scale on "waveform" y axis

[TRP]

Hi,

I have recorded some cymbal sounds under different conditions for a science experiment and am measuring the amplitude of the wave at the moment the cymbal is struck to obtain a figure for the volume (using "waveform" rather than waveform db under the audiotrack menu). My problem is how to express in my data tables what the scale on the y axis actually means - it goes from -1 to 0 to +1.

Thanks for any help.

Ben

[kozikowski]

It doesn't mean anything which is the down side of using percent instead of dB for editing. It is roughly percent. "1" is 100% of the allowable volume in either direction. So 0.5 is really 50% and so on. You need up and down, positive and negative because sound vibration goes in both directions.

Some radio personalities have a dominant "mostly positive" or "mostly negative" voice (uncompressed) and you can recognize when they speak on a timeline just by looking at the blue waves.

The problem comes when you need to map the performance to how your ear works. Half volume to your ears is very roughly -18dB which works out to 12%. 12% and we're only half-way down from maximum. The limit of human hearing is about -60dB which is half and half again ten times.

I can see where you would give blank looks by recording a percussion instrument and then trying to analyze it. Those can look like they only go one direction, either up or down, but if you look closely in dB range, the cymbal is really ringing with the up and down waves rapidly dying toward -60 or wherever your system is set to display. You can get a better idea of how this works by striking a triangle. Those ring at a specific frequency and you can see by magnifying the waveform what it's really doing.

Another oddity of sound is that recorded sound can only handle the vibrations or condensing and expanding blobs of air down to about 20 vibrations per second. If you had a theoretical system that went down to 0 vibrations per second, that would give you a speaker system which could deliver wind (or vacuum) and blow out candles. Each engineering class goes through a thought exercise where they design the "perfect" sound system that has no frequency limits. You could do it with mechanical valves and compressor pumps instead of paper and metal loudspeakers.

Koz

[kozikowski]

Here's another thought experiment. Say you held a piece of printer paper up in front of somebody's face like you want them to read a message on the paper. Now bring the paper closer and farther away by a couple of inches or so, generating slight high and low pressure zones between the paper and the person (+ and - on the waveforms).

If you could do that 20 times a second (super fast hands), that would be the lower limit of human hearing -- one of those big pipes of a really large church organ.

If you could move the paper back and forth 440 times a second, the paper would make the same tone as the oboe at the beginning of an orchestra performance. That is exactly how the paper cone in a loudspeaker works.

Koz

[steve]

My problem is how to express in my data tables what the scale on the y axis actually means - it goes from -1 to 0 to +1.

The scale on the vertical axis is the numerical value of the samples. (linear scale)

If you click on the name of the track and select "Waveform dB" the vertical scale will change to "dBfs" (dB relative to "full scale").

-6dBfs is about equal to 50% on the linear scale.

[kozikowski]

<<<amplitude of the wave at the moment the cymbal is struck to obtain a figure for the volume >>>

I thought we had all missed the point here. Volume is the difference between the + and the - values. You can't get anything close to a volume measurement with just one or the other number.

You get into the perceived volume and loudness thing here, too, so I'm not going there, but you do need to take into account both numbers. That will probably give you what you want for the experiment.

This is all wave theory, so I'll do waves. A tsunami isn't just a large wave hitting the shore with enough force to alert CNN. The first thing that happens is the tide goes way out and stretches of shoreline that haven't been seen in years are no longer covered with water. Then the wave hits shore. The tsunami is properly measured between the maximum low tide and the size of the wave hitting shore, not the overall average water level (ignoring tides for a minute) to the peak of the wave. The tsunami event is both the low tide (-) and the wave (+).

Same with sound events. They almost always have air moving both directions.

You can't make actual loudness measurements this way, this is just a scientific curiosity. The real world takes into account the oddities of your ear, propagation through the air, humidity, velocity, etc. For example, the BBC PPM sound meter for radio shows is designed to not respond to very brief sounds no matter how loud they are. Your ear is never going to hear them, so why bother measuring them?

Books have been written.....

Koz