Large Ensemble Recording Problem

Profile picture for user Henrique Fantato
Submitted by Henrique Fantato on Sun, 01/14/2018 - 18:15

Hello all,

I need some tips for recording an Large ensemble( something about 60-70 musicians) in an large auditorium (300 people capacity).

It's a charity job for a project of my church, so the equipment is, unfortunately, low budget.

I actually have a Roland R26 recorder, a Audio Technica AT2020 and two 5 meters height tripods.

If it's really needed, i have a $200 budget that can be used to improve the equipment.

It isn't needed to record the auditorium's ambiance(it is an lecture auditorium, so the ambience isn't good) , and i'll use some Reverb plugins in the mixing.

We (me and the ensemble) already made some recordings, I used the roland recorder positioned at 5 meters away from the center-front of the ensemble(45 degrees down, in relation to the tripod), and the at2020 you can see at the picture( the idea was to try to reinforce the winds group, srry for the noob try xD).

Although, i will send a sample of this attempt...

Obs: i used Altiverb 7 (reverb), with the Berlin Philharmonic hall, using the 8m(on at2020 track), 12m(X/y recorder mics track) and 20m(Omni recorder mics track) mics simulations.

Any thoughts?

[MEDIA=audio]https://recording.org/attachments/hino-21-mp3.17951/[/MEDIA]

Attached files

hino 21.mp3 (7.7 MB) 

That's pretty good given the constraints you were working under. The excerpt you chose to post did not illustrate much in the way of dynamic variation, but we can certainly get an idea of (a) the quality of the players and (b) the acoustics (natural and otherwise).

You haven't said how you mixed in the AT2020. The AT2020 is a microphone that is not very well suited to this type of recording, and since you only have one of them, there is going to be a lot of off-axis sound from it.

What I would like to hear is the Roland X-Y unidirectional mics on their own, panned hard L and R, and with no additional reverb. That would give us more an idea of what you are dealing with by way of hall acoustics and the balance of sound from the performers.

Hi Henrique, Welcome to RO !

I'm no expert in big ensembles but here is a bit of food for your thoughts
With your 200$ you could buy another AT2020 and record with the 2 in stereo X/Y.
Or you can rent better mics. Many one shot shooters forget about renting gear. If you could rent a Royer SP24 stereo mic or a match pair of condensers (AKG, earthworks...)
That way you could offer better quality within your budget.

I'll let my more experienced friends suggest what mics you could rent...

Boswell, post: 455067, member: 29034 wrote: That's pretty good given the constraints you were working under. The excerpt you chose to post did not illustrate much in the way of dynamic variation, but we can certainly get an idea of (a) the quality of the players and (b) the acoustics (natural and otherwise).

You haven't said how you mixed in the AT2020. The AT2020 is a microphone that is not very well suited to this type of recording, and since you only have one of them, there is going to be a lot of off-axis sound from it.

What I would like to hear is the Roland X-Y unidirectional mics on their own, panned hard L and R, and with no additional reverb. That would give us more an idea of what you are dealing with by way of hall acoustics and the balance of sound from the performers.

Ok, I'll post a sample with only the X/Y without reverb.

And what do you mean with at2020 mixing? The positioning? (Sorry, I'm really excited about the recording universe, but I'm still very newbie, and here on Brazil this is a very hard bobby to learn :/)

pcrecord, post: 455069, member: 46460 wrote: Hi Henrique, Welcome to RO !

I'm no expert in big ensembles but here is a bit of food for your thoughts
With your 200$ you could buy another AT2020 and record with the 2 in stereo X/Y.
Or you can rent better mics. Many one shot shooters forget about renting gear. If you could rent a Royer SP24 stereo mic or a match pair of condensers (AKG, earthworks...)
That way you could offer better quality within your budget.

I'll let my more experienced friends suggest what mics you could rent...

The musicians are recording monthly, so I guess that buying something is a better choice. But I'll try to estimate how much it would cost to rent those mics.

Henrique Fantato, post: 455078, member: 51102 wrote: The musicians are recording monthly, so I guess that buying something is a better choice. But I'll try to estimate how much it would cost to rent those mics.

