House curve - Reference Levels - Calibration

Discussion in 'Archived Threads 2001-2004' started by Ned, Apr 16, 2002.

  1. Ned

    Ned Supporting Actor

    Joined:
    Feb 20, 2000
    Messages:
    838
    Likes Received:
    0
    This idea of the house curve for equalizing the 20-80hz range brings up a few questions for me.

    First off, +6-8db at 20-32hz is quite aggressive. I tried it today using my ART351 EQ by adding

    20hz +6db

    25hz +6db

    31.5hz +6db

    40hz +4.5db

    50hz +3db

    63hz ++2db

    80hz +1db

    The THX Phantom Menace trailer was hitting peaks of 110db and the red clip light was flickering (first time I've seen it flash).

    Now, using this "house curve", do we want to be calibrating without the extra "hot" amount (2-5db on LFE)that we typically add? Perhaps even a bit under considering that the rat shack meter readings are incorrect in the lowest few octaves?

    What about listening levels? If we always listen at 10db below reference, what sort of setup do we want? What about if we listen at full reference?

    I'd like to add that the SVS ISD upgrade has really pushed my 20-39CS to a much higher level of performance. With the old driver this intro was just a bunch of clacking or else it wasn't loud enough to get any impact. Now it incites giddy/demonic laughter when the LFE hits.
     
  2. Harold_C

    Harold_C Stunt Coordinator

    Joined:
    Apr 1, 2002
    Messages:
    198
    Likes Received:
    0
     
  3. Ned

    Ned Supporting Actor

    Joined:
    Feb 20, 2000
    Messages:
    838
    Likes Received:
    0
    Well I run a subsonic filter under 20hz and the driver hasn't reached its limit yet. The discussion is about "house curve" and how it relates to reference levels and calibration.

    So what's your point?

    "150 clean watts" is rather arbitrary. For each speaker? What sensitivity must the speakers have? How big a room? What type of room acoustics?

    I'm well aware of the new demands created by equalization (hence why the clip light finally flashed).

    This intro is a great test because it has a very sudden high peak of low frequency bass at full intensity followed by a sustained note of low bass at similarly high intensity.

    Anyway, get back on topic.
     
  4. Wayne A. Pflughaupt

    Joined:
    Aug 5, 1999
    Messages:
    6,100
    Likes Received:
    33
    Location:
    Katy, TX
    Real Name:
    Wayne
    Harold,

    Please keep in mind that the actual demand on the sub’s amplifier depends as much on its gain setting as well it does any applied equalization. For instance, if Ned found all that EQ boost created a need to turn the sub down, then there is not a demand for an additional 350 watts of output.

    Also, boosting a specific EQ band demands more power at and around that frequency. Most amps can deliver higher peak power at specific frequencies than their RMS (broad-band) power ratings indicate.

    Ned,

    Harold does have a valid point that any kind of equalizing places demands on a sub system, both to amplifier and driver. The only way you can effectively equalize is to have plenty of headroom going in.

    It’s hard to determine if your house curve is on-track without knowing your pre-EQ readings or having some idea of the size of your listening area. You’ve dialed in what appears to be a relatively small 6dB house curve, which is sufficient only if you have a very large room.

    You really need to base your equalizing on some in-room response data; otherwise anything you do is arbitrary or “seat of the pants” at best. For example, from the EQ values you’ve posted it appears you had ruler flat response before dialing in your house curve – highly unlikely.

    In addition, note that all your sliders are set to boost. This is poor “EQ etiquette.” The reason: Although it would seem (by looking at the slider settings) that you have dialed in a smooth boost that levels off at 31Hz, that’s not the way a graphic equalizer actually works. In reality response sags between the sliders, giving a “comb-filter” or “saw-tooth” effect. The sag becomes more pronounced the further you boost or cut.

    In your situation, Ned (again assuming the curve you’ve dialed in is what is actually needed) it would be better to split the overall curve above and below the “0dB” line. In other words, 31Hz would be set at +2.5dB and 80Hz at –2.5dB. This would retain the 5dB differential and give functionally smoother response by minimizing between-filter sag. The loss of overall output could be regained at the sub amp volume control.

