Radio Shack SPL correction questioned

Discussion in 'Archived Threads 2001-2004' started by Pieter_L, Jul 27, 2001.

  1. Pieter_L

    Pieter_L Guest

    i finally got a test tone CD (Stryke) & ran my system through it's paces. after recording my measurements using the RS analog SPL meter, i factored in the adjustment for the RS, using the often published table (see below).
    what i question are the adjustments specified for 10khz and above - the table states that the RS meter is essentially flat.
    HOWEVER, my RS measurements drop like a rock starting at 10khz, and checking the response sensitivities of the RS meter in the RS manual (p.8), i saw a similar dramatic drop for 10 - 20khz that matches my results, but contradicts the table below.
    the RS manual graph seems to indicate the correction for "C weighting" at 20khz should be closer to +20 than +1???
    what gives?
    --- here is the table ----
    10Hz +20.5
    12.5Hz +16.5
    16Hz +11.5
    20Hz +7.5
    25Hz +5
    31.5Hz +3
    40Hz +2.5
    50Hz +1.5
    63Hz +1.5
    80Hz +1.5
    100Hz +2
    125Hz +0.5
    160Hz -0.5
    200Hz -0.5
    250Hz +0.5
    315Hz -0.5
    400Hz 0
    500Hz -0.5
    630Hz 0
    800Hz 0
    1KHz 0
    1.25Khz 0
    1.6KHz -0.5
    2Khz -1.5
    2.5Khz -1.5
    3.15Khz -1.5
    4KHz -2
    5KHz -2
    6.3KHz -2
    8KHz -2
    10Khz -1
    12.5KHz +0.5
    16KHz 0
    20KHz +1
     
  2. Patrick Sun

    Patrick Sun Moderator
    Moderator

    Joined:
    Jun 30, 1999
    Messages:
    38,724
    Likes Received:
    464
    You can safely ignore the corrections for over 10KHz when using that corrections chart.
    Here's what one set of readings look like (and I just ignore the reading over 10KHz unless they are totally whacked) when I use the BassZone CD and RS SPL meter:
    My latest FR graph .
    Notice how the high end dies? That's just the RS SPL meter's microphone limitation.
    Nowadays, from 500Hz to 2000Hz, I don't really adjust the values much at all.
    For me, this is what I do for corrections:
    2.5KHz, -1dB
    3.2KHz, -1.5dB
    4KHz-6.5Hz, -2dB
    8KHz, 0db
    and I ignore the rest (but just plot their readings, but knowing it's not as bad as it looks).
    The real meat for speakers is the midrange in the 1KHz-6KHz anyhow. And you'll know when your 10KHz and up frequencies are too zingy, your ears will tell you.
    For stuff under 500Hz, I go with the corrections on the list above.
    ------------------
    PatCave ; HT Pix ; Gear ; DIY Mains ; DIY CC ; Sunosub I + II + III ; DVDs ; LDs
    [Edited last by Patrick Sun on July 27, 2001 at 07:53 AM]
     
  3. Manuel Delaflor

    Manuel Delaflor Supporting Actor

    Joined:
    May 25, 2001
    Messages:
    657
    Likes Received:
    0
    Is there a place to test our beloved spl meter? I mean, a lab or something which can bring us a detailed and realistic resonse chart to make our own corrections file...
    ------------------
     
  4. Pieter_L

    Pieter_L Guest

    whew - glad it wasn't my system that is cruddy!
    looks like the RS correction table should read more like +20dB at 20KHz - using the official correction table and adding the stated 1dB at 20KHz will give a very inaccurate reading.
    strange that this official, but wrong (>10KHz) table has not been challenged before??
     
  5. ThomasW

    ThomasW Cinematographer

    Joined:
    Nov 6, 1999
    Messages:
    2,282
    Likes Received:
    0
    Manuel
    Sure you can send your RS meter off to a lab and get it calibrated. They compare it against a known standard, provide a correction curve, and usually a floppy with the correction data ready to load into a computer .
    The cost generally runs about $300-$500......
    FYI, the little electret capsule in the RS meter costs less than $2.00ea retail in small lots (so you can imagine what RS is paying. They are fairly accurate using the standard correction tables. But the HF and LF do roll off.
    [Edited last by ThomasW on July 28, 2001 at 09:32 AM]
     
  6. Jones_Rush

    Jones_Rush Stunt Coordinator

    Joined:
    Jan 19, 2001
    Messages:
    198
    Likes Received:
    0
    If there is a place where you can rent high-end SPL meters, you can do the Radio Shack SPL meter calibration by yourself for a lot less money than $300-500 (though it might not be that accurate).
     
  7. Steve Kinkead

    Steve Kinkead Auditioning

    Joined:
    Aug 16, 2000
    Messages:
    12
    Likes Received:
    0
    You can purchase a calibrated mic for $250 from http://www.etfacoustic.com/.
    I purchased the RTA software from them but have not personally tried the mic.
    Steve
     
  8. Greg Monfort

    Greg Monfort Supporting Actor

    Joined:
    May 30, 2000
    Messages:
    884
    Likes Received:
    0
    I have a 1st generation RS meter #42-3019 and don't know if it applies to the later versions, but compared to a calibrated B&K:
    10 +20.5
    12.5 +16.5
    16 +11.5
    20 +6.2
    25 +4.4
    31.5 +3
    40 +2
    50 +1.3
    63 +0.8
    80 +0.5
    100 +0.3
    125 +0.2
    160 +0.1
    200 0
    250 0
    315 0
    400 0
    500 0
    630 0
    800 0
    1000 0
    1250 0
    1600 +0.1
    2000 +0.2
    2500 +0.3
    3150 +0.5
    4000 +0.8
    5000 +1.3
    6300 +2
    8000 +3
    10000 +4.4
    12500 +6.2
    16000 +8.5
    20000 +11.2
    GM
    ------------------
    Loud is beautiful, if it's clean
     
  9. Manuel Delaflor

    Manuel Delaflor Supporting Actor

    Joined:
    May 25, 2001
    Messages:
    657
    Likes Received:
    0
    Greg that seems fine, at least by intuition! I have a newer model (33-2005) and also intrigued to know if your calibrations apply on my model...
    A quick question, do I have to invert +/- signs in the calibration file? I want to try yours.. it can't be worst than the "official" one...
    HMM I wonder... WHO says is official???
    ------------------
     
  10. Richard Greene

    Richard Greene Stunt Coordinator

    Joined:
    Mar 5, 2001
    Messages:
    148
    Likes Received:
    0
    Greg, what you posted, with the exception of the frequencies below 20Hz., are the exact conversion factors to convert C-Weighting to U-Weighting.
    If a Rat Shack meter measured perfectly using C-Weighting and another sound meter measured perfectly using U-Weighting, the conversion factors you posted (above 20Hz.)
    would be absolutely correct ... what you have posted IS a rough approximation of correction factors for an "average' Rat Shack meter -- NOT correction factors for a specific Rat Shack meter, as you suggest -- the Rat Shack meter is not that good -- its specifications are +/- 2dB and it has a peak in the treble from 2-8kHz followed by sharp roll-off above 10kHz. You can use the standard C-Weighting to U-Weighting conversion factors you posted up to about 2kHz. and be pretty close -- I've done just that for many years. I had my own Rat Shack meter compared with a very expensive meter many years ago and it was accurate to within it's specifications of +/- 2dB up to about 2kHz. -- that's close enough for me.
    The treble peak makes me advise others to avoid using the Rat Shack meter for frequencies above 2kHz. -- Even with a perfect sound meter, treble should always be evaluated by ear because high frequency hearing ability and treble roll-off preferences vary so much among audiophiles (Most people would describe a treble frequency response that measured flat as "too bright".)
     