It all depends on what you expect regarding quality and what you are going to do with the recordings.
When you said it was for a charity, I though it was a one time event. That's why renting seems appropriate.
Boswell asked about the AT2020 because it's a mono mic. Mixing a stereo mic with mono mic is a bit tricky because of the possible phase cancellation.
As I said, I'm not an expert but if you really like to buy, I'd go for a second AT2020 and use only those two to record the ensemble.
with either of those config :


Of course 200$ is a very thin budget for ensemble recordings.
If you were to sell CDs or broadcast on radio, most engineer would invest in better preamps and mics.
It could go 4k fast in this kind of job.. ;)

pcrecord, post: 455079, member: 46460 wrote: It all depends on what you expect regarding quality and what you are going to do with the recordings.
When you said it was for a charity, I though it was a one time event. That's why renting seems appropriate.
Boswell asked about the AT2020 because it's a mono mic. Mixing a stereo mic with mono mic is a bit tricky because of the possible phase cancellation.
As I said, I'm not an expert but if you really like to buy, I'd go for a second AT2020 and use only those two to record the ensemble.
with either of those config :


Of course 200$ is a very thin budget for ensemble recordings.
If you were to sell CDs or broadcast on radio, most engineer would invest in better preamps and mics.
It could go 4k fast in this kind of job.. ;)

Yeah,$200 isn't nothing for this hahaha, actually, the charity is recording for free, because the equipment used in those recordings are from my uncle, and the "budget" is from myself....

In reality, I did a poor job on mixing(didn't know what to do), because I just got together the three audio files generated by the recorder(2 stereo files, one for X/Y mics, and another for the Omni mics; and 1 mono file from the at2020), normalized the tracks and applied some reverb. So any tip will be tremendous helpful :)

Boswell, post: 455067, member: 29034 wrote: That's pretty good given the constraints you were working under. The excerpt you chose to post did not illustrate much in the way of dynamic variation, but we can certainly get an idea of (a) the quality of the players and (b) the acoustics (natural and otherwise).

You haven't said how you mixed in the AT2020. The AT2020 is a microphone that is not very well suited to this type of recording, and since you only have one of them, there is going to be a lot of off-axis sound from it.

What I would like to hear is the Roland X-Y unidirectional mics on their own, panned hard L and R, and with no additional reverb. That would give us more an idea of what you are dealing with by way of hall acoustics and the balance of sound from the performers.

Posted the X/Y track

Henrique Fantato, post: 455160, member: 51102 wrote: Posted the X/Y track

Yes, thanks. I'm away from my studio at the moment, so although I can hear your track, I won't make any judgements based on computer speakers. I should be back in the studio by tomorrow, so will listen to it there.

Boswell, post: 455163, member: 29034 wrote: Yes, thanks. I'm away from my studio at the moment, so although I can hear your track, I won't make any judgements based on computer speakers. I should be back in the studio by tomorrow, so will listen to it there.

Ok

Sorry I haven't been able to get back to the studio to listen to this until now - we've had more snow here than we are used to, and things ground to a halt for a day or two.

Yes, the X-Y only track is interesting, partly because it is very dry. This surprised me, as from the photo you posted I thought that unwanted reflections would be a problem. I'll try adding some hall acoustic and see how that compares with your Berlin Phil hall. I have a Focusrite VRM box, and it's a good way of moving quickly between different acoustics for trial.

I think it sounds pretty good. I do hear the space altering the tone some, but I don't think it's necessarily inappropriate. With the right hall or whatever reverb on top of it, it might be just right.

Boswell, post: 455330, member: 29034 wrote: Sorry I haven't been able to get back to the studio to listen to this until now - we've had more snow here than we are used to, and things ground to a halt for a day or two.

Yes, the X-Y only track is interesting, partly because it is very dry. This surprised me, as from the photo you posted I thought that unwanted reflections would be a problem. I'll try adding some hall acoustic and see how that compares with your Berlin Phil hall. I have a Focusrite VRM box, and it's a good way of moving quickly between different acoustics for trial.

Yeah, the acoustic treatment of the auditorium is very good(the hole wall besides the musicians is treated with acoustic foam, and the rest of it is of wood), but so much dry...

Another question: The auditorium have a gallery and 3 floor that i can access, would it be a good place to set the at2020?(I only have some bad pics of it, sorry)

Attached files

bouldersound, post: 455332, member: 38959 wrote: I think it sounds pretty good. I do hear the space altering the tone some, but I don't think it's necessarily inappropriate. With the right hall or whatever reverb on top of it, it might be just right.

