Studio SOS: Winchester Talking Newspaper

John Richards (left) , Paul White (centre) and Simon Applebaum (right) trying out some new mic positions at the Winchester Talking News headquarters.

The SOS team are tasked with improving speech recordings for an audio newspaper aimed at the partially sighted.

John Richards is part of a team of volunteers who run and produce the Winchester Talking Newspaper (www.winchestertalkingnews.com), an audio news service for blind listeners in the area. The organisation is supported almost entirely by charitable donations, so they have to work to a tight budget, but they manage to produce a weekly audio newspaper and a monthly audio magazine. The idea is that teams of volunteers gather in a makeshift 'studio' each week to read around 50 selected items from the local newspaper. The recordings are then copied onto USB memory sticks, which are distributed, by Royal Mail, to about 100 blind and partially sighted people in Winchester.

The timescales are pretty tight! The Hampshire Chronicle is published on Thursday morning and, since the whole point of news is that it is current, the Winchester Talking Newspaper volunteers record their readings on Friday morning. The duplicated USB drives are then collected by the postman in the afternoon, and delivered to the listeners on Saturday morning.

Speed Reading

The necessary speed of the operation means that each item has to be recorded in one pass — there just isn't time (or enough skilled technicians and technology) to edit out mistakes. If a reading goes horribly wrong then a complete retake is made, but the listeners are happy to accept occasional stumbles as a normal part of the voluntary and amateur aspect of the enterprise — it's all very human and friendly.

Each news item is recorded as a separate file, using specialist software running on a Windows PC, and the audio is eventually transferred to the USB sticks by a dedicated duplicating machine. Recently, the team have started introducing 'roving reporter' inserts, recorded using a Zoom H2 handheld Flash recorder.

With the mics set up as they originally were, the voices that they were intended to pick up were arriving off axis, while the most sensitive parts of the microphones were pointing towards the reflective walls. This was contributing to an excessively 'roomy' sound.

As well as helping to organise the other volunteers, John is also one of the team's regular news readers, and his experience in teaching and drama has certainly paid off in terms of his professional delivery. However, John was not happy with the quality of the recorded sound, which was extremely 'roomy', and the organisation's IT expert, Simon Applebaum (a reader of Sound On Sound), suggested approaching us for some Studio SOS help. And so it was that we set out across the country, on a freezing cold February morning, heading for Winchester...

Room For Improvement

For their studio, the WTN team use a modest boardroom in an old listed building near the town centre. When we arrived, the room had bare walls, a large mirror above the fireplace, and a road-facing sash window backed by secondary glazing and a reasonably heavy curtain.

When recording the newspaper items, the typical setup had been to seat four people around a large baize-covered table, one per side, with the main presenter also operating the recording system. The recording software is a very straightforward Windows program called Newsbridge (www.bridgedigital.net/newsbridge_software.html), and the audio is fed into the computer via an automatic level-control device made by Audessence, called the Podblaster (www.audessence.com/PodBlaster.html). At the time of our visit, the four studio readers were being captured with just two Behringer B1 cardioid-pattern capacitor microphones, which were amplified and mixed using a small Behringer Xenyx 1202 mixer. The mics, which had their low-cut switches turned on, were set up back-to-back on a stereo bar supported from a low table stand. This was bolted to a large metal plate and positioned close to the centre of the table, with each mic aimed towards a table corner. A thin piece of folded polystyrene wrapping material had been placed under the plate, in an attempt to provide some shock isolation from the solid table.

The microphone levels are set using the mixer's preamp controls (which are never touched), with the presenter monitoring through cans connected to the mixer's headphone output. However, a stereo line output was also being fed into the mixer for listening to the recordings, and we found it was necessary to turn down the mics or kill the phantom power during playback, as the Newsbridge software leaves the mics live, even when not recording.

The Winchester Talking Newspaper use this Newsbridge software to record all of their spoken news stories.

Miking Up

The physical arrangement, with the two mics pointing towards opposite table corners, was intended to capture two readers on each mic — but meant that both readers were well off-axis, and the most sensitive parts of the microphones' pick-up patterns were aimed directly between the readers towards the reflective walls. Also, because the two mics shared a mounting bar, they were further away from the individual readers than was ideal. The arrangement created two main problems: the voices were tonally a little dull due to the off-axis response of the microphones, and the inevitable reflections within the untreated room were being picked up very strongly in relation to the wanted sound. These issues were the cause of the excessively ambient sound that John had described, and which the sample recordings he'd sent to us had confirmed. Unsurprisingly, there was also some evidence of traffic noise getting onto the recordings.

The obvious solution was a two-fold one: first, treat the room acoustics to reduce the natural reverberation as far as we could within the constraints of a temporary and non-destructive installation, and secondly to install two additional microphones so that each reader has a dedicated mic that can be used on-axis, and at the BBC-approved 'two hand spans' distance. John said they had a sufficient budget for two more B1 mics and some table stands, and we suggested also adding an order for some mesh pop shields, as the foam socks originally supplied with the mics tend to dull the sound a little and aren't as effective against popping as external screens. We had one SE Electronics pop shield with us courtesy of Sonic Distribution, and demonstrated how it should be set up, so that the three new ones could be deployed in the same way. Thankfully, the Xenyx mixer had four microphone inputs, so other than a couple of new microphone cables no further expenditure would be required.

