Testing frequency response (Avia)

Discussion in 'Archived Threads 2001-2004' started by Oren, May 29, 2002.

  1. Oren

    Oren Stunt Coordinator

    Joined:
    Aug 10, 1999
    Messages:
    178
    Likes Received:
    0
    Last night while doing a frequency sweep (200hz-20hz) test (Avia) on my speakers with an SPL meter, I noticed that there were dips (around 82hz) and peaks (around 110, 90, and 72 hz). (I also noticed my sub only went down to 28hz [​IMG] ) (I have a sub/sat setup.)
    I also noticed that my receiver allows me to adjust the volume in 1 decibel increments, from 100 hz to 10,000 hz in 43 steps. (Unfortunately, below 100 hz it's just a single 'bass' adjustment.)
    I'd like to do a full 20,000 hz - 20 hz test sweep, but the Avia test DVD seems to only offer 200 - 20 hz. So, how can I build a graph like this:
    [​IMG]
     
  2. Ned

    Ned Supporting Actor

    Joined:
    Feb 20, 2000
    Messages:
    838
    Likes Received:
    0
  3. Oren

    Oren Stunt Coordinator

    Joined:
    Aug 10, 1999
    Messages:
    178
    Likes Received:
    0
    Thanks - that helps. I actually just hooked up a laptop directly, which seems to work ok.

    I discovered that my receiver has a 'parametric' equalizer. That is, I can't set the db boost/cut at specific frequencies; rather, I can only give boost/cut at three specific frequences (of my choosing) and the effect drops off to the left and right from there. Weird.

    I suppose I'll do a baseline chart, and then use the adjustments to (try to) raise the troughs or lower the peaks.

    There ought to be a way to plug a microphone into the receiver, place it in the sweet spot, and then have the receiver do all this sort of stuff automatically. It should play various tones, determine how the mic receives it and make any necessary adjustments. Every setup would be custom to the room. Perchance to dream...
     
  4. James Bergeron

    James Bergeron Supporting Actor

    Joined:
    Oct 9, 2001
    Messages:
    831
    Likes Received:
    1
    Thanks for the heads up on the software I'm going to try that!

    What's a good range?

    I was thiknig of makeing a track for
    10hz
    15
    20
    25
    30
    32
    35
    40
    45
    50
    55
    60
    70
    72
    75
    80
    82
    85
    90
    100
    150
    200


    Would that be good enough? I don't really have an EQ or anything I'm just thinking of getting reading for general knowledge of what my system can do, so I can see if there are holes or peaks etc. This can give me an idea of if I need an EQ or not I suppose.
     
  5. Oren

    Oren Stunt Coordinator

    Joined:
    Aug 10, 1999
    Messages:
    178
    Likes Received:
    0
    [​IMG]
    What do you make of that?
    Bass is 100-1000, mid is 500-5000, treble is 1000-10,000. In each range I can specify a particular point and give it boost or cut. It looks like I need to give a boost at 445, 2000, and 8000 hz, and lower the bass a bit.
     
  6. Joseph_W

    Joseph_W Stunt Coordinator

    Joined:
    Mar 23, 2001
    Messages:
    111
    Likes Received:
    0
    Oren, Ned, and James,
    I have a question that has been bugging me when I see graphs like these. In college (mid 70's), we had a speaker design lab that had an expensive pink noise generator (I know you are replacing this with Avia, which makes perfect sence) and a $10,000 microphone. With a cheaper microphone, you were tuning the speaker to fit the microphone, not a flat response. Now they also took great pains with wall treatments which you don't have to worry about, since you are doing the ultimate calibration - one done in the environment the speaker is used.

    Can you now get a fairly flat microphone from 10 to 40Khz without paying a fortune? Sorry to interrupt your thread, but you guys seem to know what you are doing.

    Joe
     
  7. Oren

    Oren Stunt Coordinator

    Joined:
    Aug 10, 1999
    Messages:
    178
    Likes Received:
    0
    Thanks for the compliment, but I don't really know what I'm doing. Although, I can make nifty looking charts in Excel.

    Between Avia, that NCH software, and a $50 Radio Shack SPL, I guess I'm just trying to aim for a flat response (and a neat chart). Seems to me like there are a lot of 7-10 decibel gyrations in my setup, so I'm not sure what's going on.

    I've boosted my EQ by 3 db at 445hz, 10 db at 2000hz, and 5 db at 8000hz. It sounds different but I don't know if it sounds 'better' and I haven't retested yet.

    I know...that doesn't answer your question.
     
  8. Joseph_W

    Joseph_W Stunt Coordinator

    Joined:
    Mar 23, 2001
    Messages:
    111
    Likes Received:
    0
    "I know...that doesn't answer your question."

    Oren,
    I'll take an honest answer all day long!!!

    Good luck. Making it sound better is what it's all about.

    Sorry for the interuption!!

    Joe
     
  9. Ray Stericker

    Joined:
    May 29, 2002
    Messages:
    26
    Likes Received:
    0
    Oren,
    Nice graph! What version of Excel are you using? I tried to do that a while ago, but couldn't get the x axis to be a log scale. I could get the y axis to go logarithmic but not X [​IMG]
    I don't do much graphing in Excel and I only have the 97 version. If its just a menu item I am missing I would appreciate the tip.
    Ray
     
  10. Oren

    Oren Stunt Coordinator

    Joined:
    Aug 10, 1999
    Messages:
    178
    Likes Received:
    0
    I'm using Excel 2000. Once the graph is created, I just click on the X axis, then select the menu format:selected axis. Then click the tab labeled 'scale' and check 'logarithmic.'

    But last night I retested using the software Test Tone Generator 2.6 (another freeware avail. on the net) and I discovered something that throws the whole thing into question: the decibel reading can vary significantly by moving the SPL meter just a few inches one way or another, or angling it differently. Of course, I notice this with my own ears as well. That is, just shifting around or turning my head a bit, significantly changes the perceived volume.

    As a result, the huge gyrations and dips as shown in my graph may not actually be there. If I select the highest reading after moving the SPL meter around a bit or use my 'best guess' after seeing the numbers flip around a bit, the graph is much, much flatter.
     

Share This Page