    Regarding use of the RS meter correction values, that depends on whose advice you want to listen to. Some here on the Forum (most notably Richard Green) don’t use them, setting up measured flat, uncorrected response and relying on the meter’s low frequency shortcomings to achieve the house curve. This worked well for Richard since he has stated that he didn’t feel the need to start his house curve until 50Hz (the meter’s deficiencies don’t start getting drastic until about that point), or the need to level off response at about 30Hz. And apparently the meter’s response deviations below 50Hz were perfect for him. That certainly won’t be the case for everyone.

    My room needed the curve to start at 100Hz and level off at 30Hz, so I needed to use corrected readings.

    Richard and I both agree that whichever method you use (or perhaps even a combination of the two) it is best to let you ears be the final determinant of how much curve to dial in and where it starts and stops.

    Regarding listening levels, I finalized my house curve using music (which most people listen to at lower levels than their movies). Then you can set your sub level for movies by ear or by calibration disc.

    Regards,

    Wayne A. Pflughaupt
     
  5. Harold_C

    Harold_C Stunt Coordinator

    Joined:
    Apr 1, 2002
    Messages:
    198
    Likes Received:
    0
     
  6. Ned

    Ned Supporting Actor

    Joined:
    Feb 20, 2000
    Messages:
    838
    Likes Received:
    0
    I measured my in-room response several months ago and after making the RS meter corrections it was almost ruler flat, at least in my preferred listening position (who cares about the rest [​IMG] ). I will definitely use a much less aggressive amount of boost to keep the amp from clipping. Today's setup was just a quick testing of the full "house curve" idea. My room is small and fully contained so I probably have more freedom to play with the EQ than most.
    As good as the bass response is, I'm still going to build a clone SVS with my leftover driver. There are certain LFE's where I can tell that the driver is nearing it's limits again. It's not clacking like the old one but the control is getting a bit "loose". 110db peaks are nothing to snear at though. Very impressive.
     
  7. Ned

    Ned Supporting Actor

    Joined:
    Feb 20, 2000
    Messages:
    838
    Likes Received:
    0
    I would guess the SVS are more efficient than 87db/1w/1m though. Maybe Tom can give us a rough idea.
     
  8. Harold_C

    Harold_C Stunt Coordinator

    Joined:
    Apr 1, 2002
    Messages:
    198
    Likes Received:
    0
     
  9. ColinM

    ColinM Cinematographer

    Joined:
    Dec 9, 2001
    Messages:
    2,050
    Likes Received:
    0
    About 3 months ago Tom V. extrapolated 98dB/W from Tom N.'s measurements with the 'prototype' on hand for testing.

    He also said not to ever quote him on that, as it was a rough non-scientific number...

    I am a House Curve user, I admit it.
     
  10. Richard Greene

    Richard Greene Stunt Coordinator

    Joined:
    Mar 5, 2001
    Messages:
    148
    Likes Received:
    0
    Comments to Wayne and Harold C. follow:

    To Wayne who wrote:

    "Some here on the Forum (most notably Richard Greene) don’t use them, setting up measured flat, uncorrected response and relying on the meter’s low frequency shortcomings to achieve the house curve. This worked well for Richard since he has stated that he didn’t feel the need to start his house curve until 50Hz (the meter’s deficiencies don’t start getting drastic until about that point."

    *****************RG replies:************************

    I use the "house curve" called "C-weighting" because it works well for my ears in my room. Radio Shack meter "shortcomings" are quite small between 20Hz.

    and 80Hz., which is the primary frequency range for my subwoofer. I believe what you call "shortcomings" are mainly the standard adjustments from C-weighting to U-weighting (actual unweighted SPL).

    I believe Radio Shack meters are within their rated

    +/- 2dB frequency response from 20Hz. to at least 2kHz.

