Using SPL meters

[toddalin]

OK Todd, I have never used this feature on my TEF, so please help me see if I have this right.

I set up with something a few of us might have, Jennifer Warnes, Bird On A Wire. At a comfortable level my old R.S. meter read about 70 in the vocal parts, and 80 in the louder parts with percussion. "A" weighted, slow response.

The Noise Level Analisys reads as follows

Instantanious Min 49.4dB Max 87.4dB Lmean 72.3dB
Percentiles L10 79.7dB L50 73.4dB L90 63.5dB

Weighting A Lmin 50.6dB Lmax 86.2dB Leq 76.6dB

Always anxious to learn, does that look like reality?

Thank you.

Andy is correct and this looks about right for a song with wide dynamic range (I don't have this cut).

Note that the Leq is a little more than 3 dB louder than the L50 as we previously discussed. I notice that you did not include the Lpeak.

The L10 is the value that is exceeded 10% of the time while the L90 is exceeded 90% of the time. These values have no basis in regulation, but serve as a basis to jugde the "quietness" or "loudness" of an environment. As previously noted, the L50 represents the value that is exceeded 50% of the time. Half the time the sound is louder and half the time it is quieter.

[toddalin]

Those look about right for a slow, impulsive song.

Not sure what the Lmean value is though...The value is different from both the Leq and L50 values???

The Lmean (or Lavg on my meter) is similar to the Leq except that a different "exchange rate" is used.

The "exchange rate" refers to the level of increase that represents a doubling of the power. While we tend to think that doubling the acoustic power adds 3 dB to the sound (and this is the basis for the Leq), other "exchange rates" are also used in the industry.

For example OSHA uses an "exchange rate" of 5 dB. Thus, OSHA figures that a doubling of the power results in a 5 dB increase. When my meter reports data, a selected "exchange rate" of 3 dB displays "Leq" on the LCD display, whereas a different exchange rate (selectable from 3 - 6 on my meter) shows "Lavg" on the LCD display.

Audiohack, you may go into your meter's menu/manual and see what exchange rate is being used for that display.

[toddalin]

On Monday I did monitoring of Dire Straits Money for Nothing both on the original CD and 20 year remastered CD. This song is very highly compressed with a real world dynamic range (ignoring the fade-out, Lmin - Lmax) of only about 15-20 dBA!

I initially went into to the chorus part (~2 min in) and set the volume to an almost uncomfortable (relatively loud) SPL of ~80 dBA. This was just the two L200 cabinets (2235/LE175-HL91/2402, custom crossovers ~800Hz/7,250Hz) and all subs, shakers, and surround speakers were off.

After the first ~5 seconds, the opening levels are ~67-73 dBA during the synth pans and I want my MTV (slow meter response). The last 5 seconds of the fade out go to <40 dBA. The rest of the song is pretty much the same.

Leq = 78.3 dBA, Lmax = 86.6 dBA, Lmin (during fade out = 41.6 dBA = threshold floor of meter on 50 - 110 dBA scale), L10 = 80.6 dBA, L50 = 78.2 dBA, L90 = 67.4 dBA, Lpeak = 104.4 dBA

Note that the peak requires 400 times the power (26 dB louder) of level for the Leq. Also note that in this case, the Leq and L50 are less than 1 dBA apart.

I then monitored the 20 anniversary re-mastered edition and came up with the following (again setting the SPL at ~80 dBA ~2 minutes in). This version is even more highly compressed and the Leq ~= L50!

Leq = 78.6 dBA, Lmax = 86.0 dBA, Lmin (during fade out = 41.6 dBA = threshold floor of meter on 50 - 110 dBA scale), L10 = 81.1 dBA, L50 = 78.6 dBA, L90 = 66.2 dBA, Lpeak = 99.3 dBA

Here the peak requires a little over 100 times the power of the Leq.

Interesting that the remastered edition was a full 10 dBA "hotter" than the original, and when I put it on (SPL ~90 dBA) it about blew out my ears until I turned it down. This of course infers that even if you go to great lengths to set your preamp volume levels to some standard (e.g., dolby digital standard), the media that you play may not be to the standard so your settings are for naught.

I also have this disk on DVD-Audio and am curious as to the dynamic range (haven't tried yet). This may not be a comparable test though because it requires the full use of all channels. Still, dynamic range is dynamic range regardless of how many speakers are cranking it out.

[Andyoz]

Also note that in this case, the Leq and L50 are less than 1 dBA apart.

Yep, that's what I get with rock'n'roll!

Have you tried comparing the dBC and dBA. I bet the dBC Leq and L50 are more like 3-4dB different.

[Andyoz]

The Lmean (or Lavg on my meter) is similar to the Leq except that a different "exchange rate" is used.

Oh, no wonder I haven't seen it. We don't use that in Eurpoe, it's a 3dB exchange rate for occupational hearing assessment.

[toddalin]

Oh, no wonder I haven't seen it. We don't use that in Eurpoe, it's a 3dB exchange rate for occupational hearing assessment.

Alternatively, maybe on your meter Lmean is a simple (non-logrithmetic) average. You would need to refer to your manual.

[toddalin]

Yep, that's what I get with rock'n'roll!

Have you tried comparing the dBC and dBA. I bet the dBC Leq and L50 are more like 3-4dB different.

I haven't tried, but this sounds reasonable. Linear and impulse weighting should increase this difference by even more. I'm also thinking that a "fast" response would show a greater difference. "Slow" response is averaged over 1 second whereas "fast" response is averaged over 125 ms (if memory serves).

[1audiohack]

Lmean and Leq as defined in my manual.

"Leq is the dB level derived from the average level of Intensity of the sound over a given period, where the intensity is the level expressed in terms of Pascals squared."

"This differs from Lmean, which is the average of the dB level of all of the individual Leq;s that have been recorded during the measurement."

"This unit takes an Leq reading every 85ms or 703 times a minute. Each individual Leq result is calculated over a 4096 sample block at a 48kHz sample rate."

I do not see an Lpeak anywhere in the manual, or the screens.

I have both of the Dire Straights recordings you mentioned. I hope to get a chance to NLA them tonight, just to compare our results.

It looks like this would be a good tool for comparing the dynamic capabilities of loudspeakers.

Thank you both for the education, I look forward to more!

[1audiohack]

"

Sounds like fun. How are we going to set the level? Use a pink noise source?? I am assumng we need this played back at the same level to see how different the meters are??

I have three test CD's with pink noise tracks and they all play back at different levels. From 87.7 to 91.3dB. Only on the Stereophile Test CD #2 were they kind enough to state that it is at -20dBFS, it is also the loudest. Even the Alan Parson Sound Check states no level, and is the quietest. hmmm,,,

I ran both Dire Straights tracks last night and get very nearly the same results as did Todd.

Also the "Exchange level" set in my software is indeed 5dB. For what we are doing is 3dB a more appropriate choice?

[toddalin]

"

Also the "Exchange level" set in my software is indeed 5dB. For what we are doing is 3dB a more appropriate choice?

Yes.