You are here

Session Notes: Blue Rose Code

Recording A Live Show By Sam Inglis
Published October 2014

Blue Rose Code on stage. From left: Jessica Moncrieff (violin and backing vocals), MG Boulter (pedal steel and slide guitar, barely visible!), Samantha Whates (backing vocals), Ross Wilson (guitar and vocals), Daniel Paton (drums) and John Parker (double bass)Blue Rose Code on stage. From left: Jessica Moncrieff (violin and backing vocals), MG Boulter (pedal steel and slide guitar, barely visible!), Samantha Whates (backing vocals), Ross Wilson (guitar and vocals), Daniel Paton (drums) and John Parker (double bass)

We overcome unfriendly acoustics and troublesome spill to capture a memorable gig.

To weary passers–by and workers in London’s busy King’s Cross area, St Pancras Old Church and its grounds provide welcome relief from the pressures of the world. And to musicians seeking an unusual location for a concert, it makes a dramatic alternative to conventional music venues. Among the artists who have taken advantage of this striking space, hidden away in a quiet churchyard behind St Pancras station, is singer–songwriter Ross Wilson, who records and performs under the name Blue Rose Code. Ross hit on the Old Church as the perfect venue in which to launch the second Blue Rose Code album, The Ballads Of Peckham Rye.

For his tour in support of the Arts Council–funded long player, Ross had assembled an excellent five–piece backing band, and was keen to have the album launch recorded for posterity. Having recorded a couple of Blue Rose Code shows before, I was happy to oblige — but when I started to think about the practicalities, it soon became clear that this was going to be quite a challenge.

With no permanently installed PA, live sound in the church is handled by Joel Cormack using a pretty basic setup centred around a small Behringer Eurodesk mixer and a pair of Mackie powered speakers. The interior of the building is almost entirely faced with stone, creating a pretty hostile acoustic environment for anything remotely loud or uptempo. And the band’s line–up would involve no fewer than six vocal mics, not to mention a full drum kit and potentially awkward instruments such as violin and double bass.

Scratch & Sniff

Happily, I was able to contact Joel well before the event, and he proved friendly, helpful and calm. As I don’t own a microphone splitter, and there was no budget to hire one, my plan was to ‘sniff’ signals from the insert points on the Behringer mixer. This requires special balanced cables with tip and ring wired together at one end, which I already had, and of course is only an option if the front–of–house engineer isn’t using the inserts, which Joel wasn’t. This setup, on paper, would give me line–level feeds of all the channels that were going through the PA, and by pairing a Creamware A16 Ultra A–D converter with an RME Fireface 802 interface, I cobbled together the necessary number of line inputs.

However, Joel explained that he doesn’t usually mic up drum kits or other loud instruments, so I knew I would need mics and stands of my own. He also pointed out that his mix position is near the back of the church, so a reasonably lengthy snake would be needed to connect these to my recording rig. It was about this point I abandoned the idea of getting to the gig on the train...

Session NotesBy the time I’d made it through the London traffic, Joel’s setup was almost complete, and as it turned out, he was using one drum mic: a Shure SM57 in front of the kick drum. The drum in question was a small ‘cocktail’ specimen with a full resonant head, so this sounded rather soft and lacking in impact; I thought about adding a second mic on the beater side, but it proved physically impossible to position one. I miked the snare with a Beyer M201, and used a not–very–widely spaced pair of Neumann KM84s as overheads. I chose these because I knew that I would be fighting significant bleed: KM84s have excellent off–axis response and their small size makes them easy to position, so I could get them fairly low over the kit.

The other vital source that I couldn’t rely on capturing via Joel’s FOH setup was the audience. Though I realised I was already likely to be fighting an overload of room ambience at the mix, I also wanted a true record of how the band sounded in the hall, so I set up a Pearl ST8 variable–pattern stereo mic near the mix position to capture both. As some audience members were sitting around and behind this mic, theory suggested the use of a stereo configuration with a 360–degree acceptance angle, such as M/S with an omni Mid mic. However, most of the action was coming from the front, and as applause isn’t something that demands to be reproduced with pinpoint stereo imaging, I compromised by using a subcardioid Mid and figure-of–eight Sides setting. One of the great advantages of M/S is that as long as the Mid mic is positioned appropriately, you’ll always have a decent mono signal to fall back on.

