Leicester recording - session 6

During this session, I used one of the iMacs in MIDI lab (no other studios are available due to refurbishment).

Checking work from previous session

With pleasure I listened to the organ pieces that I mixed before - I love organ music! I found nothing to change in matter of sound. There was a couple of pauses in which noises from the audience went through. To eliminate them, I used Noise Gate. Threshold of -20 dB separated the noise from music. I set the attack to 5 ms so that the gate would open right at the beginning of the note. Hold was really long (630 ms) in order to omit short pauses. Release was long too (430 ms) to leave some time for reverb which might be below the threshold.


Noise Gate applied to mastered track

This seemed to work on both organ pieces. Automation was used to disable the gate for any other tracks.

Starting with other pieces - organ and piano

As no studio with good monitors was available, I decided to do some rough mixes for some new tracks. I started with piano and choir song 'Cantique de Jean Racine'. Using automation, I adjusted gain on reverb and piano so that they would jump up to 0 on the loudest part. As mentioned before, sound of the piano is quite bad. It has no attack and no bottom, and is out of tune a little. The only thing I could do with it was EQ - I used low cut below 94 Hz to cut out bass rumble, and a band filter at 510 Hz - down by 9 dB.


EQ on piano track

This apparently cut out most of the frequency content, so I had to raise the gain even more. At least it doesn't sound like on an old recording now that it has more bottom and top.

I managed to set the faders in a way that sounds OK on headphones, but this will have to be adjusted on proper monitors. That's it for this one.


Mixer settings for choir and piano track

Mixing Missa Brevis (choir, soloists, strings and organ)

This is the first track of this concert with soloists singing to their microphone, so I started with having a close listen to this channel. I was extremely happy with NTK performance - it got the tops crystal-clear, while maintaining balanced bass and middle bands.
As soloists never sing together with choir, I decided to cut the choir pick-up from the solo channel. This was quite a lot of work, because some parts of this piece are composed in a 'dialogue' between choir and soloists. However, if I wanted to keep the mix clear, I had to do it. To make the solo pieces appear without nasty digital 'click', I faded each region in and out very shortly.


Arrange window showing cut solo track

After cutting solo channel in pieces, I found the loudest part and adjusted gain plugin to raise the levels. I did the same thing with reverb and organ / strings channels - choir bus was apparently reaching 0 without any adjustment. Then, I tried to make a rough mix to get the proportions right. To make soloists stand out from the choir, I widened choir pan to +30 and -30.
I didn't apply any compression or EQ, but after setting the faders I realised that most of the choir channel is really quiet, with one or two loud peaks - I think that for this piece it would be OK to use compressor.


Mixer set-up for Missa Brevis

I finished with mixer looking like on the picture. Sound-wise, reverb is dominant (even though its fader is below choir's fader - for the reason described above) and instruments, choir and soloists 'float' in it.

Leicester recording: 5th Logic session

This session's mix was done in Studio B - Mastering Suite.

Previous mix

I had another listen to Wesley track and I liked it a lot. Valve track, as usual, was warmer and 'closer' than unprocessed mix. Solos sounded a bit 'spaced out', but not much can be done about it since the soloist was away from the microphones and a lot of reverb was picked up. In previous mix I managed to adjust the base of the organ so that it now sound deep. It is also quite wide, as reverb microphones picked up a lot of it, but that's very good because even though it's loud it still doesn't interfere with the choir, which sits on the middle of the mix.

Mixing organ solos

I started off with adjusting pan on reverb bus. As you remember, the organ was placed on the right side of the altar, so on reverb microphones it was significantly moved to the right. To compensate it, I changed the pan to 15 left.
Next, I searched for the loudest bit in both tracks and adjusted gain plugin so that both reverb and organ channels were jumping right up to 0. Then, I tried to search for the best proportions between organ microphone and reverb: I must say that the reverb channel sounded great just on its own. When I added the organ channel though, it filled up a kind of empty space on the middle, and added some detailed sounds of keys. I ended up with blindly done setting with both channels set to 0, which I sent to valves.


