Microsoft Songsmith – why?

I love music, really, it’s as important to me as food, air, or my Apple Mac. At age 10 I thought it was what I’d end up doing with my life. My job during my undergrad years was playing in a band, so I could say (tentatively) that I’ve been a professional musician at some stage in my life. However, I’m well aware of my limits, and I recognize that I don’t have enough talent to make a living performing. Having at least an approximate grasp on your limits is important, will I compete in the Tour de France this year, enter X-Factor or start a new career in pro-wrestling? The answer to all is sadly no.
Nowdays I’m an academic, I work on new approaches that helps make video games better, and I understand the difference between an application that is intended for research purposes and one intended for commercial release. It appears that Microsoft don’t however.

Songsmith is a ‘music’ application that came out of Microsoft Research Labs, but I just can’t work out why it’s made it to commercial release. My first exposure to the app was not the terrible muzak that it generates (more on that later), but the publicity ads surrounding the product. Scoble interviewed the two Microsoft engineers (see video below), and like most of the videos covering Songsmith, I found this terribly painful to watch. I felt like a kid watching Doctor Who all over again as I had to watch this video with my hands covering my face, peeking through my fingers now and again. The key difference being that Doctor Who made me scared, Songsmith induced agonizing embarrassment. If there had have been a sofa in my room, I would have been behind it.

Let’s rewind to the beginning. Songsmith is an app that let’s you sing into your computer’s microphone, it analyzes the recording and auto-generates a backing track according to some preferences that you set.

UI and Interaction
I feel that Songsmith does everything wrong. Everything. If I have an idea for a song in my head, I want to get that idea into the computer as quickly as possible. Not so with Songsmith. On launching the app you have to choose your style of song, then choose the BPM. How on earth anyone knows the BPM of what they want to record is beyond me. One of the ways of interacting with the UI is using a jog-wheel or dial approach, this made me scream out loud while trying to use it. In the Scoble video one of the engineers talks about how easy it is to use, I sense he’s toying with us.

Step 1 - choose your song style, but avoid the dial

Why you need two methods of interaction, both jog-wheel and the cursors is beyond me, the app already looks terribly old-fashioned, do they think dials are jazzing it up? After choosing the song style you have to then choose your BPM. I laughed when I first saw this, then shouted, then shouted some more. Both the UI and interaction for setting the BPM is nothing short of bewildering, none of my HCI students would design something as bad as this.

Step 2 - choose your BPM, if you know, care, or actually are able to.

In the choose BPM dialog, again there are two methods of interaction, the dial or cursors. There is almost no reason I can think of for having the dial here, it is very confusing to understand and control, and will almost certainly slow down the process. Nor does it add to the user experience, so how this made it past user testing is a mystery.
As for the interface itself, we can see there are two markers on the outside of the dial, is it obvious what these are? These are the maximum and minimum values of BPM, the value at the 12-oclock position is the min and the one at 9-oclock is the max. Obviously they don’t state what these values are, nor is obvious to me how the values should be chosen using this method. Is there not a common UI component that already exists that defines a numeric range and makes it easy to set a particular value? Oh yes, it’s called a slider and is also commonly found on music interfaces. Far too usable apparently, best to choose the infuriating option which makes users guess how it works.

Laying it down
We finally arrive at the recording window feeling enraged and have probably forgotten the tune that we were thinking of. Regardless, let’s record something. The first few times I hit record I wasn’t able to start in the right place, the count-in completely confused me. Once that was mastered I found singing-along to only a metronome completely, well, uncreative. After recording my vocals I was finally able to listen back to the joy of the auto-generated accompaniment. The final output is torturous, nothing else can describe it. The output seems to be using MIDI which was a protocol designed to primarily control audio equipment, and has as much potential for expressing emotional sound as Yoko Ono. Everyone else has mostly abandoned this approach for generating music, or do it very well and combine it with high quality samples (Garageband).

With vocals recorded, I explored the rest of the interface. I noticed a button labelled “Undo [1]” and when clicking it, it erased my vocals – with no ability to redo. Genius. I would suggest a better approach may have been to, oh I don’t know, how about actually tell the user in plain English what they’re undoing, e.g. Undo vocal recording? Programs like Adobe Lightroom tell the user clearly what steps they’ve performed and which they can roll back to, again MS take the ‘let’s make it completely confusing’ approach, which is always best for novices using these programs.

I don’t even know who this app is aimed at. The promotional videos show young kids but when I was using it, quite complex chords appeared, why would they care what a C#sus4 chord is?

Do potential users know or care what the chords mean?

Coda
So there we have it, Songsmith is an app which is infuriating to use, shows music syntax that is meaningless to novices, makes them answer pointless questions before they can record, will never be used by musicians, and the output would be terrible even 20 years ago. I doubt anyone on the dev team was either musical or had any awareness of usability. In the Scoble video they mention user testing, frankly I find this hard to believe, or another possibility was that the user testing wasn’t even conducted properly (i.e. our other engineers with no musical ability loved it too).

I suggest before this research team release any more products they ask themselves one question, “What would Steve say?”.

Are video games disposable?

Being born in the early ’70’s, my first exposure to music was vinyl being played on my parents radiogram. I then remember getting my own personal music player, a reasonably large cassette player when I was around 7 or 8, followed by a Walkman at the age of 11. In 1990 I got my first CD player, eventually moving to a mini-disc player in the mid-90s.
Since discovering MP3s and my first iPod in 2003, things have never been the same. Or have they? Despite the changing media format of the music, I am still able to enjoy Fat Bottomed Girls by Queen on my iPhone just as much as I did on my parents radioogram in the late 70s.

The same is true of movie formats. Over the years I have been able to watch movies on VHS, DVD, Blu-ray or iTunes digital download, but again, I am still able watch a movie via iTunes which I originally saw on VHS in the early 80s.

However, this is not true for video games. This week I had a yearning to play Cannon Fodder that I originally played on an Amiga 500 in the late ’80s, but I couldn’t think of how to do so. My first thought was to install some sort of software emulator, but getting these working can be difficult, and I’m not exactly clear on the legal issues of doing so. Perhaps I could buy an Amiga computer and find a copy of Cannon Fodder on eBay? Forget it.
It seems that there is a vast history of video games that the current generation of gamers are simply denied access to. Yes, some older games are remade, but very few are available today, legally, in their original form.

What if today’s musicians were denied access to the likes of Bob Dylan, The Beatles, Elvis or Rolling Stones and instead all they had access to was Westlife and the Cheeky Girls? Why are video gamers forced to live only in the present, could we not also enjoy our past?