Film-Tech Forum ARCHIVE: Linear Meter

This is topic Linear Meter in forum Film Handlers' Forum at Film-Tech Forum ARCHIVE.

To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=1;t=004436

Posted by Antonio Marcheselli (Member # 373) on 01-14-2003, 08:37 AM:

Hi everyone.

Sorry Gordon if I started another thread, the argument become important... [Smile]

I had a talk with one of you by mail about measuring the pink noise level using an RMS meter. I've heard once about an "audio meter" that is linear from 20 to 20.000 while cheaper meter are reliable only around 50Hz.

measuring a Pink noise would mean have the meter measure from 20 to 20.000: who says that the meter will read all the frequencies at the same level? I had an idea. I have a Behringer 8024 equalizer that is able to provide a sine as test tone.
I set a level and I start sweeping from 20 to 10.000 Hertz, watching what the meter says. I do not mean that I use the swept tone, I made the freq change manually each time. At a random setting (-17.5dB) the results were:

20Hz:625 mV
25Hz:625mV
31.5Hz:625mV
40Hz:625mV
50Hz:626mV
63Hz:626mV
80Hz:638mV
100Hz:630mV
125Hz:529mV
160Hz:513mV
200Hz:700mV
500Hz;548mV
1Khz:650mV
4Khz:762mV
12.5Khz:880mV
20Khz:1428mV

Can I suppose that my behringer is a reliable source of sine? What does it mean? Isn't my meter linear 20-20.000?
Please note that at maximum output (0.0dB) the output of my device is measured at 50Hz as 4.67V. Spec sheet says +16dBu that converted in Volt (I use a web page, I hope that it is right!) is 4.89V... Not too far!

Oh, I've connect the meter to the LINE output and not in parallel to a speaker.

Bye

Posted by Dave Macaulay (Member # 813) on 01-14-2003, 11:02 AM:

Using tones for auditorium EQ brings up a lot of difficulties.
- how do you apply the C-weighting curve? Lots of calculations.
- Tones excite room resonances and the SPL measurement will be horribly inaccurate. Standing waves form, so there are additions and cancellations. Moving the microphone a little you will see huge changes in SPL.
- HF loudspeaker components are easily destroyed when driving the system with pure tones, even LF drivers can be ruined at certain resonance points as the coils can go out beyond the magnet gap - and probably miss when they try to go back in.
Pink noise avoids these problems. Any given frequency is present at random times and at random phase so standing waves don't form, speaker resonance is avoided, and applying enough power to damage HF drivers is difficult - the amps may have trouble doing it (because they are driving the entire frequency range) and the noise gets so damn loud you notice it may be excessive. A pure 18KHz tone at high output may be inaudible to many of us but will cook an HF driver coil in very little time.

A sweep oscillator is useful for tracking down mechanical resonances in the screen structure and A/C ductwork but using tone for EQ is a bad plan.

Posted by Antonio Marcheselli (Member # 373) on 01-14-2003, 05:51 PM:

Dave,

Perhaps I was not clear, I'm sorry.

I've connected the meter directly to the LINE OUTPUT of the equalizer. No room, no speaker!!!
And, I was not going to equalize a theater with sine wave, just to understand how a METER can be considered "linear"

Bye

Posted by Steve Guttag (Member # 268) on 01-14-2003, 07:21 PM:

Ok...first off (and I should have brought this up in the other thread). The 300mV reference level is for a single tone, not wide band pink noise...PN with C-weighting will be closer to 360mV and will be difficult to read without a meter with very slow ballistics.

Using pink noise REQUIRES weighting for any sort of meaningful result. Pink noise should be flat from "DC to Daylight" if the generator is that good. If your meter can read that wide a band, it will read higher. Once you put in a weighting, like the "C" weighting, you have band limited the system. If your meter is flat over the pass band defined by the "C" weighting then it should read properly. The "C" weighting basically applies a 20Hz - 20KHz passband filter over the meter so it doesn't pay attention to frequencies outside of that range.

As to another's comment...playing tones with "C" weighting makes no sense since you have absolute control over the range of interest. You could, frequency by frequency play the appropriate tone, use your calibrated SPL meter and proceed to tune up the room...this method would show some of the flaws of the ever-present Radio Shack meter.

Pink Noise is convienient since you can do all frequencies at once. If you had the appropriate measuring system, you could do tone bursts and nix the use of the whole "X" curve thing too. In many rooms, with high RT times, you'd be better off with this method too.

Steve

Posted by Antonio Marcheselli (Member # 373) on 01-15-2003, 05:45 AM:

Steve,

First of all: what is that things in your picture???

Thanks for you reply. I understood about measuring PN with a meter and about weighting the measure. Ok.

But I believe that you missed my center question. In some instances I had a talk with a very qualified tech about measure 300mV on DTS player. This has caused me to make the test I've show above: connect a meter (Voltmeter) to a signal generator. I've connect the meter directly at the LINE OUTPUT of the generator, no speaker or ampli connected, like a DTS player!

I've played several test tone, at the same level (-17.5dB on the signal generator) and the results are above. If the signal generator is reliable, shouldn't be read the same value for all frequencies?

I mean, if the generator is set to -17.5dB shouldn't I measure, say, 280mV to all frequencies from 20 to 20.000??

Bye