Using SPL meters

[toddalin]

Really...?

Here's my equation:

=10*LOG((10^8.8)+(10^8.9)/2)

It looks as you suggest - I am not seeing the error. Excel insists this equates to 90.1. Where is the error - one of hierarchy of instructions?

FWIW, ironically, that was the meter at FOH last weekend. It was not mine - it was something the contractors had available.

Agreed. One "tool" I never travel without is the Drawmer 1968 - I hand-carry a 4RU rack of Drawmers on every flight. Only my clothes get checked...

As posted elsewhere I insert the 1968 across the Mix buss. It seamlessly takes 3 to 9dB off the transients, entire Mix (obviously settings dependent). I credit this one device, properly used, with providing the most pleasing experience for our fans. Loud never gets too loud, and the subtle compression is inaudible...

Sorry, use this.

=10*LOG(((10^8.8)+(10^8.9))/2)

[boputnam]

Sorry, use this.

=10*LOG(((10^8.8)+(10^8.9))/2)

Dammit - I tried every other permutation of hierarchy. Missed that one!

Thanks. Got it...

[Andyoz]

We find the difference between the L50 and Leq for live and recorded music to only be about 1dB.

That's for rock music that tends to have a fairly high constant level (even live music once the guys shove all the compression in there)

As said before, for traffic noise the difference is typically 3-4dB depending on alot of factors though. Over any given measurement period, you generally get the biggest variation between L50 and Leq where you have a limited number of discrete loud sounds in amongst a much lower overall level. The logarithmic averaging nature of the Leq penalizes those loud events hard baby, the L50 tends to let them slide baby!

[boputnam]

We find the difference between the L50 and Leq for live and recorded music to only be about 1dB.

I don't think that's true, Andy - it is not a linear relationship.

I noticed when Todd provided his example, having a wider range (spread) than mine, his Leq number was 'biased' toward the upper range limit, more than was my example. That is, Leq tends to approach dB(max) as the range (spread) increases.

In-haste (headed out to a weekend of gigs...), I prepared this graph, showing how L50 (mean) varies wrt Leq for the range 85dB to 106dB.

Is this close, Todd?

[toddalin]

I don't think that's true, Andy - it is not a linear relationship.

I noticed when Todd provided his example, having a wider range (spread) than mine, his Leq number was 'biased' toward the upper range limit, more than was my example. That is, Leq tends to approach dB(max) as the range (spread) increases.

In-haste (headed out to a weekend of gigs...), I prepared this graph, showing how L50 (mean) varies wrt Leq for the range 85dB to 106dB.

Is this close, Todd?

Agreed, as the spread increases, the Leq will show more difference from the L50. But it is based on the relative volume levels of the spread and not some absolute number (e.g., 85 dB where the paths diverge. (Maybe it was in the previous example and that's what you are illustrating.)

Andy could be right if the music is highly compressed and this will vary with each and every case.

Tell you what.

Pick a song and if I have it in my selection, I'll/we all can monitor and report the Leq, L50, Lmin, Lmax, and Lpeak and any other L value you desire. My noise floor here is under 40 dBA. So long as we don't exceed ~100 dB, dynamic compression shouldn't be much of an issue with my L300 wannabes.

Lets pick a song lots of people have access to and monitor it with their meters too so they can have an idea of what the levels are relative to what they've been monitoring.

Anyone with suggestions?

[Robh3606]

Lets pick a song lots of people have access to and monitor it with their meters too so they can have an idea of what the levels are relative to what they've been monitoring.

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??

Rob

[boputnam]

Andy could be right if the music is highly compressed and this will vary with each and every case.

Yeah, I thought of that when I was headed-off to my first-of-three gigs. How Leq relates to L50 is dependant on the dynamic range.

Tell you what.

Pick a song and if I have it in my selection, I'll/we all can monitor and report the Leq, L50, Lmin, Lmax, and Lpeak and any other L value you desire. My noise floor here is under 40 dBA. So long as we don't exceed ~100 dB, dynamic compression shouldn't be much of an issue with my L300 wannabes.

