[IkeH]

Roger,

As you have recommended, I got the Spl meter from Radio Shack.  I got the digital one (Model #: 33-205)  for $49.95 which was $5.00 more than the analog meter (Model #33-4050)   

I was very surprised to learn that at my normal listening volume, the average sound level was around 70db.  Peak sound level at the loudest volume I would ever listen to was 85db during the fff passage of an orchestral piece.   

This experiment helped me to choose EM7 2.5W version instead of 5W version (to work with two speakers I have with 90db and 95 db efficiency.)  I would say that the investment in the SPL meter was well worth it because I was able to confirm that I need not spend $200 more and potential hassle of adjusting bias for the 5 W version.

BTW, I borrowed  a radio shack analog meter from my audiophile co-worker.  I was able to confirm that readings were the same with both meters.

 Isaku

[Roger A. Modjeski]

Thanks for your post. I think I have helped Radio Shack sell 100 SPL meters by now. Anyone who has spent over $100 on a cable or accessory has no excuse for having one of these. It's absolutely essential.

Over 90% of those who have bought an SPL meter has informed me that they are listening 10-20 db lower than they had expected. Remember if you are 10 db lower that means your amp is 10 times more powerful than you are using. If it's 20 dB less that's 100 times. If you err in the other direction just reverse the numbers and remember 3 dB is twice times the power, 10 dB is 10 times, 20 is 100, 30 is 1000.

If anyone has questions on why or how to use one please write here. Let's carry on this most important topic.

Roger

[hartwerger]

I have the Radioshack digital SPL meter.  I bought it before they reissued the analog meter.  When the topic of SPL meters comes up, it seems the  recommendation or preference is for the analog meter.  Is the analog meter better? and why or how would it be better (more accurate, perhaps)?  I acutally use my digital meter not only for listening volumes, but also to roughly measure frequency response in my room.   Wondering if I should pickup an analog meter or just stick with my digital. 

Also, I've seen posts where others have suggested that the Radioshack meters should be re-calibrated.   I have an idea what is meant by the calibration, but do you think its necessary for casual use?  Thanks for any advice.

[mfsoa]

If anyone has questions on why or how to use one please write here. Let's carry on this most important topic

OK, I've got two:
Can you address the taking of SPL measurements with either both or only a single speaker playing for a stereo system?


How omnidirectional is the mic of the RS SPL meter? If I'm calibrating a 6.1 home theater, do I need to rotate the meter for the rear channels?

Thanks

[geezer]

Another question: I'm having a running argument with my wife about what the appropriate volume level should be when listening to classical music . Can anyone tell me approximately what the spl might be (on the loudest passages), in the dress circle say, at a symphony concert?

If I can get an answer to this question, and the question about the analogue vs digital Radio Shack meter I intend to get one.

[gitarretyp]

Classical can have peaks in the 100-105 dB range.

[Russell Dawkins]

How omnidirectional is the mic of the RS SPL meter? If I'm calibrating a 6.1 home theater, do I need to rotate the meter for the rear channels?

Thanks

I don't have my SPL meter manual anymore, but I remember that the mic is meant to point at a 90 degree angle to the direction of the source of the sound being measured. This suggests pointing it vertically and you shouldn't have to change anything to measure the rear channels.

It is calibrated to measure flattest at this angle of incidence.

[Russell Dawkins]

Another question: I'm having a running argument with my wife about what the appropriate volume level should be when listening to classical music . Can anyone tell me approximately what the spl might be (on the loudest passages), in the dress circle say, at a symphony concert?

If I can get an answer to this question, and the question about the analogue vs digital Radio Shack meter I intend to get one.

I find the analog version to be easier to interpret - you can get information from the rapidity of the needle movement, whereas the refresh rate of the digital is not fast enough.

I would estimate that the typical CD recording of a symphony orchestra has between 6 and 15 dB of compression applied to it. Since John Atkinson reportedly measured 105 dB on the tympani hits in the audience on a loud piece of music, that would suggest that, with the meter set to C weighting, fast response mode I would expect to see roughly 90 - 99 dB on those same hits, presuming meter ballistics similar to J.A.'s meter.

If theory matches reality to some extent, this would result in average levels similar to the concert.

