Curated by Francesco Prisco
Everyone has happened to not know or not remember what was the title of a song heard in a pub, in a car or on the street in passing. Thanks to technology, this too is now an outdated problem . Today it is enough to have an app such as SoundHound or Shazam on our devices for song recognition and even the most unknown song will become part of our playlists. But is it always like this?
SoundHound: You can hum songs to recognize
True to its name, SoundHound (hound means hound in English) scans its music databases in search of the song we want. Founded in 2005, the American company has always paid great attention to the innovation of its services. With SoundHound you have the possibility to whistle or hum a song to find out the title and this demonstrates a great ability of “audio and speech recognition”. Also, for some time now the app has been integrated with the voice assistant called Houndify. By simply saying “OK, Hound” you can play a song on Spotify, Apple Music or similar services. SoundHound is the industry’s first app available on wearable devices.
Using SoundHound’s voice assistant, you can ask us to add songs to a Spotify playlist, play music videos through YouTube and finally dissolve our musical doubts (which album a song belongs to, when an album was released or who are the members of our favorite band).
Shazam: You can also buy your favorite songs
Shazam is a company with a very long commercial history: a service born in 1999 and then acquired by Apple in 2017, which gives the possibility to identify a piece of music by listening directly to the aforementioned through the microphone integrated in smartphones. In full compliance with the trends of recent years, today Shazam allows not only to identify a song but also to buy music on digital platforms, share our discoveries with other users or deepen the biography of the artists.
As mentioned, Shazam was acquired by Apple, but fortunately for the vast majority of mobile users the app is also available for gadgets and Android devices .
SoundHound vs Shazam for song recognition: costs
Shazam is presented on the stores with two apps Shazam and Shazam Encore. The latter was, until recently, the paid version without ads. But since the ads were removed from both versions, the two apps look pretty much the same . It’s unclear why Apple, which owns the company after its purchase in December 2017, still keeps them online.
SoundHound also appears online in two versions. One is free, the other at a cost of € 7.99 for Apple devices and € 5.49 for Android devices. Basically the difference between the two versions is that the free version has a limited number of monthly searches .
SoundHound vs Shazam for song recognition: reliability
As for the reliability of the search for the desired song, the two apps promise a lot. SoundHound is renowned for finding songs even whistled and hummed. The function was tested for a few hours in a row: it will be that I am not at all skilled in humming or whistling but I was not even recognized a song.
As for the databases of the two services, I think that of Shazam is more extensive. I made both apps listen to a Neapolitan neo-melodic song, just to pick up a lesser known one. Well, Shazam immediately managed to trace the title of the song unlike her colleague who, once again, failed miserably .
As for listening to songs, SoundHound also allows the recognition of the songs you listen to with headphones, Shazam instead requires the use of the device’s microphone. The blue app is not for use in places where you might be disturbed by loud music.
SoundHound vs Shazam for song recognition: how do they work?
All services like Shazam or SoundHound offer a truly unique functionality. Their operation is essentially based on the analysis of a spectrogram .
You can think of any piece of music as a time-frequency graph called a spectrogram. On one axis of this graph there is time, on the other the frequency, and on the third the intensity. Each point on the graph represents the intensity of a given frequency at a specific point in time.
Therefore, the algorithm creates a fingerprint of a song , returning a 3D graph like the one shown in the figure, managing to identify the frequencies thanks to the intensity of the peaks. For each peak point, Shazam keeps track of the frequency and time elapsed since the start of the song.
The algorithm doesn’t just mark a single point in the spectrogram. Hence, the key is not just a single frequency, but is a hash of the frequencies of the dot sequence. This means, except for hash collisions, that every time you search for a song the speed with which it is given as a result increases allowing the algorithm to take advantage of the constant search time (O (1)).