This article is the second in a three-part series on methods for song analysis. The article describes my approach for transcribing music from audio to score, the equipment I use and some of the challenges involved in the process.
What is a ‘transcription’?
Transcription is essentially the process of notating a song or a noise/sound which was (for whatever reason) previously unnotated*.
It is not common practice for pop songs to be released with supplementary scores when they enter the charts. However, in some cases, publishers eventually release officially licensed piano/vocal/instrumental reductions of a song, if it was considered to be particularly successful.
Simplified reductions are by no means a bad thing – the song is ultimately being made accessible to a broad spectrum of people who want to learn how to play it. Unfortunately, these reductions are generally arranged to suit a specific type of ensemble or they are significantly edited to make the song much easier to play by the majority of consumers. In this case, loss of detail is sadly inevitable.
* Transcription should not be confused with Transposition, which is the rearrangement of a musical score up or down in pitch by a consistent interval.
Why are transcriptions useful?
Providing that they are accurate, audio to score annotations can help to highlight points of interest or substantiate claims made during a musical analysis. Going beyond academia, they are also useful as a foundation for writing arrangements of songs, which better suit your performance requirements without compromising on detail.
Once you get the hang of it, writing a transcription is immensely rewarding. It helps develop your ability to identify pitches and rhythms on the fly, which can in-turn help to contextualise aspects of music theory, as well as help with working on orchestration projects or even writing out original compositions. Furthermore, just the act of reading and understanding a musical score whilst listening to a song can expose a totally new ‘dimension’ to something you may have thought you already knew inside-out.
That said, getting started with transcription requires a lot of patience. Transcribing from audio to score can be difficult and time consuming – I have been practicing transcription for about 13 years and often get delayed by issues stemming from poorly-edited vocal mixes (drowning out the backing or being drowned out by the backing!) to establishing key-signatures in heavily augmented extracts.
However, I also find the challenge and the practice highly therapeutic – if you’re just looking for a relaxing form of entertainment, transcriptions make an excellent (albeit challenging) alternative to cross-word puzzles!
What you will need
- 5-line Manuscript Paper (can be found at most good stationary stores)
- (optional) computer: I regularly use ‘Audacity’ for audio playback.
- (optional) notation software: Sibelius/Finale Notepad/Lilypond
- (optional) midi-keyboard/musical instrument capable of playing chords
- (optional) speed-editing software: I recommend ‘Amazing Slow Downer’
- (optional) good quality headphones or speakers
Feel encouraged to try… but please manage your expectations
The approach below is a great exercise for people wanting to improve their sense of rhythm and pitch. If you’re new to reading and writing musical notation, please ‘have a go’, but I urge that you manage your expectations and don’t try to take on everything at once (start small, pick out distinct 1 or 2-bar single-line melodies or riffs and once you’re confident with them progressively ‘expand’)…
that’s all very well and good… but… ‘how’?
Start small and progressively ‘expand’.
The key to doing a transcription is knowing ‘where to start’ and ‘where to look’. As a general rule of thumb, I start by picking out the section I want to focus on (usually the intro/verse or chorus!) and start to jot it down by no more than 4 bars at a time:
- Find the timecode for the section of the track using a ‘BTS diagram’, then focus on it with your music software (audacity/iTunes/vlc etc) and play it back several times
- Don’t try to write everything all at once(!) – focus on either picking out the main melody (tune) or the rhythm to begin with. If you’re new to this, be prepared for a lot of trial and error. If it helps, I suggest playing along with your chosen instrument to aid in accurately pitching the notes.
- When writing extracts for an analysis article I always acknowledge the key signature in the ‘track info’ section but as a rule I always transcribe a score without a key signature written on the staves. Instead I use accidentals (sharps/flats) alongside the relevant notes in order to make it absolutely transparent which notes are being played at any given time.
- Eventually you should have a several bars of melody and rhythm sketched out like below:
(fig A – vocal melody to verse 1 of Lee Jin Ah’s ‘Random’)
- If you have notation software like Sibelius or Finale, you can conveniently check your accuracy by playing back the extract. If not, you will be reliant on your musical intuition.
- Once you are satisfied with the rhythms and pitches, move on from the ‘melody’ to the next most identifiable element of the same extract (tip: it’s usually the bass line) and repeat the process all over again, now focusing on that new element (I appreciate that the process of physically focusing on a different aspect of an audio extract is easier said than done when starting out, but it definitely gets easier with practice)
(fig B – picking out the bass lines)
- Now, try identifying the backing harmony (chords) – In order to keep this article simple, I’ve picked out a relatively straightforward passage. It’s plausible that you’ll eventually come across songs where there are chord switches on nearly every beat(!). That said, the passage below is not without complication – the full chord symbols include directions like adding an additional 2nd degree or a sustained 4th or omitting part of the chord to achieve the exact effect heard on the recording. I don’t recommend getting bogged down with this level of detail when starting out. Instead, try to identify the basic chord in it’s simplest form and then, when you are confident, ask yourself… what’s missing? What position/inversion is the chord actually in? etc.