Showtime

All of my mics went through my Focusrite ISA828 preamp and into line inputs on the Fireface. This, in turn, was hooked up to an Apple MacBook Air running Pro Tools 11. To keep things simple and avoid the need for USB hubs, I recorded directly to the system drive. As well as the five mic signals of my own, I ended up with 14 active signals from Joel’s desk. Jessica Moncrieff’s violin, John Parker’s double bass, and Ross’s acoustic guitar were all coming into the desk through DI boxes, along with Cyrus Shahrad’s stage piano (in stereo) and a second acoustic guitar that featured on a few songs. MG Boulter also played pedal steel through a Fender Deluxe amp; this was miked up, as were the six vocals and the aforementioned kick drum. Finally, it turned out that Ross had a surprise up his sleeve in the shape of legendary double bass player Danny Thompson, who would be taking over Parker’s instrument for part of the encore.

Singing drummers are always fun to mix. Here you can see the two Neumann KM84s I  used as overheads, low over the kit, and Beyer M201 on snare drum.Singing drummers are always fun to mix. Here you can see the two Neumann KM84s I used as overheads, low over the kit, and Beyer M201 on snare drum.Thankfully, both support acts were solo artists, so most desk channels could be left unmolested between soundcheck and showtime. Recording the soundcheck allowed me to confirm that all the signals were reaching my hard drive, but also threw up a few red flags. Several of the sources I was sniffing from Joel’s desk were clipping, so I had to ask him nicely to turn the gain down at the preamps. More of a problem were the weedy kick-drum sound and the Fender amp, which was surprisingly noisy. My biggest concern, though, was the double bass. This was amplified using a miniature microphone, presumably an omni, mounted to the bridge — and if I hadn’t known better, I’d have assumed that it was actually miking the drums! We were able to improve the pedal-steel situation slightly by repositioning the mic, but there was little I could do about the other problems except cross my fingers.

Having recorded Ross before, I was also forewarned of a few other likely issues. In the heat of performance, both he and his female backing singers Samantha Whates and Jessica Moncrieff have a tendency to wander away from their microphones; had I had the budget and time, I’d have been tempted to attach a lavalier radio mic to him. Once he gets going, his percussive guitar playing also puts out huge low–frequency thumps, and I warned Joel to leave plenty of headroom.

Front–of–house engineer Joel Cormack (rear) and I  set up our equipment on a  table at the back of the church.Front–of–house engineer Joel Cormack (rear) and I set up our equipment on a table at the back of the church.Recording unrepeatable events to a laptop with no real–time backup isn’t the most comfortable thing in the world. From software crashes to cables becoming dislodged, there’s endless potential for things to go horribly wrong. My nerves weren’t eased when Pro Tools decided to stop recording and report a non–existent CPU overload during one of the support slots, but when it mattered, it managed to get through Blue Rose Code’s ecstatically received 100–minute headline set without falling over. The house lights were barely up before I was copying 20GB of audio files to a portable hard drive, and as ever, breaking everything down was a hundred times easier than setting it up.

Record In Haste, Mix At Leisure

The next day, copying the files onto the Windows machine I use for mixing confirmed that every source had been captured all the way through, with no dropouts. However, it was far from being a pristine studio multitrack. I usually find that levels are all over the place with signals that are ‘sniffed’ from the FOH desk, and so it proved here: some were healthy to the point of occasionally clipping, while others were barely present at all. Low levels are not, in themselves, a problem in a 24–bit recording, but for some reason the stage piano was both vanishingly quiet and audibly noisy. The vocal tone constantly changed as Ross and his backing singers moved around, and, as I’d feared, my crisp drum sound very quickly went to the dogs as soon as I raised the fader on the double bass.

Placing a room mic in the middle of the audience can be disastrous if people sitting nearby want to hold a conversation or eat crisps, but the punters had been very well behaved, and the applause sounded good. Its audience perspective on the sound of the band, however, was enlightening. Despite the presence of 200 human bodies, the ambience in the stone church still sounded hard and mid–rangey.