Mixer window for organ tracks


Valves

After setting up levels (I always find it very difficult to adjust the gain on both strips to that mono signal is received at the same level at left and right channels), I compressed the sound a little - as usual in this project, to act only on the loudest bits.


Valve compressor

EQ goes as follows:

• Low cut at around 40Hz
  
• Top cut at almost 25kHz
  
• Low shelf up significantly around 250Hz
  
• Low band pass down a lot around 400Hz (some very resonant sound there)
  
• High shelf down a bit from about 6kHz (to reduce hiss present in the recording and raised by compression, plus to smooth the mix a little).
  

After EQing, I did some final levels adjustments using EQ's master knobs (green ones on the right), and bounced the whole thing.


Low- and high-cuts and low shelf of the valve EQ


Middle band pass, high band pass, top shelf and master volume knob of valve EQ


Output and input levels during bounce


Spectrograms of output and input during bounce

During bounce I realised that the choir track was not switched off, but it was so quiet that when I listened to the difference it makes in the mix afterwards, I couldn't tell the difference.

DREADFUL PEAK

There we go - I was setting the levels up on what I thought to be the loudest part of both pieces. Well, apparently there was one a bit louder part in another track, which I discovered when it was already being bounced down. The audio interface clipped it in most nasty way possible, so I was really close to going through the whole work with setting levels up again, but at first I decided to try to remove it.


Nasty clip

I located 2 clipped peaks on Logic's sample editor and... used pencil to smooth it up! I had a listen and I could not believe my ears - the nasty clip was not only smooth now, but also it was completely inaudible.


Fixed clip

Having everything else recorded correctly and with good levels, I decided to keep it this way.

Leicester recording - 4th Logic session

This post describes the session in studio C with proper monitors and valve equipment.

Intro
As usual, I started with a quick listen to the organ / choir pieces mix I've done during last session. On speakers, it sounded a bit flat and without space, so I switched off all the automation from the Wesley track and increased reverb volume for this one up to -3.8. With that, solos gained some depth without really loosing any clarity.

'In the box' adjustments
Without any EQ applied, the recording is a bit flat on bottom side. When the choir goes fortissimo, it sounds metallic. I decided to correct that with valves. I also noticed some noises made by bored choir members (I'm obviously joking, there is no way they were bored!) flipping through pages and moving around, so I applied some more EQ on solo track - gentle (12 dB/oct) top cut at 16.8 kHz and -9 dB on top shelf at 10 kHz.


EQ on solo track


Mixer window for the Wesley track

After this little adjustment, all things that I wanted to have done 'in the box' were finished, so I started with valves.

Valve EQ and compressor
I sent both organ tracks to the same stereo valve configuration: compressor and EQ. I needed to re-patch cabling as by default the signal goes to EQ first and then to compressor, but remembering that EQ is a bit noisy, I wanted to feed it good signal, and then adjust the output levels with master gain on EQ.
Remembering that I was working on classical piece of music, I made the compressor to act only on very loud bits. Ratio was set to 1:5 with fast attack and slow release. I had to lift the gain make-up quite high to drive EQ well, and with that I achieved good signal to noise ratio in the EQ.


Valve compressor settings

To deal with the sound issues I mentioned above, I used following settings:
Bottom shelf up below 300Hz up - to bring the bassy organ sounds up and make the mix warmer.
Mids around 600 Hz down just a bit - to get rid of unwanted resonances.
3kHz down quite a lot and wide - this is where most of the metallic sounds was.
Top shelf above 8kHz up - to add some air.
Low and high cuts on - from 40 Hz to about 18kHz to get rid of unwanted noises.


Low- and high-passes and bottom shelf on valve EQ


Two middle bands, high shelf and master volume on valve EQ

Result: Clearer, warmer sound.
With settings as above, I recorded both tracks back to Logic.