Great idea. I would think you'd want something with wide dynamic range?

Use a pink noise source??

I don't think Pink Noise would work - it has no dynamic range. It's kind of like a banjo in that regard...

[Robh3606]

I don't think Pink Noise would work - it has no dynamic range. It's kind of like a banjo in that regard...

Hello Bo

I thought this was log related. If that's so would it make sense to have the output all at the same relatve SPL with a pink noise source so we are all at the same level?? If we listen at different levels it would affect the Leq and max numbers no??

Rob

[boputnam]

Oh, I see. I was too focussed on measuring / calculating Leq - that would require dynamic range.

As far as crosschecking measurement of SPL by various meters, you are right - steady Pink Noise would be perfect.

[toddalin]

The sound that you hear on a moment by moment basis is called the SPL (sound pressure level). This is what you guys are observing on your meters/displays. The level that is exceeded 50% of the time is the L50.

When you guys are reading the Rat Shack (or any other non-integrating meter), I referred it to the L50 because I figured that based on what you guys were reporting, this was basicly at the center of your volume range, with the needle bouncing back and forth pretty much equally to both sides.

But by strict definition, the L50 is the level that is exceeded 50% of the time. If your sound is "strobing between 60 dB and 90 dB, with equal time at each end, the L50 is 60 dB as this value is being exceeded 50% of the time. But the Leq is 87 dB.

Now, it it were strobing between 60 and 120 dB, the L50 is still 60 dB, but the Leq is now 117 dB.

Not having a Rat Shack meter myself, I don't know if I can exactly duplicate what you guys may read (unless someone wants to bring one over for the test and compare it with a professionally certified meter).

But regardless of what level anyone listens at the relation of the Leq and all the other L values will be the same as this is based on the dynamic range and not the absolute volume. If the peaks are 20 dB louder than the average, it makes no difference what average you listen at (unless you are to the point of dynamic compression), as the relationship is retained.

[Andyoz]

I don't think that's true, Andy - it is not a linear relationship.

I was actually surprised myself when I looked back at measurements done at venues and nightclubs. L50 is a parameter we rarely use in our game. It's normally the L90, Leq and Lmax that is reported. L10 is also used for traffic noise and sometimes it's used to assess music noise.

Note that my comment relates to several minute samples of bands, etc. really "going for it". If you were to measure over an entire gig including slow songs, breaks, etc. the logarithmic nature of the Leq will come out more and the Leq value will just keep sitting up high whilst the L50 starts to drift down to a lower level.

Also, note that if you are only looking at the overall dBA levels, then the measured levels will be dominated by the mid-frequency energy. This is exactly where rock music maintains a high "average" energy level because you have guitar, vocals, keys, snare, etc. fighting for that frequency range. If you were to isolate just the 63Hz octave band (i.e. "bass"), you see a bigger difference between the L50 and Leq as the signal is more impulsive.

Try doing the test using dBA and dBC measured levels and you should see what I mean.

[robertbartsch]

What was the typical SPL at a Deep Purple concert - say 20 rows back - in the 1970s?

PS - This was loud enough to cause ringing in the ears for several hours.

In the modern era, I have not come across anything like the SPLs in the old days. I suppose this is a good thing.

[Robh3606]

What was the typical SPL at a Deep Purple concert - say 20 rows back - in the 1970s?

I brought my old Radio Shack analog to a Greenslade show in the mid 70's. We were in the 15 row and when the band started the meter pegged at full scale on the 110db range. Full scale was 113 or 115db can't remember. We got the hell out of there and moved as far back as we could and it was still way too loud. It was one of the loudest shows I was ever at. Ended up using toilet tissue to make ear plugs. Always carried some kind of ear plugs after that one.

Rob

[1audiohack]

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.

[Andyoz]

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???