  11. Greg Monfort

    Greg Monfort Supporting Actor

    Joined:
    May 30, 2000
    Messages:
    884
    Likes Received:
    0
    >A quick question, do I have to invert +/- signs in the calibration file? I want to try yours.. it can't be worst than the "official" one...
    >HMM I wonder... WHO says is official???
    ====
    No, you just add these values to whatever the meter reads. WRT the other chart, I first saw it a couple of years ago on the Basslist and it seems to have good credentials. I just noticed that he said he pointed the meter directly at the speaker, while mine was done per the booklet's recommendation that the mike be at 90deg to the source. Not having made any comparisons, I don't don't know how this may affect the results.
    Here's the post:
    ----------
    First, I would like to thank Eric Busch (The DUMAX technician)
    and Dave Clark from DLC Design for the low frequency data,
    Tom Nousaine for supervising the calibration, Bill Duddleston
    at Legacy Sound in Springfield, IL for letting PSACS meet at
    his business, and Paul Barton from PSB for furnishing the
    loudspeaker calibrated at the NRC anechoic chamber in Canada.
    The Radio Shack meter is a wonderful unit. About a two and a half years
    ago, I did a calibration curve for the RS meter using lab equipment,
    with
    Eric Busch from DLC Design adding the low bass
    down to 10 Hz. This was published in PSACS Sound Bytes in two issues.
    These are the corrections that should be added to the meter readout in
    order to achieve the correct SPL. These corrections are only valid
    for the meter set to C weighting, using 1/3 octave pink noise (easily
    available from various CDs), with the mic pointed at the speaker.
    Both my analog meters and my digital meter measured the same
    in October, 1996.
    These are corrections, they are to be added to the meter readout for the
    correct response in dB SPL.
    10Hz +20.5
    12.5Hz +16.5
    16Hz +11.5
    20Hz +7.5
    25Hz +5
    31.5Hz +3
    40Hz +2.5
    50Hz +1.5
    63Hz +1.5
    80Hz +1.5
    100Hz +2
    125Hz +0.5
    160Hz -0.5
    200Hz -0.5
    250Hz +0.5
    315Hz -0.5
    400Hz 0
    500Hz -0.5
    630Hz 0
    800Hz 0
    1KHz 0
    1.25Khz 0
    1.6KHz -0.5
    2Khz -1.5
    2.5Khz -1.5
    3.15Khz -1.5
    4KHz -2
    5KHz -2
    6.3KHz -2
    8KHz -2
    10Khz -1
    12.5KHz +0.5
    16KHz 0
    20KHz +1
    To measure the in-room response of your speakers with the SLM,
    find a quiet and undistracted time, obtain a 1/3 octave pink noise CD, mic
    stand, and worksheet. With the speakers in their normal positions and
    using a mic stand for the SLM (I've had good results just holding the meter
    too, pink noise is forgiving), place its microphone where your ear would be
    at your favorite position or "sweet spot". Set the meter to "C" and
    "slow". Play the 1 Khz, 1/3 octave pink band and set the level on the amp
    or preamp, and the meter range, so the meter reads 80 dB at 0 dB on the
    meter. Higher levels might cause driver compression in the frequency
    extremes, rolling off the response. Now go back to the
    first ISO center 1/3 octave band on your CD (25 Hz on mine) and record
    the response: 3.5, or -1.0 , or -5.0, or whatever
    it is. Now step through the pink noise
    bands, recording the meter level each time. This is the raw data.
    Keeping the meter in the same measuring position, using an accurate
    CD, having a quiet room, repeating the measurements for accuracy,
    understanding what you are measuring, fresh batteries, not talking
    while measuring, having your meter calibrated for overall level
    (relative level comparison is unaffected), and doing only one
    speaker at a time if possible to avoid comb
    filtering (variations of plus or minus 2 dB are possible) are all
    important factors. Sometimes if I have
    to measure two speakers at a time, I move the meter around in a figure
    eight pattern, slowly, and try to obtain an average reading, as the
    microphone moves in and out of the combing peaks and troughs. Best
    accuracy would be obtained from taking several measurements at different
    listening positions, if one has the time and patience.
    Now take the raw data and make the corrections on the work sheet,
    entering the new values in the appropriate column. Time to either moan
    or marvel, since this is the actual in-room 1/3 octave
    pink noise frequency response of your speaker/system at that listening
    position. Plotting the results on graph paper in different colors for
    left, right, center, etc. makes it look cool.
    Yes, this way is tedious, but it is very inexpensive. And very
    accurate. For the third octave pink noise, I use Carver's Amazing
    Bytes CD, GRP Z-9907; other CD's with 1/3
    octave pink noise ISO centers are:
    My Disc, Sheffield 10045-2-T
    Sheffield/Coustic Test and Demonstration Disc, Sheffield 10040-2
    Autosound 2000 CD #103, $18, 800-795-1830
    IASCA Setup and Test CD
    HI-FI News and Record Review "CD-II", $30
    Japan Audio Society Audio Test CD-1, YDDS-2
    --- these last two available from DB Systems, 603-899-6415.
    For the others: http://www.audioxpress.com/ http://www.mcmelectronics.com/main.html http://www.parts-express.com/
    There are other CD's that have warble tones on them, but I am not
    comfortable using them. I also extensively use track 19, disc 2,
    from Delos "Surround Spectacular", DE 3179.
    This has a slow sine sweep from 160Hz to 20Hz with the voice of
    David Ranada announcing the frequencies as they
    descend. A fantastic help when used with the Radio Shack SLM, you can
    easily tell where the room peaks and dips are.
    http://www.delosmus.com/
    You have permission to copy and distribute this information freely, as
    long as no commercial gain is involved.
    Radio Shack is your friend.
    Here is how I did my calibration:
    The $34.99 Radio Shack 33-2050 analog sound level meter has been
    around for over 25 years. Its predecessor, the 33-1028, was reviewed
    favorably in Stereo Review, (Julian Hirsch, "Equipment Test Reports",
    Stereo
    Review, August 1972). It has a much different curve than the ones I
    tested.
    To verify the accuracy of the newer version, I compared it to an
    Audio Control 3050 RTA, the same one Tom Nousaine used for years in his
    test reports for Car Stereo Review until he bought MLSSA. The overall
    SPL accuracy of all three of my Radio Shack sound level meters -- 2
    analog
    and 1 digital 33- 2055-- were within 1 dB of 75 dBC SPL compared to the
    Audio Control. I then checked the frequency response, comparing it
    while
    set to C weighting and slow, with pink noise, 1/3 octave band by 1/3
    octave
    band, to the Audio Control RTA in the SPL mode. Using the same official
    PSACS calibrated PSB loudspeaker and a pink noise CD, I made a
    calibration curve that can be subtracted from the results obtained by
    the
    Radio Shack in your living room to obtain accurate, repeatable
    measurements
    for about $60, including pink noise CD. Make sure your meter is set to
    C
    weighting. The digital meter and my second analog meter (6 years
    newer than the test unit in 1996) were the same as the test SLM.
    Response below 25 Hz done by Eric Busch with sine waves and
    B&K equipment at Dave Clark's DLC Design in Michigan.
    Michael Sims
    Prairie State Audio Construction Society
    ---------
    ====
    >Greg, what you posted, with the exception of the frequencies below 20Hz., are the exact conversion factors to convert C-Weighting to U-Weighting.
    ====
    What's U-weighting, and when is it used? I'm only familiar with A, B, and C.
    ====
    >If a Rat Shack meter measured perfectly using C-Weighting and another sound meter measured perfectly using U-Weighting, the conversion factors you posted (above 20Hz.)
    would be absolutely correct ... what you have posted IS a rough approximation of correction factors for an "average' Rat Shack meter -- NOT correction factors for a specific Rat Shack meter, as you suggest --
    ====
    Gee, I asked the technician at the plant where I worked at the time to compare mine to the calibrated one, but who knows, he may have had a drawer full that he'd already done, so gave me an 'average' correction curve. Not! If it matches any others, then I guess they were at least consistant.
    ====
    >the Rat Shack meter is not that good -- its specifications are +/- 2dB and it has a peak in the treble from 2-8kHz followed by sharp roll-off above 10kHz.
    ====
    In the booklet that came with mine, it claims to be flat from 35Hz-7kHz (+/-3dB), -4dB/10kHz, and -7dB/20kHz. As I noted, mine is relatively ancient, and I don't know how it compares to later units. What you're indicating implies that it doesn't and I shouldn't publish it anymore, lest it confuse more than help.
    WRT MS's corrections, the criteria for how it was determined should be published with it in the future IMO.
    GM
    ------------------
    Loud is beautiful, if it's clean
     