Valve-processed sound comes back to Logic


Levels comparison between sound coming out (top one - it is above 0 because the meter is pre-fader) and coming back from valves (bottom)


Spectrum analysers for the sound coming out (right) and back in from valves (left).

Leicester recording - Logic sessions 2 & 3

Session 2 (in Mastering Suite)

Pre-works

In this session, I finished cutting the concert into regions. After that, I had a good listen to tenor solos, and decided to stick to the solo mic even though they were moving. Even with that the solo microphone sounded better than any other one.

First mix

To keep the whole mix consistent, I decided to sort tracks by instruments (choir only, choir with organ, organ solos, vocal solos etc.) and then mix all pieces in groups together.
After I had them put next to each other, I started mixing the simplest material - choral pieces. First, I switched off all of the compressors I used before and applied gain plug-ins instead to keep the peaking parts close to 0. I found the loudest part of all the choral pieces and set the gain on choir pair and reverb pair so that both of them would jump right up to 0.


Gain plug-ins used

I didn't use any other tracks. I panned reverb hard left and right and choir pair only by +9 and -9 (on a scale from -63 to 64), which underlined and clarified direct sound from the choir, but preserved space of the venue. Having these levels set up, I sent the mix into valve processors (first compressor and then EQ).


Mixer set-up for choral pieces

On valves, I set the levels up with mono signal so that left & right were the same back in Logic. I adjusted compressor to act in loudest parts only, and even then not to compress too much. EQ settings were: shelf below 200 Hz up a bit, band just below 3k down a bit, shelf above 6k up, high cut around 20k. All settings on both compressor and EQ were the same for both channels (I used 'Stereo' button).


Valve compressor settings


Valve EQ settings - first two modules


Valve EQ settings - last three modules

After setting all this up, I adjusted the level of the reverb so that it would add some space, but leave the main microphone pair clearly on the front.


Level meters measuring levels of signal coming in and out of the valves during bounce-down.

Result

Dynamics stayed virtually unchanged, but lifted up a little. Sound clearer, sort of closer, beautifully warmed up by valves.

Session 3 (in iMacs room on headphones)

First, I had a listen to mixed choral parts and decided that they sound great, and found that valves did good job. Then I went onto organ / choir tracks. During first listen, I had a closer look on the clipping on the organ channel during "Like as the Hart". At close examination it didn't show very severe, and was apparently inaudible, as it had only a couple of clipped peaks in area shorter than 1 second.


Close-up of one of the clipped peaks


Clipping affected area. Note that only some of the peaks are 'touching' the edges.

The balance and pan set up for choral pieces worked quite well with organ track switched on as well, and the only thing that should be done with those pieces is just to fine-tune the level on organ channel and then valve processing.


Mixer set-up for choir / organ tracks

None of that could be done using headphones, so I will have to take this to the Mastering Suite.
More mixing work was possible to be done on the Wesley track though - there is a soloist there, but not by the microphone, picked up quite well by the choir pair. She only sings alone, and never together witch the rest of the choir. To make her sound clearer, I decided to mix down the choir microphones into a mono track, increase its volume for the solo parts and then crossfade it with the choral pair.


Arrange view of the Wesley track

To increase the difference between choral and solo parts, I widened pan on the choir and turned the volume of the reverb down.


Automation on Wesley track

The solo mixdown came out with plenty of organ, so I EQed it to cut off most of it (low cut below 186 Hz (24db/oct) and low shelf below 590 Hz (-15 dB)) - this made solo on the middle, much louder than the rest. Thanks to EQ, the organ loudness doesn't jump.


EQ applied to solo track

Finally, I adjusted gain on choir (this time up by 1.3) using automation for both organ / choir tracks.


Gain plug-ins for organ tracks

Fitzwilliam Choir recording in Leicester

Intro
Last Saturday I went to St Philip's Church in Leicester to record a joint concert of Fitzwilliam College Choir and St Edmund's Hall Choir.

