Updated: August 20, 2016
As you all know, I am not too keen on mobile use. It's not because I reject technology, change and such, it's because I embrace comfort and efficiency, and working on a mobile device is as efficient and comfortable as pleasing yourself while wearing oven gloves. However, there are exceptions.
One such is when a bunch of readers email you and swear by their unborn children that there's this app what does wonders, and may you please test it, pretty please. The name of the game is SoundHound, and it comes in two flavors. The first, Hound, is a voice search assistant, which was likened to Cortana, and which we will explore separately. The second one is a namesake tool for guessing and finding music for you, in case you don't know the lyrics and whatnot. Well, readers swear, I test. As simple as that.
Bell the doggies
Toll the Hounds would be a more appropriate title, but 'tis a name of a book in the Malazan empire, so we can't use it. Anyhow, I installed the application on my test S5 device, which I rarely use for serious work, as I'm not a big fan of Android. Then, while I waited for the Play store to do its serious byte crunching, I was thinking of all my previous attempts to use would-be artificial intelligence to guess my music. Cloud-based distributed harmonics analysis is not as fancy as it sounds despite the hype and fad, and I've always had different results from the generic praise-per-click reviews out there.
I had previously tried the likes of Shazam, Spotify and others, and they all failed me so miserably I didn't even get to the point where I'd save a few screenshots and show what they can do. Since you're reading this article, it seems that SoundHound managed to get past the initial barrier.
The interface is simple and riddled with ads. I can't shake the impression that the app is designed to gently goad you toward searching, exploring and buying music. As it happens, when it missed, and we shall elaborate on that in a jiffy, the alternative offers always had a rather here-and-now spin. But let's focus.
Let us sing!
I decided to hum and sing a bunch of fairly popular and then rather obscure songs, some with lyrics and some without, to see whether SoundHound can actually do some good. The first few searches were not really encouraging. The samples included me whistling and grunting William Tell by Rossini and Ballade Pour Adeline by Clayderman. Then, the app suddenly recognized my interpretation of Morning Mood by Grieg.
After that, I went 80-90s hiphop with a few trashy but popular singles, and the app missed all of them, giving me modern stuff instead. C'mon, Bailando by Paradisio is such an easy track to detect. Ah, well.
Ramping it up with real music
The best way to evaluate the tool is to actually play real music and see what it can detect. My own singing is lame at best, but you can't really argue with a laptop playing Youtube clips. Which is what I did next. Now, you should take into account various parameters like distance and position of the audio source - laptop speakers - and the S5 mic, sound quality, sample rate, and such. But overall, it should be 100x better than me.
Again, I had rather mixed results. Identifying modern and popular music is trivial. But not so when you go international with some rather obscure but valid searches. My choices of the day include Sandu Ciorba, DJ Zeki, Jozin z Bazin by Ivan Mladek, Chi Mai by Ennio Morricone, and My Name is Trinity by Annibale E I Cantori Moderni, the last being the theme song of the super cool Italian-dubbed-English western with Bud Spencer and Terence Hill. I added some more classic music too, like The Toreador Song from the opera Carmen, and a few others. As random and international as I could think. One of the selected songs is a remix of a Kazakh version of the Finnish Freestyler hip hop single.
Anyhow, the SoundHound did the following: It identified Carmen well, and it didn't have any issues with Ennio Morricone's music, either Chi Mai, although it identified it as a film music, or the theme song from My Name is Nobody. Again, film music.
Speaking of westerns, SoundHound did not successfully identify the Trinity track. It scored well with the techno music I used in my Turbo driving clip, but not the crazy remix in the awesome Megane RS 265 clip. Likewise, it didn't like Ciorba's Pe Cimpoi melody either. Then, it managed well with the quirky Jozin song. I welcome you to copy & paste the names and titles mentioned here and try for yourself with whatever music search tool you use. You might be amazed. Or something of that sort.
Problems
Well, there were some other issues with the app. One, it loads way too many ads, some of which are plain moronic. Two, it was telling me where the search was done, location wise, at the bottom of the screen. Why the hell would you tell me that? I know where I am. Perhaps I don't know the song, but I'm not an imbecile.
Conclusion
Well, there you go. Yes, I know, I'm not a typical smartphone user, and therefore, this review may be completely irrelevant to your needs. Furthermore, if your music taste revolves around whatever is on the radio or such, again, my review is probably useless to you, oh you young fellow with developing musical needs. Then again, if you want to know how advanced cloud music search thingie is, this article might be of some use.
Song popularity and country origin definitely count. Which is understandable. But don't expect any great miracles, because even if you can sing like a champ, there's still about 50% chance you might get it wrong, based on my short, limited but sweet testing. Add a bit of gently aggressive advertisement and subtle commercial hints, and you might not be all too pleased. All that said, SoundHound is more accurate than all other similar apps I've tried before. So there's progress. And hope. And that means we shall expand our testing to the voice assistant, too. It might actually work and not suck. For the time being, you might want to make a fool of yourself and sing to your smartphone. With SoundHound, there's some chance of minimizing the humiliation. 6/10. See you around.
Cheers.