  12. Pete Mazz

    Pete Mazz Supporting Actor

    Joined:
    May 17, 2000
    Messages:
    761
    Likes Received:
    0
    Yes, but, according to his info, you have to replace the internal mic of the meter with the external mod. He talks about using the same mic as I posted from the other site (SpeakerBuilder.net).

    Pete
     
  13. Patrick Sun

    Patrick Sun Moderator
    Moderator

    Joined:
    Jun 30, 1999
    Messages:
    38,717
    Likes Received:
    463
    C12 is the change for HF only. You can leave this one alone if you only care about bass. The HF mod does seem like it's better with the external mic mod he suggests.
     
  14. Chris PC

    Chris PC Producer

    Joined:
    May 12, 2001
    Messages:
    3,975
    Likes Received:
    1
    Whats a good mic to use with any software such as ETF using my PC and Sound Blaster PCI 128 sound card and how do I calibrate it? I mean, I've only got speakers in my own room so I'd need to know the mics response before hand, whatever the mic might be, right?

    Although, ultimately, I myself am more concerned with bass and in comparing and not what the ultimate measurements are compared to reference. I just want to measure and find the response and compare with other sub placements etc.

    I guess that in order to truely determine the peaks and valleys in my room, I'd need a calibrated mic, right?
     
  15. Pete Mazz

    Pete Mazz Supporting Actor

    Joined:
    May 17, 2000
    Messages:
    761
    Likes Received:
    0
    If you just want to measure the LF, the RS meter is fine. Just add the corrections to the FR, and you're in business.

    I think there is info on a calibrated mic on Wallen's site.

    Pete
     
  16. Chris PC

    Chris PC Producer

    Joined:
    May 12, 2001
    Messages:
    3,975
    Likes Received:
    1
    Where do I get the simple LF corrections for the RS SPL meter again? There was a link somewhere [​IMG]
    thanx
    So that ETF software is really worth it?
     
  17. Pete Mazz

    Pete Mazz Supporting Actor

    Joined:
    May 17, 2000
    Messages:
    761
    Likes Received:
    0
    Correction chart is Here.
    The ETF software is great for someone(like me) that is forever tweaking things in their system. IMO, it should be used with a calibrated mic, unless you just want to play with LF. You can D/L a trial demo that is fully functional, with the exception of saving files, and limiting single session measurements, I believe.
    Pete
     
  18. BobL

    BobL Agent

    Joined:
    Sep 6, 2001
    Messages:
    32
    Likes Received:
    0
    Here is part of a paragraph from my RTA manual (Goldline DSP30).

    Your sound system has ideal frequency response when all the lights are within one or two divisions from each other (+/-2dB deviation). The lights should form a horizontal flat

    line, give or take one division. You should find that above approximately 8kHz the line will have a downward turn; this is normal in most listening rooms, and should not be

    corrected. You might also find that your sound system does not produce a flat line below 8kHz and that it is rather bumpy, especially in the low frequency region, in the left half of the display. This is often a problem and is due to low frequency resonance in the listening room.

    It goes on further in the manual to state that between 8khz and 16khz there is typically a 3-6db drop in a room. They consider that normal and don't suggest correcting for it. It states that if you do correct it to flat the room will often sound very bright and unnatural. I hope this helps.

    Bob
     
  19. Dennis B

    Dennis B Stunt Coordinator

    Joined:
    Nov 1, 2001
    Messages:
    189
    Likes Received:
    0
    Bob,

    That's a very good point you made.

    I'll try to remember that when messing with the HFs.

    Thanks,

    Dennis
     

Share This Page