How I Transcribe

Over several months of transcribing Oliver Schroer tunes, analyzing them, and posting them here to my blog, I’ve had a couple of people ask how I do it. I use a several tools – most of them free – to help me, but the process is not all that complicated. Here's a quick summary of the steps I go through.

Before Transcribing

By the time I sit down to start a transcription, I’ll usually have listened to the track at least a few times and have a sense of what I’m up against. Before I start copying down notes, I’ll usually give the track one or two more close listens, with an ear for a few general features:

• Does the tune have a simple form, like AABB? Even if I can’t immediately grasp the form, can I tell when new sections begin, or when the entire tune start over? How many times is the tune repeated?
• What instruments are there? Are there several instruments that play the same line, are there countermelodies, and are there instruments for which I can just note that the instrument enters (e.g. a note above the fiddle line that “shaker starts here”)?

The Transcription Process

When I start to notate the piece, I mostly use VLC Media Player to listen to the track. It has a loop function, which I use occasionally but not often, and it also allows me to isolate one stereo channel if I need to (this feature has been useful for tracks like Horseshoes and Rainbows, where two different countermelodies are played in the left and right channels, or Enthralled, where the fiddle is more prominent in one channel and the flute in the other, allowing me to more easily pick out the details of either line).

For tracks with a moderate tempo, like Bright Eyes, I can usually transcribe the track at its original speed, but it’s often useful to be able to slow a track down. If I need to take a close listen to one specific moment, VLC’s playback speed feature usually does the trick, but for fast tunes, I’ll often use Audacity’s Change Tempo feature (making sure to select “Use high quality stretching”) and export the entire track at a slower speed.

For most tracks, I’ll reduce the tempo by about 30%. For a complex tune like The Humours of Aristotle, I might reduce the tempo by as much as 50%.

For tracks like The Hub of the Wheel or Toby’s Reel where I’m trying to transcribe a bass line, it can be useful to adjust the EQ for a track. This can be done in Audacity, but I usually do it in Logic Pro, since I can hear the adjustments I'm making in real time.

I use MuseScore to notate what I hear. If it’s a tune in a familiar style (e.g. Bright Eyes or The Hub of the Wheel) or if I’ve heard it many times before (anything from Jigzup), I may try to learn passages by ear on the mandolin and then notate them. But for many tracks, I’ll work through a full repetition of the tune, copying out a few notes at a time. Once I reach a passage that’s a repeat of something I’ve already written out, I’ll copy and paste the notes, and listen for variations. Depending on the tune’s length, complexity and familiarity, it can take anywhere from 15 minutes to an hour to copy out a single repetition of the tune, and it takes roughly the same amount of time to listen through three or four additional repetitions and note any differences.

Preparing my Transcription

To share what I've discovered, I’ll usually prepare a few figures using MuseScore. I’ll export the figures as .png files, which come with transparent backgrounds which look nice with my website’s current theme. The images tend to come out with wide margins, so I’ll crop them in Preview.

I prepare my audio examples using Audacity, which mostly consists of trimming the audio and adding fade-ins and fade-outs.

On a somewhat more technical note, I draft my blog posts in markdown, using Obsidian. It used to be a pain to embed all my audio and visual examples, so I wrote a small python script to insert them – I still need to manually add captions and alt text, but now at least all the file names are automatically sorted out. I used to use dillinger.io to convert from markdown to html, but I’ve recently switched to pandoc.

Final Thoughts

If you’re interested in transcribing music, a couple of tips: first, don’t feel that you need to use a bunch of fancy software to get started - if you can find the track on YouTube, you’ll be able to adjust the speed using the video controls. You can notate it on a piece of staff paper, or even just play it on your instrument without writing it down. And it’s fine to skip over or simplify a difficult passage – in my transcriptions, I’ve often chosen to not notate all the ornaments an instrument plays or to leave out an accompaniment line, just so that I don’t get bogged down in the details.

I always find transcription to be a rewarding experience, and I’d recommend it to just about anyone interested in learning about how music works. It feels a bit like watching a video in 240p and then switching to 1080p – after spending a couple hours listening closely to a track, I’m consistently amazed at how small nuances, a subtle variation played on the third repetition of the tune or the ornament added to one particular note, jump out at me where they used to go by unnoticed.

If you have any questions, comments or suggestions, don’t hesitate to get in touch! In particular, if you transcribe a piece or a part of one and find something cool, I’d be interested to hear what you've discovered!