Beeston’s Second Time Around Folk Club kindly invited Carolan along to a club evening to talk about its recent stint as ‘guitar in residence’ (Posts 81, 82, 87, 88, 90, 95, 101). I grasped this as an opportunity to put together a Carolan show that combined performing tunes and songs with explaining the wider concept and history of the project. This also provided an opportunity to tackle a challenge that had been niggling at me for several years – how to gracefully embed the scanning of Carolan’s interactive artcodes into a live performance. I’d briefly flirted with this idea when recording a set for the Nottstopping online festival, when I’d included a segment that demonstrated scanning Carolan’s artcodes to conjure up a backing track that I then played along to (Post 78). However, that had been pre-recroded; attempting such a thing live on stage felt like a quite different proposition altogether.

After some experimentation, I devised a 35 minute show, split into two sets at the club’s request, that progressed from initially playing the guitar in a conventional way, to revealing different aspects of its history, to finally scanning its artcodes to draw in material from its archive. Here’s a quick highlights video.

On balance, I feel that the show went pretty well for a first outing of an experimental new format. The technology was sufficiently reliable to deliver a coherent show without seriously awkward moments. The balance of content, between tunes and songs and discussing research, felt about right for this audience. It was definitely fun to perform and I think that the audience appreciated it too (though no doubt they would have been polite even if not – but you can still tell). They certainly sang along, laughed, and commented at the right places, and there were plenty of questions and comments at both the interval and after the show.

The remainder of this post documents in (way too much) detail how I designed the performance …

The format of the show

The first half of the show unfolded like this:

• Introduce the inspiration for the project through a personal story about my daughter inheriting an old violin many years back.
• Play a two tunes on Carolan – The Wild Rose of the Mountain and Planxty Fanny Power, the latter being one of Turlough O’Carolan’s compositions. I was fortunate to be joined by Penny on her flute for these.
• Talk a little about Carolan as a guitar, mentioning Turlough O’Carolan as its muse and inspiration for its inlaid Celtic knot work decoration.
• Scan Carolan to trigger the playing of a short ‘making of’ video edited specially for the show, featuring a sound track performed by Steve Hicks (a favourite guest of the Club) on Carolan (see Post 48). However, I gave no explanation of artcodes at this stage, saving it for the second set.
• Close the first set with chorus song that everyone could join in on. Somewhat unusually, Penny and I choose a pop song called W.E.E.K.E.N.D (a family favourite from Arling and Cameron’s album Music for Imaginary Films) that we think works surprising well as a chorus song, especially on a Friday night.

The second half went like this:

• I began by talking about how Carolan had encountered many other guitarists, including well-known professionals that the audience would likely know, but also club members through the recent residency
• I then scanned Carolan to conjure up specially edited ‘showreel’ video summarising the residency.
• I now explained the artcodes scanning, explaining the concept and showing how different surfaces of Carolan can be scanned, ending up by triggering a ‘How to Draw Artcodes’ video.
• The final section of the show turned to the question of ‘why do this?’, explaining the kinds of stories that guitars might tell if they could talk to you.
• The final example was of the guitar being able to summon forth videos of absent fiends and bandmates to play along to. I triggered a final video, of the Beeston Ceilidh Collective, who are regular guests at the club (and indeed who were the final guests before the global pandemic forced closure of the club for many months). For this I reedited a band video that we’d produced during lockdown, removing my guitar part and Penny’s flute part so we could play along live (though I left in our banjo and whistle parts) .

You’ll find links to the various pre-prepared videos at the end of this post.

Hardware set-up

The most significant hardware choice was to use a mobile phone for the artcodes scanning and playback of digital media. I had also tested with a laptop with an external webcam, however, my iPhone’s camera appeared to cope better with low lighting conditions (which I anticipate might be the case in performance settings) while its small size makes it less obtrusive on stage.

Early on, I realised that I wanted to work with a mode of interaction that involved holding the guitar and pointing it at a fixed camera, rather than trying to hold the phone and point it at the guitar as has usually been the case when previously demonstrating artcodes. This felt like a far more naturally performative style of interaction on stage that didn’t involve having to swap between holding guitar and phone or even try and hold both simultaneously. It felt less awkward to do and less intrusive on the structure of the performance. More generally, it seemed to offer a potentially graceful solution to the problem of ‘encumbered interaction’ in which guitarists try to control digital media and devices while simultaneously playing (see Post 65). After some experimentation, I chose to mount the phone on a microphone stand positioned quite low down and to one side of me that sat in a sweet spot where scanning seemed to work reliability without the camera sitting directly between me an the audience.

Another key element of the tech set up was the use of a short-throw projector that could be readily deployed, alongside a portable screen and PA system, in the relatively small and crowded club room where the performance took place. The resulting technical set-up was light weight, had a small footprint, and could be set up in under an hour.

Mapping artcodes to videos

Further testing explored the scanning of Carolan’s various artcodes using this fixed-camera set up. Scanning the larger artcodes on the front, back and top sound hole of Carolan worked reliably, even at some distance, whereas the smaller codes on the headstock and in the cutaway did not, requiring careful positioning of the guitar and holding it steadily in place which might be difficult to achieve given the nervous tension likely to be be felt on stage. Moreover, repositioning the guitar to access the codes on the front top and back by flipping it over horizontally while sitting and holding it in two hands felt quite easy to achieve and even relatively natural compared to the lateral side-to-side movements required to bringing the headstock into view of the camera which felt more awkward. The guitar could be quickly spun around to reveal front, back or top, perhaps with a suitable air of relaxed casualness or even something of a flourish. At least this was the case when seated (how I usually perform); interacting while standing and using a guitar strap would need to be quite different.

The net result was that I built the interactive elements of the performance around scanning just three codes: the front, top and back. Testing also revealed that it was good to manually press one ‘play’ button to trigger videos even if this mean reaching down to the phone screen as this avoided false positives that could happen whenever codes came into view of the camera (especially the one on the front).

The videos were played via Youtube, and upgrading my account to Premium removed all adverts and allowed me to download local copies that I could still work with in case the networking in the venue was inadequate (though in end the pub’s WiFi was fine on the night). I created a ‘splash screen’ video that showed a default background image on the projector and also included this at the end of each video so that this would be left on the screen whenever a video finished playing.

My performance plan required me to play four videos, but I had chosen to only work with three Artcodes. This required me to create two different Artcode mappings, one for the first half and one for the second.

Artcode location	First Half Mapping	Second Half Mapping
Front	Splash screen video	Residency showreel video
Top (sound hole)	Making of video	How to Draw Artcodes video
Back		Ceilidh set to play along to

Structuring the show as two halves helped me here as I could manually swap mappings in the Artcodes app between the two halves. Had it been one continuous show then I would have needed something more graceful that allowed me to move between mappings during the show itself. One option would be to extend Artcodes so that scanning a code can load a new mapping so that a performance could be set up as a sequence of mappings. One of the three codes would then move you along the sequence, leaving two free at each stage to be mapped to videos.

Final reflections

So on reflection, this first outing of a Carolan interactive performance felt quite promising and I learned a great deal about how to set up the triggering of artcodes on stage. I wonder if the show might be extended and even be tailored to different audiences, for example folk audiences who are primarily interested in the music but would enjoy some of the backstory, versus ‘science’ audiences who might be mostly interested it the technology and research but appreciate an element of musical performance.

Videos from the show

Here are the videos that were shown during the performance.

A set of jigs to play along to from the Beeston Ceilidh Collective