Mixing a live recording often involves quite a bit of extra effort compared with working on well–recorded studio tracks, but this labour is usually front–loaded: while many of the sources will require a lot of corrective EQ and other processing, you can normally rely on being able to use similar settings across the entire set. In other words, once you’ve successfully mixed one song, the rest will follow fairly easily. In this case, however, things didn’t really work out that way, because no two songs used the same line–up or arrangements. The drum kit featured only on about half the songs, and even then, the drum sound varied as drummer Daniel Paton switched between sticks, brushes and beaters. Thankfully, the buzzing piano was used only on a couple of tracks, and then in a background role.

‘Sniffing’ the insert points is a  cheap but rather inflexible way of recording the signals going through the FOH desk. It also has the potential to generate some serious cable spaghetti!‘Sniffing’ the insert points is a cheap but rather inflexible way of recording the signals going through the FOH desk. It also has the potential to generate some serious cable spaghetti!Anyone who’s mixed a live recording will be familiar with the problem posed by guitars and other instruments DI’d via a piezo pickup. No matter how much mid–range you cut, there always seems to be too much; and no matter where in the mid–range you cut it, some other part of the mid–range always sticks out. Thankfully, the violin had been captured using a clip–on mic, while MG Boulter’s slide guitar actually sounded pretty good; but Ross’s guitars were typically spiky, requiring multi–band compression as well as EQ to bring them under control. Ross, as predicted, did his share of guitar–whacking, but played other tunes fingerstyle, and also swapped between two guitars which sounded somewhat different from each other. To add to the fun, one of these had been dropped on a stone floor before the set, and emergency repairs to a machine head hadn’t prevented it drifting out of tune.

Meanwhile, reining in the variable vocal level and tone was not easy, and my eventual solution involved an equaliser to remove unnecessary mud, followed by a de–esser, and a multi–band compressor to keep the mids and low–mids in line. This then fed a conventional compressor, with plenty of level automation also required. On top of all that, there were quite a few places where I needed to apply a high–pass filter momentarily to address plosive pops. Similar treatment was required for several of the backing vocals, too.

Session NotesThe question of vocal reverb posed a dilemma. The room mic and spill from the vocal monitors onto other sources had ensured that plenty of ‘reverb’ was captured, but unsurprisingly, it didn’t sound very good. On the other hand, if I added enough artificial reverb to overshadow this spill, the vocal ended up sounding too wet, and the distinctiveness of the venue was eroded. In the end, I decided that the best reverb plug–in in the world was never going to make this sound like it had been tracked in Abbey Road, so I erred on the side of using the room mic, albeit heavily equalised. This was augmented with a short ambience and a larger chamber from Acon Digital’s Verberate and Exponential Audio’s Phoenixverb respectively.

Spills, Thrills & Bellyaches

Although all of these issues took a fair amount of trial and error to address, and some required solutions which compromised the overall quality, none of them threatened to make the tracks unmixable. The biggest problem in this respect was the double bass (see box). This can be a difficult instrument to record and mix at the best of times, and in this case, the level of spill from other sources caused real problems. It’s heartbreaking when an otherwise decent drum or vocal sound suddenly turns to mush as soon as you unmute the bass channel!

Ultimately, though, when you’re focused purely on technical considerations, it’s easy to worry too much about them. A live recording is a ‘warts and all’ record of a moment in time, and there’s more to be gained by accentuating the ‘realness’ of the performance than by tidying everything up to the point where it becomes sterile. When I began sending Ross rough mixes, it was clear that he was happy with the sound quality. More than that, he was overjoyed to have a document of a one–off event that meant a great deal to him, and is hoping to put the recordings out as a live album or EP for fans. The sound of a large and appreciative audience really enjoying themselves does a lot to compensate for imperfections!

Double Trouble

By far the biggest headache I encountered in mixing the Blue Rose Code show was John Parker’s double bass. The problem had nothing to do with his playing, which was excellent, but with the sound that was coming from the bridge–mounted miniature microphone. In its raw state, the signal sounded like a fairly trashy room mic, with as much drum and vocal spill as bass; and what it had captured of the double bass did not sound very natural. The low frequencies and the percussive ‘snap’ of the strings on the fingerboard were emphasised, while the mid–range ‘growl’ that is so characteristic of the instrument (and so essential to making it cut through on small speakers) was largely missing. Any attempt to boost the mid–range using EQ simply delivered more and nastier–sounding drum spill!