The venue
The church is quite large (to seat about 150-200 people easily) witch wooden arc-shaped ceiling, soft seats and carpet floor. This causes the acoustics to be really dry for this sized space. I was very happy with it though, because it gave nice and soft sound.

Performers
Over the course of the concert, the following appeared: Choir, organ, piano, string quartet, instrumental trio (organ, recorder and bassoon) and soloists. Unfortunately, when I was arranging the equipment and microphones for the event I was only aware of the choir, soloists and organ. Therefore I had to split the NTKs that I wanted to use as a stereo pair for soloists and use one of them for the piano. Luckily, the string quartet, bassoon and recorder were placed next to the organ and rarely playing together, so I could use one microphone for them all. It didn't come out as best set-up ever, but given the equipment I had it was the best I could do.

Detailed microphone set-up
For the main choir, I used a crossed pair of AKG C414s both set as cardioids, placed as close to the middle as possible without interrupting the conductor, to his left. I extended it up high just above his head so that it wouldn't interfere. Here are pictures:










Unfortunately, I couldn't place the reverb microphones (Røde NT5s – their small diaphragm would not pick up too much of noise rumble from the street, plus I have seen them in action as reverb microphones and they did excellent job) on the middle of the church simply because the boom microphone stand that I wanted to use is too bulky to be put right in front of the choir. Therefore, I put it as close to the middle as I could, extending the arm as long as it goes. Eventually I reached the place a couple of meters from the middle, approximately in the middle of the total height of the church. After some struggling up with the handles, I managed to set them up as a wide spread stereo pair. Here is how they looked like:











For the organ and other instruments I used another C414, but this time set up as a hypercardioid for it to concentrate only on the area where the instruments were. There was nothing behind it to spoil the effect. It was sitting on another boom stand, extended so that it would be almost above the instruments, facing downwards a bit (about 45 degrees):



The soloists were miked up using one of the Røde NTKs, placed on the height of their mouths, approximately 1.5 metres away from them. I instructed them to make a sort of semicircle so that the difference in their distances from the microphone would be minimal:



I placed another NTK by the piano. The instrument sounded a bit out of tune and very, very dump. To get the most of the (virtually non-existent) sharp attack, I placed the microphone facing the hammers section of the piano:







All above microphones were sent through a multicore to the room on the back, where the mixing desk and noisy recorder were. I used the kit that was freshly put together in ARU studios, consisting of Yamaha DM1000 desk and Fostex DV242LV. I chose to record in 24 bit resolution, 48 kHz sampling rate.


Hello!



Program:

• Beati Quorum Via – C V Stanford, choir only
  
• Like a Hart – H Howells, organ and choir
  
• Ubi Caritas – M Duruflé, choir only
  
• Haec est Regina Virginum – G F Handel, soprano and piano
  
• O my deir Hart – H Howells, soprano and piano
  
• Prelude in E – BuxWV141 – D Buxtehude, organ
  
• Let all mortal flesh keep silence – E Bairstow, choir only
  
• Ave Verum Corpus – W A Mozart, choir only
  
• Missa Brevis in G – W A Mozart, choir, organ and strings

INTERVAL

• Cantique de Jean Racine – G Fauré, choir and piano
  
• Sonaten in B flat – G P Telemann, recorder, bassoon and organ
  
• Bogoroditse Devo – S Rachmaninoff, choir only
  
• Miserere – G Allegri, choir and 4 soloists
  
• Praeludium in e – Nicolaus Brühns, organ
  
• Loveliest of Trees, Is my team ploughing – G Butterworth, tenor solo and piano
  
• Whither must I wander – R V Williams, tenor solo and piano
  
• Blessed be the God and Father – S S Wesley, choir and organ
  

Notes taken during recording:

• Piano sounds very wet and dump, no attack. Will have to be heavily EQ-ed.
  
• There is no clipping on any of the microphones during applause.
  