    My own meter reads about 4dB too high in the treble in the 4kHz. to 8kHz. range and then has a very sharp roll-off above 10kHz. based on a comparison with an expensive meter (I wish I had compared both meters in the bass frequencies but at the time I was only trying to understand the treble errors). My Owners Manual also shows both of these treble errors in a frequency response chart. My meter's treble errors do not match the often posted RS meter corrections in the treble frequencies (done for someone else's meter)

    So I never use my meter for measurements over 2kHz. becauses accuracy is poor at those frequencies. Even a perfect sound meter would not be very useful above 2kHz. because hearing ability varies so much in the treble frequencies along with preferences (what one person calls "detailed" sound another person may call

    "bright" sound).

    RS meter bass frequencies accuracy is good down to 20Hz.

    Below is a comparison of "correction (misleading term) factors" to convert SPL readings from C-weighting to

    U-weighting (unweighted) for a theoretically perfect sound meter compared with the "correction factors" for a specific RS meter often posted on the internet. The differences between these two correction factors are the actual RS meter "errors" (beyond the standard adjustment from

    C-weighting to U-weighting):

    SPL Adjustments from C-weighting to U-weighting:

    ....."Perfect Meter".....Sample RS Meter .... RS error

    10Hz.....+14.3dB............+20.5dB...........(-6.2dB)

    12.5.....+11.2dB............+16.5dB...........(-5.3dB)

    16Hz......+8.5dB............+11.5dB...........(-3.0dB)

    20Hz......+6.2dB.............+7.5dB...........(-1.3dB)

    25Hz......+4.4dB........... .+5.0dB...........(-0.6dB)

    31.5Hz....+3.0dB.............+3.0dB...........(-0.0dB)

    40Hz......+2.0dB.............+2.5dB...........(-0.5dB)

    50Hz......+1.3dB.............+1.5dB...........(-0.2dB)

    63Hz......+0.8dB.............+1.5dB...........(-0.7dB)

    80Hz......+0.5dB.............+1.5dB...........(-1.0dB)

    100Hz.....+0.3dB.............+2.0dB...........(-1.7dB)

    *****************RG replies:************************

    To Harold C.

    You mentioned JBL drivers -- I believe the largest XMAX

    for any JBL driver is 9.5mm for the 2245 18" subwoofer driver. So the benefits of the larger cone (about 50%

    more area than a 15" cone and 150% more area than a

    12" cone) are partially offset by the relatively small

    (by today's standards) 9.5mm xmax. My own Adire Audio

    15" Tempest driver with 16.4mm DUMAX has a slightly

    larger displacement (Vd) than a JBL 2245.

    I fully agree on the dangers of boosting very low frequencies with an equalizer, especially with ported

    subwoofers. I also recommend people with ported subwoofers use a 20Hz. high-pass filter with an 18 or 24dB/octave slope if there is no other over-excursion protection.

    If you had experience using a parametric equalizer as a notch filter to CUT bass bass output at specific room mode frequencies, I think you would be a big fan of subwoofer equalizers, as I have been for about 20 years.

    To summarize my experiences with subwoofer equalization:

    In most listening rooms I would expect a relatively inexpensive subwoofer (let's say $1000), when used with a parametric equalizer, to outperform (bass quality, not quantity) any other unequalized subwoofer at any price.

    That's how important room acoustics are to bass frequency response in most home listening rooms. Without parametric equalization, a subwoofer is horrible in my own listening room with frequency response deviations of +/-9dB before equalization including a huge and very audible peak between 40 and 55Hz. (+/- 4.5dB after applying only two bands of parametric equalization, with no peak at all between

    40 and 55Hz.)

    I prefer listening at 75-80dBA with my subwoofer about

    3 to 6dB louder (so my own subwoofer rarely goes above

    85DbC). Most people I know listen about 5dB louder than

    I do. So it seems to me that most people do not need to reach Dolby reference levels (and most home subwoofers won't reach Dolby reference levels, so it's a moot point).

    If you use a 150wpc Audio Research amplifier for your subwoofer, you're a lucky (or rich, or both) guy.

    I also use a 150wpc basic amplifier for my DIY sealed

    low Qtc DVC subwoofer ... but it is an Onkyo M501 basic amplifier that cost me about $300 on sale.