So what can I do? Do I use only the X/Y track with reverb, or all the tracks, or perhaps only the omni track?

Boswell, post: 455330, member: 29034 wrote: Sorry I haven't been able to get back to the studio to listen to this until now - we've had more snow here than we are used to, and things ground to a halt for a day or two.

Yes, the X-Y only track is interesting, partly because it is very dry. This surprised me, as from the photo you posted I thought that unwanted reflections would be a problem. I'll try adding some hall acoustic and see how that compares with your Berlin Phil hall. I have a Focusrite VRM box, and it's a good way of moving quickly between different acoustics for trial.

is it like an "portable soundcard"?

BTW if you use a portable recorder like your Roland R26, deactivate any auto volume or compression functions (if it's not already done)
the recordings will be cleaner and more natural.

pcrecord, post: 455449, member: 46460 wrote: BTW if you use a portable recorder like your Roland R26, deactivate any auto volume or compression functions (if it's not already done)
the recordings will be cleaner and more natural.

Ok, but I guess that I'm not using anything like that. I'll check it, thx!

Henrique Fantato, post: 455451, member: 51102 wrote: Ok, but I guess that I'm not using anything like that. I'll check it, thx!

I thought it would be good to say for anyone coming here from a google search ;)

Henrique Fantato, post: 455448, member: 51102 wrote: is it like an "portable soundcard"?

It's more than that, as the original purpose of the VRM box was for checking mixes by emulating different loudspeakers in different acoustics. However, I also use it from time to time with a laptop as a straight headphone output without the VRM processing, as it's so much better than the internal soundcards in most laptops. I'm travelling quite a bit at the moment, so it's convenient to have a USB-powered small box and a pair of headphones to take with me along with my laptop.

It's a pity that Focusrite discontinued the VRM box as a product; I guess its sales volumes were not sufficiently great to warrant future investment, such as moving to USB3. The VRM software is, of course, still available to run with their main audio interfaces.

Boswell, post: 455453, member: 29034 wrote: It's more than that, as the original purpose of the VRM box was for checking mixes by emulating different loudspeakers in different acoustics. However, I also use it from time to time with a laptop as a straight headphone output without the VRM processing, as it's so much better than the internal soundcards in most laptops. I'm travelling quite a bit at the moment, so it's convenient to have a USB-powered small box and a pair of headphones to take with me along with my laptop.

It's a pity that Focusrite discontinued the VRM box as a product; I guess its sales volumes were not sufficiently great to warrant future investment, such as moving to USB3. The VRM software is, of course, still available to run with their main audio interfaces.

What headphones do you find most useful when using the emulation software?

dvdhawk, post: 455458, member: 36047 wrote: What headphones do you find most useful when using the emulation software?

It depends where I am. If I'm in the studio and want to do real VRM work, I'll try a couple of different ones (Sony MDR7510, AKG K702) and spin round the loudspeaker emulations to check for problems. If something shows up, I'll make sure that it's not a headphone effect by trying others (I have an old pair of Panasonics that are surprisingly good but a bit bassy). Generally speaking, if one set of phones shows a possible problem, it's there on the others as well.

I don't like carrying studio headphones on a flight, so I have a few cheaper pairs to use when travelling, and also earbuds (which I hate wearing).

Observations from the initial mixed track to the dry one. The photo reveals some interesting points. The woodwind are much closer to the conductor and main mic position so the strings sit nicely until the woodwind join in, so the balance gives them a little too much presence in the blend. Maybe you could alter the physical layout to compensate, adding a bit more depth centre, and bringing back the conductor a little. You have instruments spread over 180 degrees around the mic position, that's a little too much for convention, so bringing the mic, and the conductor back closes this up a bit. The dry space has few real issues with unwanted reflections it seems, so if it was me - I would buy another 2020 and record in X/Y, which will also help with the somewhat prominent centre instruments. I thought your reverb choices worked well, and would happily listen to this bunch of musicians recorded by you. In fact - for a beginner, that's a pretty decent recording, far better than many I hear!

