Shazam, Soundhound And Google Search Music Identification Comparison

Do you know what the name is of the playing song? Which app should you use to identify its name? Which should you use for humming and which for direct playing?
Talking about music identification, I’m reminded the web-based service, midomi. As a forerunner of music identification, I am very grateful that I had knew it before I missed those songs you love at first listening (Midomi is known as SoundHound now).

With the rapid development of mobile phones, the regret of missing a song decreased dramatically. There are several apps that can tell you the name of a song being played, we will have a comparison among the 3 mostly known apps: Shazam, SoundHound and Google Search (for music identification only).[sc:cb]

UI

Shazam Main SoundHound Main Google Search Main

SoundHound, obviously takes the lead thanks to the recently polished UI, with a scrollable list of Song Stream, Featured Video, Featured Album, Popular, Hottest Charts, etc. ; Shazam has a main screen of simplicity, with just a big button for music identification; Google Search is not a dedicated app for music identification, therefore the element of music determination is greatly weakened.

Screenshot_2014-11-22-18-15-57

One thing to note, SoundHound has 3D transition effect. As mentioned before in Inbox by Gmail: an Early Test Drive,

“while the swift animation of loading and transition reminds you, the age of technical art has come.”

 

This is a good example that programmers are willing to learn some art and make their app much friendlier on the exterior. An essential key to success of an app is, you make it function well, you make it look good.

Comparison of identification

We’re not downloading these apps for no reason! Regardless of the UI, we care much about the overall performance of their capability in music detection. First thing first, there’re some technical details which power the apps. All of these three apps uses Acoustic Fingerprint to confirm a song, that’s very similar to how we find a person by physical fingerprint. But there’s also some optimization for respective apps.

Let’s find the pros and cons by testing them with real songs. Here I have songs from three countries: the largest country in population (China), the largest country in geographic sense (Russia) and the strongest country in general national power (USA). Various genres of songs are provided, and the time of identification is recorded.

Warm up

For the first test, I choose the Anthems of these three countries. If any app fails to tell the name, then it’s a tremendous joke.

Test is done three times, and if any attempt succeeds, then the result will be a “Success”. Average time will be the final score. Time of all failed attempts are colored red.

Anthem of China (Instrumental)

Performed by Military Band

[table]

,Shazam,SoundHound,Google Search
Result,Failure,Success (3/3),Failure
Time 1,12.201,15.992,18.153
Time 2,12.763,14.779,17.429
Time 3,11.546,13.23,17.622
Average Time,12.17,14.667,17.735

[/table]

Anthem of Russia

Sung by Larisa Dolina (Лариса Долина)

[table]

,Shazam,SoundHound,Google Search
Result,Failure,Failure,Failure
Time 1,11.914,15.325,17.382
Time 2,11.327,17.816,17.793
Time 3,18.895,14.64,21.046
Average Time,14.045,15.927,18.74

[/table]

Anthem of USA

Sung by Jennifer Hudson

[table]

,Shazam,SoundHound,Google Search
Result,Success (3/3),Success (3/3),Success (3/3)
Time 1,5.597,13.215,8.045
Time 2,4.105,4.632,7.753
Time 3,5.773,5.6,8.01
Average Time,5.158,7.816,7.936

[/table]

Comments

I have no comment on the result.

Hits

Next we test hits in local area.

Hits in China

Songs for evaluation:

  1. Chopstick brothers – Little Apple (筷子兄弟 – 小苹果)
  2. Pushu –Ordinary Way (朴树 – 平凡之路)
  3. Soda Green–Tiny love song (苏打绿 – 小情歌)

[table]

,Shazam,SoundHound,Google Search
Result,Success (3/3),Success (2/3),Success (1/3)
Time 1,5.385,4.295,19.794
Time 2,3.931,15.889,7.753
Time 3,12.568,6.805,21.414
Average Time,7.295,8.996,16.32

[/table]

Hits in Russia

Songs:

  1. Mumiy Troll – Vladivostok 2000 (Мумий Тролль – Владивосток 2000)
  2. Lyapis Trubeckoy – Warriors of the Light (Ляпис Трубецкой – Воины света)
  3. Dance Minus – The half (Танцы Минус – Половинка)

[table]

,Shazam,SoundHound,Google Search
Result,Success (3/3),Success (2/3),Success (2/3)
Time 1,10.18,5.429,8.548
Time 2,8.026,8.976,17.718
Time 3,11.754,15.393,15.804
Average Time,9.987,9.933,14.023

[/table]

Hits in USA

The recent hits:

  1. Lady Antebellum – Bartender
  2. Alexi Murdoch – All My Days
  3. Blake Shelton – My eyes

[table]

,Shazam,SoundHound,Google Search
Result,Success (3/3),Success (3/3),Success (3/3)
Time 1,8.438,10.288,7.852
Time 2,8.082,3.909,6.062
Time 3,2.72,6.587,10.746
Average Time,6.413,6.928,8.22

[/table]

Comments

For hits in the USA, all apps performed well. Shazam is the best one who recognized all hits in all 3 countries, congrats! Google Search performed worst, having the longest time and also lowest recognition rate.