    Just wanted you to know what us po' folks use!
     
  11. Harold_C

    Harold_C Stunt Coordinator

    Joined:
    Apr 1, 2002
    Messages:
    198
    Likes Received:
    0
     
  12. Harold_C

    Harold_C Stunt Coordinator

    Joined:
    Apr 1, 2002
    Messages:
    198
    Likes Received:
    0
     
  13. ColinM

    ColinM Cinematographer

    Joined:
    Dec 9, 2001
    Messages:
    2,050
    Likes Received:
    0
    That number WAS derived including room reinforcement, FWIW.

    PS - I also stand there with the refrigerator door open as I glug milk straight from the carton. (For my mother...)

    I was running the 20*39 un-eq'd with an old 40wpc into 8 HK amp, getting 100db's + from 8 feet away. No problem. Now I cut 80/63 and boost 31.5/25/20 with 250w, and I don't worry about a thing. No bottoming with the old driver, no smoke spewing forth from thy amp.

    All of this may make an audiophile gasp, and I understand why. I don't think EQ's should be used by inexperienced folks. But used responsibly and within reason, they can make a very VERY modest system sound like something 5 times the cost.

    I calibrate with VE, sub at 74/76, rest at 75.
     
  14. Wayne A. Pflughaupt

    Joined:
    Aug 5, 1999
    Messages:
    6,100
    Likes Received:
    33
    Location:
    Katy, TX
    Real Name:
    Wayne
     
  15. Richard Greene

    Richard Greene Stunt Coordinator

    Joined:
    Mar 5, 2001
    Messages:
    148
    Likes Received:
    0
    Acoustics FAQ:
    http://www.campanellaacoustics.com/faq.htm#basic_dba
    see section 8 for A vs. C vs. U weighting comparison
    or see a partial excerpt below that I've typed in a way
    that I hope will be readable after I post:
    Weighting = Purpose:
    A-weighting = check for potential to damage hearing
    C-weighting = measure subjective loudness for average human ears at normal listening levels
    U- weighting = actual sound pressure level with no adjustments (microphone+amplifier attached to a volt meter with no equalization)
    A, C and U weighting comparison:
    Nominal......"A".........."C"........."U"
    Frequency...Weight... Weight..Weight.
    10Hz........-70.4dB...-14.3dB...0.0dB
    12.5........-63.4.....-11.2.....0.0
    16..........-56.7.....- 8.5.....0.0
    20..........-50.5.....- 6.2.....0.0
    25..........-44.7.....- 4.4.....0.0
    31.5........-39.4.....- 3.0.....0.0
    40..........-34.6.....- 2.0.....0.0
    50..........-30.2.....- 1.3.....0.0
    63..........-26.2.....- 0.8.....0.0
    80..........-22.5.....- 0.5.....0.0
    100.........-19.1.....- 0.3.....0.0
    125.........-16.1.....- 0.2.....0.0
    160Hz......-13.4dB...- 0.1dB...0.0dB
     
  16. Wayne A. Pflughaupt

    Joined:
    Aug 5, 1999
    Messages:
    6,100
    Likes Received:
    33
    Location:
    Katy, TX
    Real Name:
    Wayne
    Since “C” weighting is...
     
  17. Harold_C

    Harold_C Stunt Coordinator

    Joined:
    Apr 1, 2002
    Messages:
    198
    Likes Received:
    0
    Wayne:

    That gets at the heart of the whole issue of frequency response in an audio system.

    If you manage to EQ a theoretical system dead flat from 20 to 20,000 using a C-weighted meter, what you have really done is set the system to provide a rather substantial rising bass boost (electrically), in theory to compensate for the ear's inability to hear bass.

    Is this the right thing to do? That's a philosophical question.

    Certainly if I hire someone to play a cello concerto in my living room, there is no "compensation" -- I just hear the cello the way it is played. If I now record that cello on a theoretically perfect recorder and play it back on a system that is "electrically flat", I should hear the cello just like it were live -- the lack of bass perception in my hearing would be the same for the real cello and the recorded cello.

    Or, to hear the cello recording, do I need the system to have a substantial bass boost to measure flat on a c-weighted meter?

    This just highlights how little the engineers really understand about correlating the available measurement techniques to what we really hear from a sound system. The whole area of room coupling with bass drivers is a very complex issue. Basically, we are trying establish some kind of black and white quantitative measure for something that we don't really understand in the first place. I don't believe there is any engineer on the face of the earth who knows for sure what measured frequency response sounds the most "accurate".
     
  18. Rick Guynn

    Rick Guynn Second Unit

    Joined:
    Mar 23, 1999
    Messages:
    473
    Likes Received:
    0
    OK, I have enjoyed reading this thread and found it very educational. I have a question for Harold and Richard...

    What I am gathering from the weighting and oft-published RS 'correction factors' is that if we indeed use those correction factors to 'flatten' our subs, we are in fact building in a house curve to our calibration, electrically speaking...

    So setting it 'flat' with the RS meter and correction curve in fact (for example) sets me up for a 12dB (electrically) bass boost at 16Hz? This then causes my amp to need 16 times the power at 16Hz?

    Am I understanding this correctly? Did I do my math properly?

    Thanks,

    Rick
     
  19. Wayne A. Pflughaupt

    Joined:
    Aug 5, 1999
    Messages:
    6,100
    Likes Received:
    33
    Location:
    Katy, TX
    Real Name:
    Wayne
    Thanks for the interesting response, Harold.
     
  20. Richard Greene

    Richard Greene Stunt Coordinator

    Joined:
    Mar 5, 2001
    Messages:
    148
    Likes Received:
    0
    House curves are so subjective and so sensitive

    to average SPL levels (our ability to hear/feel

    bass varies at different average SPLs) that there

    can be no Rule of Thumb for a "house curve".

    The A and C weightings were originally developed

    to replicate average human hearing ability

    at low volumes (A) and at high volumes (C).

    U-weighting makes no attempt to compensate

    for imperfections of average human ears.

    It just tells us the actual SPL.

    I guess that it makes too much sense to use

    U-weighting with sound meters, so with most

    sound meters we get only A and C weightings.

    I recall reading about the history of these

    SPL weightings when I bought my first sound meter

    about 30 years ago. Unfortunately I don't

    remember the details but the weightings are

    related to the Fletcher Munson

    subjective loudness curves ... and I recall

    the C weighting represented hearing at an

    unusually high average SPL (maybe about 90dB???).

    I remember this because I couldn't believe the

    average person listened to music at such a loud

    average SPL.

    The C-weighting should sound subjectively

    flat to the average person at HIGH volumes.

    So if you like LOUD music, the C-weighting

    might work well for you as your personal

    "house curve".

    But if you listened at lower average volumes,

    as most people do, you'd need additional bass

    boost to make the bass sound subjectively flat.

    Wayne already knows this and also is smart enough

    to NOT equalize for a subjectively flat response

    below 32Hz. which could endanger speakers,

    cause listeners to become nauseous (really) and

    could damage the listening room too (nail pops/cracks/etc.)

    The important question is whether you want to push

    your subwoofer driver for the very difficult task of subjectively flat response down to 20Hz. (easy to accomplish in a car or in a very small room with lots of room gain at 20Hz. ... but usually an impossible dream in a medium-sized or large room)

    I personally do NOT want subjectively flat response to 20Hz. in my room for several reasons:

    (1) Would use up too much woofer excursion and amplifier

    (2) Very few songs have 20Hz. musical content so this is almost a moot point

    (2) My walls resonate (rattle) at 18-20Hz. so boosting

    output at these frequencies = louder rattles = not good

    So if you use C-weighting as your "house curve", as I do,

    and listen at modest SPLs (average SPL for my subwoofer is in the 80-85dB range), the bass will not be subjectively flat -- there will be some subjective roll-off that I can live with (assuming I would EVER notice this, as there is not much content under 30Hz. in music ... and music from the main speakers and rattling walls would tend to mask

    this.)
     

Share This Page