(fig C – picking out the harmony)
- Finally, if you’re confident enough, try and move on to that twiddly ‘fast sounding’ embellishment which rudely interrupts the vocals. If you’re lucky with your choice of song, the interjection may just copy a simple scale or arpeggio. If you’re less lucky, the notation will be ambiguous and you’ll need to either listen to the passage a few times or use software to slow the audio down. More on that later.
(fig D – the complete passage)
- All that remains is to repeat the process throughout the song! This will seem overwhelming at first but don’t lose courage! Many popular songs feature frequently recurring elements with only minor differences on repetition. This makes scoring a song a little less time consuming and reduces your workload.
Generally speaking, unless the song is exceptionally enjoyable, I tend to only transcribe the ‘main sections’ of a track (If I notated every repeating section, the analysis articles would be needlessly long, take ages to write and only offer fractionally more information!).
A brief overview of the software I use…
(fig E – screenshot of Audacity)
Audacity is (in my opinion) one of the most useful audio applications freely available on the internet. I use it for everything from audio playback to voice recording and even extracting and editing audio from video files. The broad choice of output formats and compression rates are convenient and the software is generally very easy to pick up. Although some may find the level of detail excessive, I personally love that you can pin-point the audio playback position to the exact thousandth of a second – Although that may not be particularly useful for BTS diagrams or transcription, it is useful for aligning multiple waveform tracks for whatever reason.
Amazing Slow Downer (free trial):
The instrumental passages in Lee Jin Ah’s ‘random’ and Dreamcatcher’s ‘Fly High’ are both good examples of where the music is either too fast or simply too challenging to transcribe accurately whilst relying solely on the ‘naked ear’ and musical intuition.
(fig F – Instrumental solo A – Lee Jin Ah’s ‘Random’)
In such circumstances, I rely on another piece software called ‘Amazing Slow Downer’, which significantly reduces or increases the speed of an audio extract without altering the pitch. It’s an invaluable tool for mapping out tricky passages, although I don’t recommend writing out a whole track using this software as it would take a considerably long time to do so! It’s also a really good tool for musicians who want to learn to play a new song with a backing track – the software allows you to slow it down whilst you learn and then improve at your own pace.
(fig G – screenshot of ‘Amazing Slow Downer’)
Sibelius (free trial):
The two big names in the notation software industry are Sibelius and Finale. I have used Sibelius products almost every day for about 20 years and cannot imagine a day in music without the software. It’s essentially the musical equivalent of Microsoft Word! That said, I know that some people are similarly entrenched in Finale Notepad. I’m not here to say one is better than the other but I’ll discuss the benefits of notation software in general:
The score playback feature is invaluable when previewing and confirming that your transcriptions are accurate. I also like the ease of plugging in an external midi/usb keyboard and ‘playing’ something straight into the software. Just about every aspect of the score presentation is customisable and it’s easy to become fluent fast. I’m wary of sounding too much like an ad.
If you are currently looking into notation platforms or which one to go for, I’d recommend that you try both free trials. Separately. That way you have at least 2 months of access to a good-quality notation platform without parting ways with any cash.
(fig H – screenshot of Sibelius 7.5 ‘in action’)
A brief discussion about hardware…
As suggested throughout the article, getting started with transcription does not need to be a costly affair at all. You can get started with just a pen, paper, passion and some patience! If you want to prop up the production values, rest assured that you can still produce most of what you see in my analysis articles without spending a single penny.
The following hardware is not essential at all – However, I’ll discuss the benefits…
- A Razer ‘Naga’ Gaming Mouse – or any mouse with multiple assignable hotkeys around the thumb-rest. I’ve assigned the numeric side keys on mine to correspond with the keypad controls on Sibelius (this saves a lot of time when editing and reduces physical movement – a good ergonomic solution!)
- Quality headphones/speakers/studio monitors – One of the biggest hurdles when transcribing a song is physically isolating the different lines whilst listening to them in the final mix! Some people prefer headphones, some prefer speakers – try to use hardware which delivers good audio balance overall (by that I mean don’t fall for the trap of buying something which only sells on the sole merit of having heavy bass. It really won’t help you!)
- USB/Midi Piano Keyboard – not an essential aspect of transcribing a score, but certainly helps to speed up the process. I use a keyboard for double-checking chords (before writing them down), working out the fingering for some piano/keys interjections (if the passage is particularly challenging – see the header image of this article!) and for quick-checcking particular aspects of the draft transcription score without having to manually playback the entire passage
Sadly, you cannot master transcribing a song overnight. It takes a lot of practice and dedication, but there are so many benefits that come with learning to do so. I hope my article gives a few good pointers to help with getting started, although, I am conscious that the guide leaves a glaringly logical question unanswered…
How do I actually use the transcriptions to support an analysis?!
More on that next time…