  12. Patrick Sun

    Patrick Sun Moderator
    Moderator

    Joined:
    Jun 30, 1999
    Messages:
    38,724
    Likes Received:
    464
    This is what I came up with as corrections above 2KHz for my own particular RS SPL meter as compared with the frequency response from a LAud set of measurements:
    Frequency dB
    2.0KHz 0
    2.5KHz 0
    3.2KHz -3
    4.0KHz -6
    5.1KHz -7
    6.4KHz -6
    8.1KHz -1
    10.2KHz +7
    12.9KHz +13
    16.2KHz +18
    20.0KHz +25
    This is for my meter only (but will be pretty representative of other RS meters - just don't use it as the gospel for meter correction values.)
    ------------------
    PatCave; HT Pix; Gear; DIY Mains; DIY CC; Sunosub I + II + III; DVDs; LDs
     
  13. Richard Greene

    Richard Greene Stunt Coordinator

    Joined:
    Mar 5, 2001
    Messages:
    148
    Likes Received:
    0
    I probably was not clear in my post, as usual, so I'll try again ... and probably be even less clear:
    If you go to the website shown below, there is a comparison of A, C and U Weighting (go to section 8.2)-- you should note all the numbers you posted, except for the 10, 12.5 and 16Hz. correction factors, are lifted directly from this A vs. C vs. U-Weighting table. The table didn't specify it was a list of correction factors for Greg M,'s Rat Shack meter ... or maybe I missed a small footnote somewhere!
    wwwcampanellaacoustics.com/faq.htm
    On that webs site you will find a comparison of A-Weighting
    (adjustments to correlate with hearing ability at low volumes) C-Weighting (adjustments to correlate with hearing ability at high volumes) and U-Weighting (unweighted measurement = no adjustments)
    I think there are several different versions of the Rat Shack analog meter because mine has an instruction booklet with different specifications than yours! After one glance at the frequency response curve that was in my Rat Shack meter's owners manual, I knew I'd ignore the treble measurements above 2kHz.
    The correction factors posted by Michael Sims at high frequencies do not correlate with my own meter that seems to be "hard of hearing" above 10,000 Hz. with every speaker I've measured over the past two decades.
     

Share This Page