1. add the second 2020
2. move the mic (and conductor) back a little

You also need to be very careful with normalisation because there are conventions for orchestral works, and if a piece of music spends most of it's time in the ppp/pp range with maybe an occasional flurry into f, then it will be presented as a finished recording as quiet, compared with a piece that never goes into p, and is forte plus throughout. So if a piece is 3 movements with one quiet, one average and one very loud - normalisation doesn't work - normalised for what???? A quiet piece needs a quiet maximum level, and a loud piece needs a much higher one. The best advice I can give is to go with convention and find some commercial recordings of quiet/medium/loud works and see how they stack up level wise. You might be very surprised how quiet they are. There don't seem to be any standards, and different producers do go up and down in levels. It's also why people like the BBC often appear to be quieter on their classical music radio station than the commercial stations - when they play quiet works, people complain that in the car, they can't hear them! Normalising, however, to a constant level, irrespective of the music, is not appropriate to this style of music. It is of course for pop/rock etc, but not for classical.

paulears, post: 455467, member: 47782 wrote: Observations from the initial mixed track to the dry one. The photo reveals some interesting points. The woodwind are much closer to the conductor and main mic position so the strings sit nicely until the woodwind join in, so the balance gives them a little too much presence in the blend. Maybe you could alter the physical layout to compensate, adding a bit more depth centre, and bringing back the conductor a little. You have instruments spread over 180 degrees around the mic position, that's a little too much for convention, so bringing the mic, and the conductor back closes this up a bit. The dry space has few real issues with unwanted reflections it seems, so if it was me - I would buy another 2020 and record in X/Y, which will also help with the somewhat prominent centre instruments. I thought your reverb choices worked well, and would happily listen to this bunch of musicians recorded by you. In fact - for a beginner, that's a pretty decent recording, far better than many I hear!

1. add the second 2020
2. move the mic (and conductor) back a little

You also need to be very careful with normalisation because there are conventions for orchestral works, and if a piece of music spends most of it's time in the ppp/pp range with maybe an occasional flurry into f, then it will be presented as a finished recording as quiet, compared with a piece that never goes into p, and is forte plus throughout. So if a piece is 3 movements with one quiet, one average and one very loud - normalisation doesn't work - normalised for what???? A quiet piece needs a quiet maximum level, and a loud piece needs a much higher one. The best advice I can give is to go with convention and find some commercial recordings of quiet/medium/loud works and see how they stack up level wise. You might be very surprised how quiet they are. There don't seem to be any standards, and different producers do go up and down in levels. It's also why people like the BBC often appear to be quieter on their classical music radio station than the commercial stations - when they play quiet works, people complain that in the car, they can't hear them! Normalising, however, to a constant level, irrespective of the music, is not appropriate to this style of music. It is of course for pop/rock etc, but not for classical.

Thanks for the advices! To modify the musician's layout, I'll need to talk with the conductor, but I guess that it will be possible. We don't have enough space behind the woodwind to bring they back, so I guess that it would be more convenient to modify the Recorder position.

And talking about the normalization, I've normalized all the audio files equally, so it doesn't alter the proportion between the piano and fortíssimo movements. Is recording with the "right volume(changing the sensitivity of the mics)" better than normalizing it on the edition?

By the way, thanks for the compliment!

That's what I mean - have you got a figure to normalise to? As in if it is a piece with some very quiet bits, that means you normalise to a different level to the loud piece that follows. If one is peaceful and gentle and the next is raucous - these need different maximums - The noise floor with no singing should be the same, no matter how loud the piece. This is the norm for choral and orchestral music.

paulears, post: 455556, member: 47782 wrote: That's what I mean - have you got a figure to normalise to? As in if it is a piece with some very quiet bits, that means you normalise to a different level to the loud piece that follows. If one is peaceful and gentle and the next is raucous - these need different maximums - The noise floor with no singing should be the same, no matter how loud the piece. This is the norm for choral and orchestral music.

I have access to the conductor sheet score, so i can make the normalization and put all the audios files in the right sequence. By the way, to make the "white noise" in the spaces without sound, I used some parts of the audio that anyone is playing. Is that correct?



So what record level would you think ppp is? In the example above of two tracks from an commercial recording - one is very quiet and the other starts quiet and gets much louder. As long as you don't normalise them both to the same level, it's correct. So you need to pick your own maximums. For me, I would have recorded them both a little lower - the loud one is VERY loud, perhaps too much?

paulears, post: 455559, member: 47782 wrote:


So what record level would you think ppp is? In the example above of two tracks from an commercial recording - one is very quiet and the other starts quiet and gets much louder. As long as you don't normalise them both to the same level, it's correct. So you need to pick your own maximums. For me, I would have recorded them both a little lower - the loud one is VERY loud, perhaps too much?