Folk songs

This section contains folk, classic and country songs, not strictly limited to “folk”

Chinese folk

  1. Tracy Huang – Jasmine (黄莺莺 – 茉莉花)
  2. Tong Li – Lullaby in North-eastern China (童丽 – 东北摇篮曲)
  3. Song Zuying – Dragon Boat Melody (宋祖英 – 龙船调)

[table]

,Shazam,SoundHound,Google Search
Result,Success (1/3),Success (1/3),Success (1/3)
Time 1,4.507,30 (Timeout),13.777
Time 2,11.88,22.711,21.047
Time 3,11.399,15.81,17.135
Average Time,9.262,22.84,17.32

[/table]

Russian folk

  1. Chizh and So – Forever Young (Чиж & Co – Вечная молодость)
  2. Mill – The dream road (Мельница – Дорога Сна)
  3. Alice – Route E-95 (Алиса – Трасса Е-95)

[table]

,Shazam,SoundHound,Google Search
Result,Success (3/3),Success (2/3),Success (2/3)
Time 1,11.163,5.134,5.17
Time 2,5.371,14.722,7.55
Time 3,4.52,14.525,21.091
Average Time,7.018,11.46,11.27

[/table]

American folk

  1. Ugly Kid Joe – Cats in the Cradle
  2. Justin Timberlake. Carey Mulligan & Stark Sands – Five Hundred Miles
  3. Jack Johnson – Talk of the town

[table]

,Shazam,SoundHound,Google Search
Result,Success (3/3),Success (3/3),Success (3/3)
Time 1,4.249,8.731,10.841
Time 2,2.753,7.478,4.904
Time 3,4.931,3.652,10.763
Average Time,3.978,6.62,8.836

[/table]

Comment

The three Chinese folk songs are not all recognized by a single app, which was a bit disappointing, SoundHound especially failed me for letting me had 5+ attempts. The timeout result is rather reasonable and I would say SoundHound deserves more penalty. In Russian folk song test, Shazam is again the winner. In American folk test, all finished well, Shazam finished in amazing 3.978 secs in average. That was truly impressive.

Humming and singing

In this section, I sing songs from these three countries. For now only SoundHound supports finding a song by humming or singing, so I am just testing this app only.

Result

Chinese: Eason Chen – Long time no see (陈奕迅 – 好久不见), Success

Russian: Basta – Mama (Баста – Мама), Success

American: Fools Garden – Lemon Tree, Success

Summary

As we can see from the sheet, Shazam overpowered the other two apps in direct listening tests. It has advantage not only in the USA, but it’s also quite a multinational app for music identification. It, though, has simple UI, but responds to user’s query fast, sometimes in an amazing degree.

SoundHound earned my favorable impression by its humming/singing detection. If unfortunately you didn’t catch the song in time with your phone, but you still remember the melody, SoundHound is the last hope for you. However, during the test, SoundHound has the least stable connection to the server. Many times had it notified “Unable to connect to server”. It also has limited number of signature of songs, therefore you possibly can’t find the song you want. Also, the new version looks fancy, but slows down as we test it.

The performance of Google Search is the poorest among 3 apps I’ve tested. However, it can be the quickest one to launch. By activating “From any screen” and “From lock screen” function, you can open Google Search in a flash. Just say “Ok, Google” followed by “What’s this song?” and you’re detecting the song name.

I’ve made a table so you can decide which app you want for song exploration under certain circumstances.

[table]

,Shazam,SoundHound,Google Search
Recognition Speed,Fast,Ok,Slow
Song repo,Large,Fair,Fair
Humming/Singing Support,No,Yes,No
Activate convenience,Good (By one-tap listening on opening),Good (By shortcut on launcher),Best (By Google Now)
Sharing support,Yes,Yes,No
Overall Efficiency,Good,Bad,Good

[/table]

Music has no language barriers. You can be alone sometimes, but music is always there for you. Missing a music is like missing a partner. Grab your phone, keep the music playing, and get one or more apps above.

Without music, life is a journey through a desert. – Pat Conroy