• There was a slight clip during Like as the Hart on organ track.
  
• Another one happened during Haec est Regina Virginum on soloists microphone. Levels taken down as soon as possible on both tracks.
  
• During Ave Verum Corpus, a table was moved in the far part of the church by one bored person who was selling tickets before. May cause problem.
  
• For instrumental trio, it will be good to use the choir microphones.
  
• There might be a clip on one of the reverb microphones during Miserere. Weird.
  
• Both tenor soloists were moving around a lot and didn't care to stand close to the microphone – they were right 90 degrees from its front – picked up plenty only some of them singng, will be very hard to edit. Preferably find another microphone to use.
  
• Organ microphone picks up a lot of choir, to be remembered when mixing.
  
• During last piece, soloist was standing together with the choir. Recommended to fiddle with pan width for solos / choral parts.
  

First Logic session
The first thing to do was to get the whole record from Fostex to Logic. What appears to be a simple task (just plug the Fostex optical output to the MOTU Firewire interface optical input), came out much harder. With some help from the studio technician, we finally managed to sync the clocks of Fostex, computer and MOTU, only to realise that all the recorded voices sound like everyone inhaled helium. After some more experimenting with clocks and sample rates we realised that although I am sure that I've done the recording with 48kHz sampling rate, the transfer between Fostex and MOTU only works properly when the sample rate on both of them is set to 44.1kHz. Therefore, to my disappointment, I now have the whole project in lower sampling rate than I was expecting. Still, the resolution is high – 24 bits, and after first listen to the record I am fully satisfied with the quality.
While writing all of the above, I was at the same time listening to the whole concert, searching for odd sounds and clipping. Also, I cut almost the whole record into piece-long regions. To make my first mix easier, I applied slight compression, mainly to increase overall levels. Some screen-shots (click to enlarge):







How does it sound?
The choir: Even stereo spread, natural and balanced sound. This pair also picked up a lot of organ and soloists, and obviously piano, which apparently has been picked up by all microphones.
Reverb: Wide stereo spread, sounds better than any Space Designer. No rumble, and very clear top.
Organ / strings: Thanks to hypercardioidal characteristics, it picked up a lot of organ and strings, even though the choir was very close. There is still a lot of choir present, but considering that they appear together rather rarely, it shouldn't be a problem when mixing. This microphone picked up a lot of organ keyboard sounds, but I think it will sound natural if I keep them in the mix. String sound natural and are clearly distinguished from the organ even when they play together – I guess that's because of good microphone positioning (directed towards strings, but still close to organ).
Soloists: Except for tenor pieces, where they didn't stand on the front of the microphone, NTK provided crystal-clear quality. Because soloists were standing right in front of the piano, their microphone picked a lot of it, but thanks to another microphone for the piano, it should be easy to separate them in the mix.
Piano: The instrument itself sounded, as I said before, very dump. there was virtually no attack to the notes and very little bottom. All of the content oscillated around 500Hz-2kHz and will have to be EQed to get it to sound nice. It might cause problem when mixing with reverb microphones which are full of it. This channel contains barely other sounds but the piano.

Problems found in the material
There was little amount of coughs and squeaking chairs. This all I think I will have to live with, although while going through the record I plan to reduce these noises as much as I can. I found one big thing though, and I am not sure where did it come from. On one of the reverb microphones (which eliminates someone hitting the stand, which would affect both of the microphones), there was a sudden 'pah' sound which peaked and clipped the signal for a half a second:



Because it basically looks like a massive, but single low frequency ripple, I used EQ with a steep (48dB/oct) bass cut below 196Hz, which reduced the 'PPAH' to little 'p':



There was also a little bell-like sound on the reverb microphones some other time, I will try to get rid of it somehow.

Summary
That's it for the first session. I am very happy with the raw material. I believe that the choral parts sound almost perfect with barely any mixing, and other pieces will require only slight interventions - a perfect example of recording well done from the beginning.