So I guess that I normalized correctly, because I've normalized both applying the same gain. For example, lets say that the ppp part is at 30dbs, and the loud one is at 70dbs, I've applied a 15dbs gain on both, making the ppp part on 45dbs, and the loud one on 85dbs; so I kept the proportion between then.

I think so, maybe?

I've just got a kind of feeling that something in your process is just not quite right. Being honest, the noise performance we now work with is so good that different techniques don't really make much difference.

People have their own systems - some, old school perhaps like me still like to record quite hot, and despite of the range available, have a natural resistance to leaving too much headroom - and I still get unknown musicians or singers to give me their loudest bit, which I still know they will exceed - and then I set their levels for recording to give me at least 6dB safe headroom at the top. Occasionally, I still get it wrong. I have colleagues who leave much more headroom, but then the battle is at the bottom end - where the danger starts to be that the wanted and unwanted quiet stuff gets tricky to manage.

Most of my music falls into0 two distinct types - tracks for live performances, and classical piano - and if I'm honest, I really do not have a specific peak level for this at all. In fact, for continuity, I always just listen to previous masters so I make sure my piano sounds the same(ish) as last time - and as the volume in the studio is never adjusted, my levels tend to be a similar level for a similar style. Often I'll notice a peak light has come on - only at that point do I think about gain, and it usually means channel or master faders are in the wrong place, but volume is right, so I'll drop one fader, raise another and all is well. I won't need to run the normaliser routine at all. For the live track stuff, I will, simply because the users need it presented differently - it could be stereo, or two track, or multiple tracks - and in this case - if there is a quiet flute - I DO normalise it to the high side, simply because the live sound op needs to see it obviously on the meters - the channel fader will be run low, of course - but the track will be hot. The average level on the meters of the brass and perhaps the harp, or even wind chime type sounds all need to be higher so they can be dealt with like live sound - where you turn the gain up until the red lights just don't come on, and if they do occasionally on a peak, nobody notices. My tracks need to replicate this.

We've drifted well of the topic really, but I was a bit worried that you'd misunderstood what I was try to say. My way of working works for me, and seems to avoid snags. Others will laugh at how I deal with things - but all I'm really interested in is music recording that works - and I just smile when I hear younger digital age people talking about very specific sound levels on a meter - I always aim for -23.5dB on the meters. I had to boot up Cubase to discover that my common working level is -9 on the Cubase meter, yet on Adobe Audition, my natural working level is more like -12! I have no idea why this is - but when I shove a fader this seems right. I don't know why the two programmes should be different? Maybe it's how the meters are designed, and react to peaks - I genuinely don't know.

paulears, post: 455565, member: 47782 wrote: I think so, maybe?

I've just got a kind of feeling that something in your process is just not quite right. Being honest, the noise performance we now work with is so good that different techniques don't really make much difference.

People have their own systems - some, old school perhaps like me still like to record quite hot, and despite of the range available, have a natural resistance to leaving too much headroom - and I still get unknown musicians or singers to give me their loudest bit, which I still know they will exceed - and then I set their levels for recording to give me at least 6dB safe headroom at the top. Occasionally, I still get it wrong. I have colleagues who leave much more headroom, but then the battle is at the bottom end - where the danger starts to be that the wanted and unwanted quiet stuff gets tricky to manage.

Most of my music falls into0 two distinct types - tracks for live performances, and classical piano - and if I'm honest, I really do not have a specific peak level for this at all. In fact, for continuity, I always just listen to previous masters so I make sure my piano sounds the same(ish) as last time - and as the volume in the studio is never adjusted, my levels tend to be a similar level for a similar style. Often I'll notice a peak light has come on - only at that point do I think about gain, and it usually means channel or master faders are in the wrong place, but volume is right, so I'll drop one fader, raise another and all is well. I won't need to run the normaliser routine at all. For the live track stuff, I will, simply because the users need it presented differently - it could be stereo, or two track, or multiple tracks - and in this case - if there is a quiet flute - I DO normalise it to the high side, simply because the live sound op needs to see it obviously on the meters - the channel fader will be run low, of course - but the track will be hot. The average level on the meters of the brass and perhaps the harp, or even wind chime type sounds all need to be higher so they can be dealt with like live sound - where you turn the gain up until the red lights just don't come on, and if they do occasionally on a peak, nobody notices. My tracks need to replicate this.