MG Boulter’s pedal steel went into his Fender Deluxe amp. Perhaps because he had the amp turned up loud in order to control his level with a  volume pedal, it was surprisingly noisy.MG Boulter’s pedal steel went into his Fender Deluxe amp. Perhaps because he had the amp turned up loud in order to control his level with a volume pedal, it was surprisingly noisy.Folk–rock royalty in the house: Danny Thompson guests on ‘This Is Not A  Folk Song’.Folk–rock royalty in the house: Danny Thompson guests on ‘This Is Not A Folk Song’.The presence of a  capacity audience did something to tame the reverberation within the church, but didn’t make an enormous difference.The presence of a capacity audience did something to tame the reverberation within the church, but didn’t make an enormous difference.I tried all sorts of fancy solutions to the problem, including an attempt to restore the missing ‘growl’ by sending a heavily filtered version of the source signal to a distortion effect, and using Melodyne to generate a MIDI bass part from the audio recording (which failed because of the amount of spill). In the event, it proved impossible to make radical changes to the bass sound without colouring the spill to a point where it ruined the entire mix, so I settled on a chain of plug–ins configured to reduce spill as much as possible while controlling the low end. Again, the heavy lifting here was done by FabFilter’s excellent Pro–MB multi–band dynamics plug–in. One of the great features of this plug–in is that you can set each frequency band’s key filter range independently from the range the band covers. Using this trick I was able to control drum spill to a certain extent by configuring a band’s key signal to focus on the high frequencies, so it would reliably trigger only on spill, yet having the band itself duck a much wider frequency range whenever the snare or hi–hat hit. At the other end of the spectrum, Pro–MB also gave me control over the rampant sub frequencies.

No matter how many times I returned to the bass with fresh ideas, though, it remained the limiting factor as to how good the overall mixes could sound. Without the mid–range ‘growl’ to help it cut through, the bass sound itself always felt indistinct and lacking definition; and it was impossible to make the bass audible in the mix without spill affecting both the vocal and drum sounds. There were, at least, a couple of songs which featured double bass but not drums: here the bass was more exposed and spill less of a problem, so I could push the mid–range a lot harder. Elsewhere, the result was inevitably a compromise.

Blue Rose Code

Edinburgh–born singer–songwriter Ross Wilson, now based in London, has built up a big following for his music the old–fashioned way, through touring and word of mouth. Despite much interest from record labels, he chose to set up his own label to retain complete creative control, and 2013 debut North 10 brought him to wider attention, both in the UK and further afield. New album The Ballads Of Peckham Rye features both the cream of Scotland’s jazz scene and double bassist Danny Thompson, known to millions for his work with Pentangle and John Martyn.

Audio Examples

To hear songs from the show, and snippets that illustrate some of the problems encountered at the mixing stage, point your browser at www.sosm.ag/bluerose.

Audio Examples

Recording and mixing a live show is always a challenge, and these audio files from Blue Rose Code’s set at St Pancras Old Church illustrate a few of the issues that can arise!

First of all, here are two complete, mixed songs from the show. I’ve chosen ‘Edina’ and set closer ‘Julie’ because I think they provide good examples of the way in which the energy of live performance and the presence of an enthusiastic audience help to overshadow technical issues with the recording. Because there are no drums in ‘Edina’, the spill onto the double-bass mic was manageable, but both the pedal steel and the piano were very noisy. To bring across the audience response in ‘Julie’ I needed to keep the room mic fairly high in level, despite the dodgy acoustics; the bass sound also suffers because of spill issues.

One of the problematic elements of this particular live recording was an inconsistent vocal sound. In the first of these two files, you can clearly hear how Ross Wilson’s vocal changes radically in level and timbre as he moves around. The second file is the same track as used in the mix, with multi-band compression to even out the timbre and extensive level automation.

The most troublesome source in the mix was the double bass mic, which captured almost everything but double bass! In the first of these examples, you can hear how the drum sound goes to the dogs when the bass track is unmuted halfway through. The second and third represent respectively the raw bass recording, and my attempt to turn it into a usable double-bass track by reducing spill.

This is what it actually sounded like in the room!

If you mix a lot of live recordings, you’ll have to find ways of making acoustic guitars recorded with piezo under-saddle pickups fit into the track. As you can hear, the raw sound in this case is very aggressive, and I’ve tried to soften it using multi-band compression and a short ambience reverb.