Of course, once you buy the meter, take it to the concert and measure the levels and match them at home, if you can do that without encountering that unpleasant strained sound of the amp or the speaker working just that little bit too hard.

[JLM]

Soundstaging, even imaging, can collaspe before noticable amp strain.

Now that you've got the spl meter, run some test tones to find out the depth of bass you're listening to and how high you can hear.  This should confirm how extreme the marketing hype of "needing" 20 - 20,000 Hz response is.  I'll concentrate on 40 -10,000 Hz long before worrying about reach the last two octaves.

Every audiofest/gathering and audio shop ought to include a spl meter.

But watch out, or you may go "too far" and get snagged by the room EQ bug.   

[geezer]

Thanks fellows. For those who haven't seen it, I refer you to a correction table for the Radio Shack meter at http://www.subwoofer-builder.com/SPL-corrections.htm

[fonz]

Hello

I am a longtime user of a RAM-9 with Michael Green Chameleons and had no complaints.  Now that I have entered the high efficiency game, I have gone from the Horn Shoppes to the PHY SAG with open baffles.  I currently use a pair of 300Bs that started life as Welbornes but have since been completely gutted and replaced with a circuit by Larry Moore and must say that the difference is staggering.

I wanted a second amp, something with a little less power in exchange for more delicacy, that's when I stumbled across the EM7.  All comments have been favorable and I'm not surprised knowing Rogers commitment to his product and tubes.

The point: Will a passive linestage (Reference Line) or Open Baffle speakers (97db) pose a problem in my 15' x 15' room with the 2.5 or should I look at the 5 watter.  My local Shack didn't have the meter and would have to order it for me and it's a 2 week wait.

[martyo]

Also, I've seen posts where others have suggested that the Radioshack meters should be re-calibrated.   I have an idea what is meant by the calibration, but do you think its necessary for casual use?  Thanks for any advice.

Thanks fellows. For those who haven't seen it, I refer you to a correction table for the Radio Shack meter at http://www.subwoofer-builder.com/SPL-corrections.htm

If anyone has questions on why or how to use one please write here. Let's carry on this most important topic.

I would really appreciate an answer to this recalibration question. The values on this chart are significant. If the meter is that far off it could certainly lead to assumptions like

This should confirm how extreme the marketing hype of "needing" 20 - 20,000 Hz response is.  I'll concentrate on 40 -10,000 Hz long before worrying about reach the last two octaves.

[martyo]

bump

[Roger A. Modjeski]

Martyo,

I have examined the values on the Radio Shack meter at the link below. They are quite acceptable. Even when the meter is 1.5 dB off at some frequencies in the upper end, it will not thwart the intention of this measurement. We are simply trying to find, within a few dB, the listening level for the purpose of selecting the right size amplifier.

http://www.subwoofer-builder.com/SPL-corrections.htm.


Nine times out of ten when a prospective customer calls and asks which amp is big enough for his needs he has no idea at what SPL he listens. "Not too loud" and "I like to turn it up when I get home from work" are not good answers. If you miss your SPL target by 6 db thats 4 times the power in watts. If you miss it by 10 dB thats 10 times and 20dB is 100 times. I am not kidding when I tell a customer who thinks he needs my 100 watt amp that he really only needs 35 watts or even 2.5. I will state without reservation that listener level differences have a wider range (65-105 = 40dB) than most speakers (82 to 102 = 20dB). There will be some exceptions in both those numbers but they are rare.

Below is a link that will show you the A, B and C weightings both by table and graph. C is what we use for this measurement. Don't sweat the roll-off at low frequencies. If the meter went down to 20 Hz flat air motion in the room could easily obscure the correct SPL at low levels. Even though the RS meter misses at the lowest frequencies that is not going to make any material difference is finding the listeners SPL.

http://www.sfu.ca/sonic-studio/handbook/Sound_Level_Meter.html

What I imagine is going on in the reader's mind (given those quotes and questions) is the assumption that one can measure frequency response in their room with this meter. The simple answer is you can't. Standing waves and reflections will cause the meter to swing wildly with small changes in frequency and the numbers will be meaningless. The best way of testing a speaker's frequency response is a tone-burst generator and an oscilloscope. We did this at Beveridge and I do it for all my speaker tests. The burst is all over by the time any reflection comes back to the mic. The burst amplitude is recorded and, even more valuable, I get to see the shape of the burst. MLSSA can not do this and frankly satisfying MLSSA has resulted in some pretty bad sounding speakers. Harold Beveridge (we called him Bev) often said to speaker makers when shown their frequency response curves "Yea, that's great, now what does it look like when you move the microphone". He could move the mic all over the place and the response was virtually unaffected.