We've drifted well of the topic really, but I was a bit worried that you'd misunderstood what I was try to say. My way of working works for me, and seems to avoid snags. Others will laugh at how I deal with things - but all I'm really interested in is music recording that works - and I just smile when I hear younger digital age people talking about very specific sound levels on a meter - I always aim for -23.5dB on the meters. I had to boot up Cubase to discover that my common working level is -9 on the Cubase meter, yet on Adobe Audition, my natural working level is more like -12! I have no idea why this is - but when I shove a fader this seems right. I don't know why the two programmes should be different? Maybe it's how the meters are designed, and react to peaks - I genuinely don't know.

I guess that I understood now, It's like everyone has your particular way to deal with it(peak levels), right? My only fear is that working with a minimal safe zone would be a risk, because the recorder is the only way that I can see the peaks levels, and it stand 5 meters high.... so the only things that I use to monitor the recording is my headphone plugged in the recorder, so in my case, I always set the sensitivity of the mics to record the fff part at -12dB, and if it seems to low on the Reaper, I select all the audios files and apply a common gain (6-8dB) to get a better volume. So I guess that, on the next recording, I will need to be more worried about the mics positioning, to not have any problems with the woodwind section and the final result. I'ill talk with the administrator of the event to see with we could get another at2020 to get a better result.

Talking about mics, what would be a good mic to record this ensemble with a good price/benefit ratio?

Henrique Fantato, post: 455621, member: 51102 wrote: Talking about mics, what would be a good mic to record this ensemble with a good price/benefit ratio?

Personally, I would not be happy using a single mono microphone for this task. It may be the case that using the money to get a single stereo microphone or a matched pair of mono microphones would not result in quite the same sonic quality as a single mono microphone, but I think the listening to the recording would be a much better experience.

I'm assuming you intend to continue using the Roland R26 recorder and therefore are looking at microphone(s) that would give the output levels needed for that device. It's difficult to give suggestions without knowing whether your strict price limit is the $200 that you mentioned in your first post.

If you were to go for a pair of small-diaphragm condenser (SDC) microphones, this would allow you the flexibility of setting them up in ORTF and other configurations that would cope better with a wide sound stage than a single-point microphone or co-incident pair. On the other hand, spending the budget on adding a second large diaphragm condenser (LDC) that has switchable patterns for use in combination with your AT2020 would give you an M-S option that would be less sensitive to microphone type differences than trying to use a mis-matched pair in X-Y configuration. My quick reading of the R26 specification does not mention anything about M-S decoding, so it may be that you could not play back an M-S recording on the R26 without transferring the files to a DAW.

A few years ago I bought a couple of multi pattern large diaphragm condensers (against the views of many friends and internet colleagues) they've always been good for trying things out, and I rather like the sound of them. Having Omni, figure-of-8, cardioids and hypers on the various mics make trying things out pretty easy. I've also got some small diaphragm ones with swappable screw on capsules - If you search out some of the bigger Chinese manufacturers, rather than the sellers on AliExpress, there are some rather nice mics out there for your budget. You'll have to take risks of course, but that's fun too.

If you want to be safe - I think the advice for one mic with switchable patterns could work, but a pair is always best - because the X/Y or ORTF techniques are just simpler.

Boswell, post: 455639, member: 29034 wrote: Personally, I would not be happy using a single mono microphone for this task. It may be the case that using the money to get a single stereo microphone or a matched pair of mono microphones would not result in quite the same sonic quality as a single mono microphone, but I think the listening to the recording would be a much better experience.

I'm assuming you intend to continue using the Roland R26 recorder and therefore are looking at microphone(s) that would give the output levels needed for that device. It's difficult to give suggestions without knowing whether your strict price limit is the $200 that you mentioned in your first post.

