<< Chapter < Page Chapter >> Page >

The final step in the project was to test the algorithm we had created so we went ahead and conducted a series of tests that would evaluate mostly correctness but also, to some extent, performance.


First, we wanted to test to make sure that our algorithm was working properly. To do this, we attempted to match short segments of the original song (i.e. “noiseless”, actual copies of the library songs) of approximately ten seconds in length. The table below shows how these original clips matched. The titles from left to right are song segments, and titles running from top to bottom are library songs. We abbreviated them from the original, so they would fit in the matrix. The original names are “Stop this Train”, by John Mayer, “Semi-Charmed Life”, by Third Eye Blind, “I’ve got a Feeling” by Black Eyed Peas, “Love Like Rockets”, by Angels and Airwaves, “Crash Into Me”, by Dave Matthews Band and “Just Another Day in Paradise”, by Phil Vassar.

This matrix shows the match score results of the six noiseless recordings made from fragments of songs in the database, each of them compared to all songs in the database

The clear matches with highest scores can be seen along the diagonal. Most of these are close to 1, and each match meets our criteria of being 1.5 times greater than the other scores (comparing horizontally.) This was a good test that we were able to use to modify our algorithm and try different techniques. Ultimately, the above results showed that our code was sufficient for our needs.

We then needed to see if our code actually worked with real world (noisy) song segments. Songs were recorded on an iPhone simultaneously with various types of noise as follows: Train- low volume talking, Life- loud recording (clipping), Crash- typing, Rockets- repeating computer error noise, Feeling- Gaussian noise (added in Matlab to wav file), and Paradise- very loud talking. There were two additional songs we used in this test to check for robustness and proper matching. One is a live version of Crash, which includes a lot of crowd noise but does not necessarily have all the identical features of the original Crash fingerprint. The other additional song, “Yellow”, by Coldplay, is a song that is not in our library at all.

This matrix shows the match score results of the six noisy recordings made from fragments of songs in the database, plus a live version of a song in the database and another song entirely not in the database

Again, the clear matches are highlighted in yellow along the diagonal. The above results show that our algorithm can still accurately match the song segments in more realistic conditions. The graph below shows more interesting results.

This plot is a visual representation of the results matrix seen above


As before, the matches in the first six songs (from left to right) are obvious, and Yellow does not show any clear correlation to any library song, as desired, but the live version of Crash presents an interesting question. Do we actually want this song to match? Since we wanted our fingerprinting method to be unique to each song and song segment, we decided it would be best to have a non-match in this scenario. However, if one observes closely, it can be seen that the closest match (though it is definitely not above the 1.5 mark) is, in fact, matching to the original Crash. This emerges as a small feature of our results. This small “match” says that although we may not match any songs in the library, we can tell you that this live version most resembles the original Crash version, which may be a desirable outcome if we were to market this project.

We were amazed that the final filter could perform so well. The idea of completely ignoring amplitude information in the filter came from the paper by Avery Li-Chun Wang, one of Shazam’s developers. As he mentions, discarding amplitude information makes the algorithm more insensitive to equalization. However, this approach also makes it more noise resistant since, since what we do from there on basically consists of counting matching peaks versus non-matching peaks. Any leftover noise will count very little towards the final score, as the number of peaks per area in the spectrogram is limited by the thresholding algorithm and all peaks have the same magnitude in the filter.

Questions & Answers

anyone know any internet site where one can find nanotechnology papers?
Damian Reply
Introduction about quantum dots in nanotechnology
Praveena Reply
what does nano mean?
Anassong Reply
nano basically means 10^(-9). nanometer is a unit to measure length.
do you think it's worthwhile in the long term to study the effects and possibilities of nanotechnology on viral treatment?
Damian Reply
absolutely yes
how to know photocatalytic properties of tio2 nanoparticles...what to do now
Akash Reply
it is a goid question and i want to know the answer as well
characteristics of micro business
for teaching engĺish at school how nano technology help us
Do somebody tell me a best nano engineering book for beginners?
s. Reply
there is no specific books for beginners but there is book called principle of nanotechnology
what is fullerene does it is used to make bukky balls
Devang Reply
are you nano engineer ?
fullerene is a bucky ball aka Carbon 60 molecule. It was name by the architect Fuller. He design the geodesic dome. it resembles a soccer ball.
what is the actual application of fullerenes nowadays?
That is a great question Damian. best way to answer that question is to Google it. there are hundreds of applications for buck minister fullerenes, from medical to aerospace. you can also find plenty of research papers that will give you great detail on the potential applications of fullerenes.
what is the Synthesis, properties,and applications of carbon nano chemistry
Abhijith Reply
Mostly, they use nano carbon for electronics and for materials to be strengthened.
is Bucky paper clear?
carbon nanotubes has various application in fuel cells membrane, current research on cancer drug,and in electronics MEMS and NEMS etc
so some one know about replacing silicon atom with phosphorous in semiconductors device?
s. Reply
Yeah, it is a pain to say the least. You basically have to heat the substarte up to around 1000 degrees celcius then pass phosphene gas over top of it, which is explosive and toxic by the way, under very low pressure.
Do you know which machine is used to that process?
how to fabricate graphene ink ?
for screen printed electrodes ?
What is lattice structure?
s. Reply
of graphene you mean?
or in general
in general
Graphene has a hexagonal structure
On having this app for quite a bit time, Haven't realised there's a chat room in it.
what is biological synthesis of nanoparticles
Sanket Reply
what's the easiest and fastest way to the synthesize AgNP?
Damian Reply
types of nano material
abeetha Reply
I start with an easy one. carbon nanotubes woven into a long filament like a string
many many of nanotubes
what is the k.e before it land
what is the function of carbon nanotubes?
I'm interested in nanotube
what is nanomaterials​ and their applications of sensors.
Ramkumar Reply
how did you get the value of 2000N.What calculations are needed to arrive at it
Smarajit Reply
Privacy Information Security Software Version 1.1a
Berger describes sociologists as concerned with
Mueller Reply
Got questions? Join the online conversation and get instant answers!
Jobilize.com Reply

Get the best Algebra and trigonometry course in your pocket!

Source:  OpenStax, Digital song analysis using frequency analysis. OpenStax CNX. Dec 19, 2009 Download for free at http://cnx.org/content/col11148/1.1
Google Play and the Google Play logo are trademarks of Google Inc.

Notification Switch

Would you like to follow the 'Digital song analysis using frequency analysis' conversation and receive update notifications?