The bottom line of this topic is that we simply want to get in the ballpark as to power readings. Anyone who has an oscilloscope can skip the SPL measurement, go right to the speaker terminals and report the peak voltage at his highest listening level. That is the most accurate way to determine amplifier power, but I doubt many people have a scope. This is the sole reason I use have encouraged hundreds of listeners to get an SPL meter. To get your peak power simply take the SPL reading at 1 meter from the speaker while playing it at your listening level. Take that SPL subtract the 1 watt speaker sensitivity given in your speaker specs and you will have the dB above one watt that you need to produce that level.

Since math skills vary from person to person let's do an example. Say you measure 85 dB at one meter on axis and your speaker is 85 dB sensitivity then you need only one watt. Two watts will give you 3 dB of headroom and it will take 10 watts to give you 10 dB headroom. Still a rather small amp and you don't need more than 3 dB headroom for most music.

That one was easy. Say you measure the same 85 dB and your speaker is 95 dB for one watt. These are realistic numbers for the FE-103 I build and for many single cone speakers available today. Again subtract your SPL reading from the speaker sensitivity and you will get -10 dB (1 meter SPL - speaker sensitivity at one watt). Putting in the numbers: 85-95 = -10 dB. Minus 10 dB means you are listening 10 dB below one watt which is 1/10 of a watt. Again, -3 dB is half the power and every 10 dB is 1/10th. That makes 20 dB 1/100. I do plan to write an article on figuring dB's by an easy method I have developed that one can do in one's head.

For those of you who have peak reading meters that will hold a reading, you can use that in lieu of an oscilloscope. We are just looking for a peak voltage value. An amplifier that puts out 100 watts at 8 ohms will have a 40 volt maximum (peak voltage). A 100 watt 4 ohm amplifier will have a peak of 28.5 volts. To go about the numbers from the speaker side; take the peak voltage measured, multiply it by .707, square that number (multiply it by itself) and divide the result by the speaker impedance. That will be the power you are using and should be a few dB below the rated power of your amplifier.

As I write this last part I recall a reviewer who wanted to review my 300 watt monoblocks. I asked what speaker sensitivity at which he listened. He already had determined he used just a watt or two so I suggested the RM-10 (my SE amps came later). He told me he could tell the difference between a 100 watt and 300 watt amp even at those levels. I did not send him a unit for review.

Roger

[miklorsmith]

Great post!   

[mfsoa]

So you must love hearing, who is it, Musical Fidelity?, telling us that anything less than 500 watts just won't cut it!!

Seriously though, thank you very much for the wonderful technical information you are providing!

-Mike

[BobRex]

Roger, maybe I missed it, but did you also mention the 3dB gain from stereo?

[Roger A. Modjeski]

Mike,

It was a good question. As to the 500 watts, are you referring to the Musical Fidelity booster that Stereophile reviewed recently? What they have done seems pretty silly and could be done to any amplifier plus a few resistors. Since they have the regular line level inputs also I assume they just provide a load for the driving amplifier and a voltage divider to get back to line level.

The car amp people have been doing this for years having both speaker level inputs and line level. The car people are smarter because they don't bother to load the driving amplifier since its solid state. A tube amp appreciates some load to keep the primary voltage swing within safe limits for the output tubes and transformer.

[Roger A. Modjeski]

Roger, maybe I missed it, but did you also mention the 3dB gain from stereo?

That's something to ponder. I will say that if you have one speaker playing (as done by swinging the balance to the right or disconnecting the other speaker) and you bring the other one in back in  the meter will go up  6dB for mono information. For uncorrelated noise (pink or white) you are correct, the meter will go up 3 dB.

As to 3 vs 6 dB are you considering the added amplifier/acoustic power that another channel produces or are you keeping total amp output constant?

[martyo]

Thank you Roger. That is indeed what I thought it could be used for, measuring frequency response in a room.