I'll give it a try. The graphs were made with Audacity, a cool freeware multitrack recorder with some nice features like effects and such.
The graph shows volume on the left and the played frequency on the bottom scales. I'll not try to explain the options in the lowest part of the picture, they're merely algorythms to display the graph in different ways using slightly different calculations.
In Audacity you can record a small bit of music and then apply these algorythms to it, the program will take a few seconds of the recording to build the graph so you can calculate the entire playing range of the guitar or just one tone.
A single singe tone like the low A in it's purest form (sine curve) will only produce a smooth peak around the 440Hz range (= 440 vibrations per second). The same goes for all the other tones. However a guitar produces harmonics (overtones) too which will display elsewhere in the graph (usually higher up the frequency scale).
The graph displays all the tones you can play on the guitar + the guitar's harmonic frequencies. These harmonics influence the original notes volumes with their own volumes, the same way when you throw a rock in water and another one right next to it.
Each rock creates it's own waves but when both waves intersect they influence eachother too.
A single pure note will produce a very smooth curve in one area, all the notes together + harmonics (notes generated by the guitar's properties) will cause a highly complex interaction in the volumes of each freq. band.
The graph shows an average of this interaction because it was applied on a longer recording in time, with all the notes played.
The GS recording was photoshopped so both scales match (Audacity auto adjusts the window depending on the volumes it detects), and one thing which is immediately clear is the difference in volume in the two models. In the frequency range there are obviously differences bcause these are two different guitars, but a quick glance shows that the frequency roll off is very simmilar (as one would expect from the same materials used).
A best scenario in the scale would be a straight line, meaning all the tones are the same in volume with a gradual roll off. Peaks in the graph show areas of extra amplification, notes which will sound louder because of the influence of the guitar. Dips in the graph are less loud notes because the guitar responds less good with these or produces unwanted harmonics. Everything above the highest playable tone (frequency) on the guitar are harmonics. The range of a guitar is from E4 to C8, this picture shows these notes frequencies.
With this knowledge it's easier to make a comparison when you don't trust your ears or the influence of the environment in which you're trying to determine which guitar sounds best.
There's obviously also a best situation for a recording but neither recording was done under studio conditions, it's just an average comparison of two Koa guitars.
The microphones used for the recordings will also somewhat influence the graph, the youtube GS suggests that the highs were boosted, the GA is much more even in that area. Also the C shows a peak throughout the spectrum, possibly the piece was played in C, normally the low C is the initial strike of the chord and therefore the loudest tone.
Hope that wasn't too hard to grasp, it's not easy being Dutch and explain things in English, my excuse if it's still gobbledygoock for some.
As it was impossible for me to travel 100ds of miles to get the guitar I wanted, I used this method to get somewhat of an idea on the guitar's behaviour, not too hard with a couple of good youtube recordings. Still it's the best way to use your hearing to get what you want.
Ludwig