If you were to go for a pair of small-diaphragm condenser (SDC) microphones, this would allow you the flexibility of setting them up in ORTF and other configurations that would cope better with a wide sound stage than a single-point microphone or co-incident pair. On the other hand, spending the budget on adding a second large diaphragm condenser (LDC) that has switchable patterns for use in combination with your AT2020 would give you an M-S option that would be less sensitive to microphone type differences than trying to use a mis-matched pair in X-Y configuration. My quick reading of the R26 specification does not mention anything about M-S decoding, so it may be that you could not play back an M-S recording on the R26 without transferring the files to a DAW.

M-S recording is just a type of mic positioning, right? I guess that I'm probably wrong, but Isn't using 2 distinct mic in X/Y or M-S config a problem?

X-Y certainly would present problems, as I mentioned. Two different mics in M-S can often be used, particularly to deal with difficult acoustics. I'm talking here about different microphone models, not just different patterns. Note that if one of the mics is a velocity-sensitive ribbon, the other has to be one too.

M/S doesn't need identical microphones, in fact, it needs them to be different - best to think about it as a mono technique that has stereo width.

One mic does the recording in what is effectively mono - so a forward facing cardioid that covers the entire space is the first type. You then use a figure-of-8 mic that has a deep dull forward, and it captures everything left or right of centre. something bang in the middle has no left or right component, but the leftmost musician gets picked up in both, as does the rightmost musician - but there is a phase relationship - so somethings add and some things cancel. The benefit is that on your mixer, you have one fader for the mono signal, and the other fader brings in stereo - as much, or as little as you want!

Downside is that before you can listen properly, you need to extract all this information back into a conventional left and right. If you listen to an M/S recording in normal left and right, especially on headphones - it's plain weird!

paulears, post: 455691, member: 47782 wrote: M/S doesn't need identical microphones, in fact, it needs them to be different.

Not at all - one of my favourite methods uses two identical fig-8s (MS-Blumlein).

Ah - I'd never considered Blumlein as M/S. Maybe I should think again, but normally Blumlein is used 45 degrees either side of centre so there isn't a centre channel, however I'm now intrigued that if you rotate the pair 45 degrees you would have a forward channel and side - and could then matrix it? This would then be M/S, bit with a huge rear component - but this could be good for some things.

Two things come to mind. My recordings with Blumlein technique were all, I thought, poor - and I blamed this on the space being too messy, so stereo field was very confusing, but it never occurred to me to try it rotated and matrix it? Does that work?

Yes, it does work well, but gives good results only in the right acoustic. Managing the rear sound field is the usual problem.

In the previous post, I had not properly made the distinction between Blumlein (fig-8s at +/-45 degrees either side of the centre line) and MS-Blumlein, which is one fig-8 mic facing front/back and the other at right angles. The MS-Blumlein needs standard MS decoding.

Years ago, I was intrigued by the symmetry of MS-Blumlein, particularly the geometric fact that you could use it to sit four performers each directly facing a mic. This meant in principle that you could arrange the polarity of the two mic signals for each player such that they were in the centre of an MS field, building up 4 stereo tracks. When mixed, these should produce a perfect sound, right? When I tried it using a B&O BM5, the result was awful. The rear signal of each player was of course the front signal of the opposite player, so while I got four nice-sounding stem tracks, the phasing was such that the sound field collapsed when mixed together. An early lesson.

Henrique Fantato, post: 455557, member: 51102 wrote: By the way, to make the "white noise" in the spaces without sound, I used some parts of the audio that anyone is playing. Is that correct?

Ok I didn't read the whole thread but why put noise in the silent part of a recording and no you don't take the recording of someone playing to make white noise.
The point of white noise is to have a noise that have the same energy at all frequencies. The noise from a recording will have a frequency curve induced by the mic response. the room acoustics, preamp, etc.
I guess what you talk about is putting room noise between playing parts so the audio doesn't stop drasticly. In that case, taking the noise when no one is playing is a good choice.. but we can't call that white noise.. technically at least.

paulears, post: 455691, member: 47782 wrote: M/S doesn't need identical microphones, in fact, it needs them to be different

Hey guys, correct me if I'm wrong. M/S and Blumlein are different, right ?
M/S would be a cardioid mic and a figure of eight mic together (could be the same mic but different patern selected) and we flip the polarity on the side mic.
Blumlein would be 2 figure of Eight mics and no polarity switch, yes ?
Or is it that blumlein becomes M/S when one of the mic is facing the source and we flip the polarity of the other one??

Just thinking out loud here ;)

M-S is a positional coding method and does not dictate what type or pattern of microphones are used. It's often thought of as a cardioid M and a fig-8 S, but that's only one implementation.

The important thing that even some professionals get wrong is that you can't mix velocity-sensitive and pressure-sensitive types of microphones. This is because the decoding goes wrong due to the 90 degree phase difference between the two transducer types.

Boswell, post: 455707, member: 29034 wrote: The important thing that even some professionals get wrong is that you can't mix velocity-sensitive and pressure-sensitive types of microphones.

Yeah I remember we talked about this and I made many tests. A condenser and a ribbon doesn't work in MS with a condenser...

Hang on - we're a bit mangled now - velocity sensitive? That's a new one on me. Maybe just British, but we have pressure gradient and pressure operated as the two types of mic, with the mics with rear vents operating in the PG mode, and single entry mics being pressure operated. isn't the velocity of sound fixed, as in 300m/s? Is this a US used term that doesn't;t travel well?

I've had no troubles with mixing a ribbon with a dynamic or condenser. My M/S preference is actually two condensers, but if I do use a ribbon (as I only have one!!) I'd match it with a condenser normally. I've tried hypers too with some success in sport, and even a short shotgun - for music they're not too nice - the side movement isn't linear. You get all dead centre, but then noises left and right leap out - great for football and sport.

Yes, there is a confusing mix of terminologies. Basically, fig-8 ribbons generate an output according to how fast the ribbon is moving through the magnetic field (velocity-sensitive). In the case of an acoustic sinewave, maximum velocity occurs at zero displacement. Most other types of microphone transducer give maximum (peak) output at maximum displacement or pressure. This is in quadrature with the velocity, hence the two types will not decode to L-R correctly when used as an M-S pair.

This is the first time I've heard of this, but your explanation does explain it very well, and I'd just never thought of this at all. Is the phase lag recoverable if the mid channel was delayed through 90 degrees?

paulears, post: 455720, member: 47782 wrote: Is the phase lag recoverable if the mid channel was delayed through 90 degrees?

I can't answer that, but it does make me ask: delayed through 90° at what frequency?

My thought are that this lag is the same as when two microphones are separated in space - but that's obviously in cm sizes. In the case we're talking about here, the phase error is less than a half cycle, so probably any electronic shifting in the time domain couldn't do this level of micro shifting in time. The difference between a ribbon side and a condenser side appears to be so small that I'm surprised it's critical - especially as physically the diaphragms of the two mics are not time aligned anyway. If the problem is the shift in the time alignment as frequency rises, so it isn't a constant - doesn't this also happen in non-ribbon diaphragms.

To be honest, my thinking is getting very circular on this - every attempt to understand why the condenser vs ribbon doesn't work also suggests that it shouldn't work in the other examples of M/S. As Yoda said - Confused now, I am.

It has nothing to do with time or spatial alignment. This is a phase issue, and it's a constant frequency-independent 90 degree difference. It's easy to see and understand with sinewave sources, less easy with complex waveforms. Whether you see it as a lead or a lag depends only on which face of the ribbon you designate as the front.

The only type of device that can align the two signals is an all-pass phase shifter. It's one of the reasons I have an Audient Mico pre-amp with a +/-180 degree continuous phase adjustment on the second channel.

Does that completely solve the problem and restore the stereo field to what it should be?

Perhaps a draft question - but is this adjustment only visible in the scope, or can you tune the phase shift by ear? I guess you'd need to have a central tone source, and then look at the X/Y display on the display, or is this too simplistic and easily wrecked by room reflections?

Learned quite a bit here today, to tuck away for future use. Cheers.

EDIT
Just been thinking more - (dangerous). I have a stereo mic - two omni/cardioid/fig-8 capsules one above the other. Made for me as a special from a Chinese supplier I've dealt with for years and I've used this for Blumlein (in the rotten room) and M/S recordings, although most of it's life is spent in X/Y mode. As it derives it's fig-8 from coupling two back to back cardioids together, there must be similar phase errors created by the coupling of the two? Obviously the ribbon is the only practical 'real' fig 8 capable design - has anyone discovered that the large diaphragm condenser different modes do similar things? Clearly, all the omni have little dips in the polar pattern where the patterns interact - is this as destructive to coherence as the ribbon/condenser combination?