Some Universal Acoustics foam hung on the walls and against the mirror, along with a duvet on the opposite wall, yielded a significant reduction in room coloration on subsequent test recordings.

Foam Was Built In A Day

The room was equipped with an old-style picture rail running around the top of the walls, so we quickly hit on the idea of using this to suspend our acoustic treatment using brass picture hooks bought from the local hardware store. Over the mirror, above the fireplace, we simply propped up a couple of Universal Acoustics foam tiles, but for the large wall opposite we fell back on our oft-used duvet solution, fixing a double-size polyester duvet to a series of picture hooks using cable ties poked through holes punched in the edge of the duvet. A second single-size duvet was procured and suspended in the same way in 'landscape' format along the wall opposite the window, above a pair of drawer cabinets. Two further Universal Acoustics tiles were fitted with picture hooks, and hung in the alcoves on either side of the fireplace.

The reduction in liveness was immediately very noticeable when speaking in the room, even though the ceiling was still untreated. Should it become necessary, nylon cord could be looped between pins fixed into the picture rail above the fireplace and into the hefty ceiling support beam to hold up a 'cloud' of foam panels, or even a light duvet — but we felt that once the new mics arrived the room treatment we had installed would be adequate. John and Simon could always add the ceiling cloud later if they felt it would be worthwhile. For the window end of the room, we simply drew the heavy curtains, which provided a reasonable amount of absorption.

Dry Run

The preamp gain controls on the Behringer mixer were initially set too low, so that the Audessence Podblaster auto-leveller (which you can see underneath the mixer) was doing little other than raising the level of undesired room sound.

A test recording with two of us speaking into separate mics — this time set up at an appropriate distance and directly on-axis — revealed a much drier sound, but even so there was still more room tone than we'd anticipated. The obvious culprit turned out to be the Audessence Podblaster, which is a 'set and forget' digital programme-levelling device intended for podcasting and streaming. This stereo unit incorporates an AGC (Automatic Gain Control) with a simple compressor, a separate limiter, a low-cut filter and soft gate. It is supplied with a range of factory presets intended for different applications, but can be tweaked manually if required. It seemed to us that the inbuilt compressor was accentuating the room tone to an unnecessary degree, so Hugh set about examining its settings, which are accessed via software running on a PC.

It turned out that the signals from the Xenyx mixer were rather lower than they should have been, and that the Podblaster was set up to introduce a considerable amount of gain to bring the output up to the device's nominal operating level. The limiter was barely doing anything at all, and so instead of the unit working to catch and pull down occasional loud peaks, its dynamic control was actually spending most of its time pushing the room tone up! After Hugh's adjustments to the software settings and to the mixer's mic-input gains, we repeated the reading test and the results were noticeably better.

A free pillow came with one of the duvets, so we placed that between the mics to give a little more absorption behind them. Thinking about the two extra mics that were planned, we then formulated another budget idea. In order to improve the separation between the four mics, and to help further reduce the amount of room tone contributed by the unused mics when only one person was speaking, we suggested improvising a barrier using squares of open-cell furniture foam. It could be made in the form of an 'X' shape standing 300 to 400mm high, with the arms of the 'X' pointing towards the table corners. This way the rear of each mic would point into a foam 'V' shape, further helping reduce the level of off-axis sound reaching the rear and sides of each mic. John and Simon thought this was a good idea and resolved to put it into practice. We also suggested placing some half-inch foam blocks under the new microphone table stands to provide more effective shock isolation, in case the readers accidentally banged or kicked the table.

John Richards (left), Simon Applebaum (centre) and Hugh Robjohns (right) trying out the new recording setup. The microphone distance had been significantly reduced, and a pillow placed between the mics to further minimise room coloration.

The Day After...

So, having treated the room to a reasonable degree, reconfigured the automatic level-control device and identified a much more appropriate microphone technique, our final task was to set up the two existing mics as best we could for the next day's recording session, as the new mics would not have arrived by then. Examining all the possibilities, we decided the best option would be to sit two readers, side by side, on each of the two long sides of the table, with each pair sharing a single microphone placed between them. To that end we moved the two mics to the end fixing holes of the stereo bar, back-to-back, and placed the bar at the centre of the table facing across the short axis. The spare pillow was again wedged between the mics to absorb some of the off-axis sound. We made some test recordings using this setup and, as expected, the results weren't as good as using one mic per reader, as the two readers sharing each mic were necessarily both off-axis. However, because of the damping materials on the walls and because of Hugh's tweaks to reduce the amount of dynamic processing, the outcome was still significantly cleaner than it had been before we made the changes.

John and Simon definitely 'got' our approach, and so we were confident that once the new mics arrived they'd have no trouble optimising their setup. Having done what we could, we all adjourned to the coffee house next door for a spot of lunch, before tidying up our tools and heading home.