Calibration Q's - pink noise vs sine waves etc..

R

rob_w

Guest
Hi Chaps,

Somethings been nagging me for a while now.

A lot of people here use the sinewave/ radioshack spl meter spreadsheet thing to calibrate/eq their subs.

When the xl spreadsheet corrects the spl meter, does it correct it so it measures the actual spl, or does it include a 3Db/octave slope to replicate a pink noise slope?

ie: if you eq'd your sub flat using the spreadsheet, would you end up with a 'flat' response that with pink noise would measure 6Db lower at 20Hz than it would at 80Hz ?

Nobody here seems to address this. I've always used pink noise to calibrate stuff with, and feel that if I used the sine wave method I'd end up with a weak bottom end.

I guess some people would call the difference between the responses a 'house curve' ?

Any thoughts?
Cheers

Rob
 
""Try asking the question again in English ""


Hehee, maybe I should've asked what the best budget sub was:devil:

I'm sure some of the 'calibrating heavyweights' will chime in soon enough........ anytime now.... any minute........ hello?. lo ...lo...lo.......echo..co..co..:D

Rob
 
C'mon folks - 75 views and no takers:eek:

Ian, I think I am talking gibberish!

Rob
 
The 'snapbug' one - I'll post a link when I get home tonight.

Cheers

Rob
 
Originally posted by robwells
Hi Chaps,

Somethings been nagging me for a while now.

A lot of people here use the sinewave/ radioshack spl meter spreadsheet thing to calibrate/eq their subs.

When the xl spreadsheet corrects the spl meter, does it correct it so it measures the actual spl, or does it include a 3Db/octave slope to replicate a pink noise slope?

It's an un-weighted (weightless?!) correction for accuracy purposes, IIRC.

When you equalise using discrete tones of the same amplitude, you're essentially referencing the same "curve" as white noise (which, of course, isn't a curve at all - it's flat across the whole spectrum). However, if you use white noise to calibrate there will be a negative skew in the measured response due to there being "more" frequencies between higher octaves. It's a little odd to think of it this way because we're dealing with a continuous variable, but bear with with me. Each component frequency in white noise is exactly the same amplitude but there are "more" of them per-interval, i.e. a greater range, as you get higher and higher. Hence, overall, more energy is delivered at high frequencies and white noise sounds "bright".

Pink noise is just filtered white noise. The -3dB slope is indicative of there being an equal amount of energy per octave, so calibration is not biased towards the high end. Also, pink noise is supposedly more representative of the sounds we hear in everyday life; especially music (makes sense, I guess). There's no need to account for this when equalising because you're only dealing with discrete tones.

ie: if you eq'd your sub flat using the spreadsheet, would you end up with a 'flat' response that with pink noise would measure 6Db lower at 20Hz than it would at 80Hz ?

You'd end up with a C-weighted curve, which approximates human hearing (in terms of frequency response) at high SPL levels. The C contour filter is largely flat but rolls off at both ends (-3dB@30Hz and 8kHz I think). This adjustment results in an empirically "flat" response - that's the idea anyway, lol.

I guess some people would call the difference between the responses a 'house curve' ?

This I find a little confusing. Most pro audio guys will use "house curve" to refer to the response of a studio or auditorium to pink noise. It also seems to be used in AV circles to describe a slight lifting of frequencies below 80-100Hz, which some people prefer to a meter-flat response. I'm not really sure whether the two concepts are related or where confusion has arisen. Perhaps it's something to do with calibrating the LFE channel using bass-managed pink noise (you have to lift the LFE to compensate).

Definitions aside, a flat response measured using the meter is a probably good starting point. :)

MuFu.
 
Hi Mufu,

A good post:smashin:

The house curve thing I was referring to is mentioned over at home theater forum. I believe it is a boost anywhere from 6 - 12Db at around 30Hz, slowly dropping to 'normal' around 80 - 100Hz. This ties in with your definition.
It was added to a 'sine wave' calibrated system, so would end up 0Db - 6Db boost with pink noise. :D I think....

Cheers

Rob

ooh, links :

snapbug:

http://www.snapbug.ws/bfd.htm

Home theater forum:

http://www.hometheaterforum.com/htforum/showthread.php?s=&threadid=54496&highlight=house+AND+curve
 
Originally posted by robwells
The house curve thing I was referring to is mentioned over at home theater forum. I believe it is a boost anywhere from 6 - 12Db at around 30Hz, slowly dropping to 'normal' around 80 - 100Hz.

I tried something like this after I EQ'd my sub. Problem was that I'd already applied quite a large amount of gain to correct dips, so boosting things further resulted in a few scary noises. :rolleyes:
 
Good on you for trying though:) - Did you like the results before it went scary?

Rob
 
Hmm... perhaps initially but in the long-run, no, not at all. Flat on the snapbug spreadsheet certainly doesn't sound bass-light, IMHO. I'm in a bedroom though - people in brighter rooms might find this "house curve" warms things up a bit.
 

The latest video from AVForums

TV Buying Guide - Which TV Is Best For You?
Subscribe to our YouTube channel
Back
Top Bottom