Introduction

Last week I had the pleasure of making a few recordings with simple videography of the Embers choir – a Norwich-based community choir led by Deva O’Neill, who sing a wide range of European folk songs. I stumbled across a few phone videos of them on-line, and thought the choir sounded great, so got in touch and offered to help produce a few videos with – hopefully – better sound: the sort of thing they could use on social media, and, of course, that are always nice to have as a record for posterity. Fortunately they were game for it. Normally the choir practices in the Martineau Hall, but, for the purposes of recording, I suggested that they book the wonderful Octagon Chapel, which is next door in Colegate. It is a fine Nonconformist chapel built in 1754-6 by the Norwich-based architect and builder, Thomas Ivory, and, happily, has excellent acoustics. The recording was to be a simple affair: setting up at the beginning of their weekly practice, recording and filming three songs once they had warmed up and got into their stride, and when the gear was ready. Nothing elaborate, and not for a commercial album release etc. The choir were so welcoming and friendly, had thoughtfully turned up in muted colours (and not a stripe – the bane of videography – in sight: I hadn’t even remembered to request this!) yet conscientious and focused. It was a delightful evening.

The mic rigs

With minimal set up time, I wanted options from which I could select in post, so thought I would try pairs of omnis and cardioids, along with one of my twin-based DMS rigs. For consistency I decided to go with mics from one manufacturer, so, as still immersed in much testing of their mics, went for the Nevaton option. With a new addition to my various Manfrotto 154b stereo bars of a 1m-long carbon-fibre tube (which, as with the stock and modified aluminium ones, I filled with rockwool to remove any resonances), I went for omnis (MC59/O) spaced at 1m (slightly toed out), with a NOS pair of cardioids (MC59S/C) between (i.e. 300mm spacing and angled 90 degrees relative to each other), with a DMS pair of an MC59 Twin and MC59/8 in the centre. Essentially an OCCO set up with DMS thrown in. The mics were all mounted in Radius Windshield shockmounts – the omni and cardioid pairs on the little field-edition mounts with 55D-shore hoops (in the prototype evil red that I love!), and the DMS pair in stiffer black 72D-shore hoops. The bar was mounted on a Manfrotto 1004BAC stand, suitably sandbagged, and the seven cables ran to a Sound Devices 788T (I had a passive splitter too, and a second recorder for redundancy). Well, that’s about it: all very simple, and when the choir kicked off they did one take for each of the three pieces they wanted to record. They were content with each of the first takes, though for safety I should perhaps have suggested a second take of each, but it is hard when people are happy, they are not being paid, you are a new face, and the stakes are modest! Anyway, we got away with it (mostly: the sharp-eared will notice a few things of course!).

The recordings

Having a few options means that – aside from covering myself when recording – this is a chance to make a few comparisons, not just with the mic pairs individually, but also in combination (OCCO pairs are often used with one pair dominant and the other pair added at a lower level). Moreover, it gives a chance to compare NOS and omni pairs to DMS decoded to stereo in different ways including those surprising options in Harpex-X for near-coincident and spaced pairs (yes, the mind boggles how this can work!): just how different does DMS decoded to NOS or AB 100cm sound to the real things? So there is much to chew over after the primary work is done, with the three short videos and recordings sent off to the choir.

In fact there are so many permutations, I’m not sure where to start. First off, perhaps best to give the NOS and omni AB 100cm pair:

Then mixing these together, as if often the case with OCCO rigs, here is the NOS + omni AB 100cm with the latter down 9dB, and the reverse:

For the DMS here is a decode to a Blumlein pair, which works well (and, of course, is a logical decoding in Harpex-X):

And, finally and less convincingly (given the difficulty of creating a virtual near-coincident or spaced pair from a coincident array), here are the DMS pair decoded to both NOS and omni AB 100cm in Harpex-X:

These files can be downloaded, mixed, compared (with levels changed as needed) etc., but, for a comparison across a song here is a video that sees the mic pairs and combinations swap on the fly:

Conclusions

There are some obvious ones: DMS decoding in Harpex-X to virtual near-coincident and spaced pairs doesn’t sound much like the real NOS and omni AB 100cm pairs and is decidedly worse in both cases. Perhaps others have found success with it, but I’ve been a bit sceptical of such options in Harpex-X and, it seems, with good reason: it is not something I have used before and I won’t be doing so again! The DMS decode to a Blumlein pair (i.e. crossed fig 8s angled at 45 degrees to the front), however, is much better, which is not surprising given that Blumlein and DMS are coincident and mathematically equivalent set ups (although, of course, DMS allows much more flexibility in post): and, of course, the use here of a twin mic for the forward and rear-facing mid mics makes for greater coincidentality than with two separate cardioids. I can’t say that I am very keen on the omni AB 100cm pair, which sounds rather swampy, but that isn’t surprising: if using an omni AB pair on its own I would not go as wide as this. The NOS and omni AB pair combinations (i.e. with one pair down by 9dB and vice versa) sound rather better and, also, the NOS pair alone sounds fine. For the finished videos, however, I went for the DMS decode to Blumlein: as readers of this blog will have realized by now, I do like MS (and related techniques such as DMS) and Blumlein, but, quite reasonably, you may prefer a different option! Anyway, I hope one or two find this useful food for thought and an encouragement to undertake your own tests.

Oh, and for completeness, here are the three videos as finished: