Compression vs Rarefaction in Sound Waves

I am currently looking into solutions for Sound Classification, and I came across Ludvigsen's methodology (if anyone wishes to refer to it).

The problem is that a sample graph of amplitudes in one of his demos shows only positive amplitudes, while Sound Waves have both negative and positive amplitudes.

I have arrived to a point to know that positives are compression, while negative are rarefaction, but I still cannot quite comprehend the difference (and the importance of that difference) between the two.

1. Would it, perhaps, make sense to handle only the absolute values (distance from zero) of either? Or would that lose properties defining a Sound Wave through rarefaction?
2. Most importantly, how do some Software show Sound Waves (if they even are Sound Waves) with positive values only? Such as: Virtual DJ, most Equalizers.

PS: The values I am hoping to achieve are similar to the first figure on page 21 (Page 9 in the header) of this paper https://www.google.com.mt/url?sa=t&rct=j&q=&esrc=s&source=web&cd=3&ved=0CDIQFjAC&url=https%3A%2F%2Fwww.uzh.ch%2Forl%2Fdiss_michael_buechler.pdf&ei=MaK6U43pIa3B7AbJgIHYCQ&usg=AFQjCNHqqOwIamFVMTul8h6f2z9J4fD2gA&bvm=bv.70138588,d.bGE&cad=rja .

I am still in progress of identifying what the Envelope Level (dB) axis represent.

• What is Ludvigsen's methodology? A quick Google failed to find anything obviousky related to your question. Commented Jul 7, 2014 at 9:44
• Further to my original question, I "think" that my proper question should be: How do I construct a Sound Wave Histogram out of a sound wave? But I am still slightly unsure if 'Sound Wave Histogram' is what it is called. Commented Jul 7, 2014 at 17:12
• Btw I have added a link to a paper in English referencing the same Ludvigsen, I did not link the original paper as it is in German, and think it best not to rely on online translation for now. :) Commented Jul 7, 2014 at 17:15

Google didn't immediately come up with anything significant for "Ludvigsen's methodology", but let me give this a shot nonetheless.

Sound is a propagating pressure wave. So as it goes by, the pressure increases, then decreases, then increases again, etc. Pressure increasing means the particles in the material (typically air) are closer together for some time. This is visualized below for a lattice.

Where the lines are close together, pressure is higher. This is a single pulse, but for a continuous sound the areas of high pressure (compression) and low pressure (rarefaction) would just continuously alternate.

As for displaying this effect, a plot of the pressure at a given point vs. time will produce some sort of sinusoidal wave, like below. I assume this is what you've been seeing. Note this figure uses condensation instead of compression - they mean the same thing here.

The a similar but all-positive plot is likely the result of just choosing a different zero. Your intuition is telling you that the 'zero' of the plot should be where there is no compression or rarefaction, but you could just as easily take some other arbitrary pressure as the zero of the plot. For example, if you chose as your zero the lowest pressure observed (most rarefaction), then all other pressure values would be higher than that, resulting in an all-positive graph.

I don't think this is a question of only looking at absolute values since that would produce discontinuities (corners) for a sinusoidal curve.

• Thank you, your explanation has helped a lot. For your perusal, here is the Ludvigsen research I have mentioned earlier, but unfortunately it is in German. :) google.com/patents/EP0732036B1?cl=de Commented Jul 7, 2014 at 12:54
• Unfortunately my German is a bit rusty, but maybe edit this into your question in case anybody else wants to take a look! Commented Jul 7, 2014 at 13:14
• If you are further interested, page 22 of this pdf (9 of the original document) google.com.mt/… are what I am hoping to achieve, but I am unsure what the Envelope Values axis represent. Commented Jul 7, 2014 at 17:08

The pressure/vacuum model is a load of crap. Sound is correctly visualized as a solid beam connecting source to ear because that's what it is, a direct mechanical link being pushed and pulled. The inclusion of latency in any visual aide is ridiculous.