How to use mic positions in sample libraries

[Linos]

There are a lot of caveats to this post. I am not a mixing engineer. It's very possible that I've got things wrong, and I hope we can correct them in this thread. What I present here is what I have learnt from more than a decade of making mock-ups, reading about recording techniques and being present at several classical recording sessions. However, I am by no means a recording engineer, and I don't even consider myself to be particularly adept at mixing.

Also, I am approaching this through the lens of classical music, the sound of an orchestra playing in a concert hall. So it's different from what most people on this forum probably do. For other genres, such as film music, some mixing techniques will be the same. Others will be different. Perhaps [USER=11252]@NoamL[/USER] can shed some light on this. I don't know anything about recording film music, so I'm not the right person to ask.

Please bear all this in mind.

Recommended reading: ‘Classical Recording: A Practical Guide in the Decca Tradition’ by Caroline Haigh

Ann-Kathrin Dern has a good short explanation on her channel:


Getting started – one library, many mic positions

You may have been in the position of loading a sample library and being faced with an overwhelming choice of mic positions. Spitfire Audio's BBC Symphony Orchestra comes with what, 20 signals? What do we do with them? Should you use them all in your mix? God, no. No, no, no, no! If you use 20 mic signals all at once in your mix, you will get unintelligible garble. A Walla Group of orchestral mixing. So, what should you do instead? In many cases, just using your main mic signal will do the trick.


Selecting your main mic signal for the mix

Sample libraries often come with one or more ready-made mixes. Your first action can always be to try them out. If it works for your track, use it. But if they are not what you are looking for, then you need to create your own. You can do this as follows:

Often a Decca Tree is the main mic signal for a classical orchestral recording. Sometimes outriggers are added to increase the width of the sound. It's not wrong, but unusual, to use two or more stereo arrays in the same mix. If you have several stereo arrays in your library - Decca Tree, ORTF, Spaced Omnis - you will usually choose the most appropriate one for your track. This stereo array will be the only signal for your mix, or at least be the loudest one. You can add outriggers or side microphones if you need more width, and/or some room microphones (often called ‘hall’, ‘ambient’, ‘room’ or similar) for a more reverberant and spacious sound.

A common mistake is to mix in the close mic signals to all your instruments at all times. This is usually not necessary and can destroy the impression of natural depth. Why are you mixing in the close mics? Just because they are in your sample library? Turn them off. Because your mix sounds too wet or blurred? Turn down the level of your room mics instead. Try using your main mic signal without any other mics mixed in. Only if your main mic signal sounds too reverberant and undefined on its own should you consider adding some of the close mic signal. However, this is usually a sign that you have chosen the wrong sample library for your track and should use a drier one instead. The close mics are meant to accentuate specific instruments for specific phrases. They are not meant to be a constant part of the mix for all your instruments. The exception to this is brass and perhaps percussion. If the brass section does not sound clear and defined enough from your main mic signal alone, add some close mic signal. But be aware that often this isn't necessary at all.


Orchestral depth

The perception of distance is based on a combination of factors:

1. The further away a sound is, the more the high frequencies are attenuated. Bow noise, key noise, transients - all of these are reduced with distance.

2. More reverberant sound. More reverberation and less direct signal we interpret as a distant sound.

3. Stereo width. The further away an instrument is, the narrower the stereo width. When a large instrument occupies only a small part of the stereo field, we interpret this as distant. Conversely, if a small instrument occupies a large part of the stereo field, we interpret it as near.

All these aspects should work together in your mix. This will give you a sense of real depth. If they work against each other, you will not get a sense of depth. For example, if a sound source has little direct signal and a lot of reverb, but also a lot of high frequency content and a wide stereo signal, we can't interpret whether the sound is near or far. This can lead to confusion and inconsistent depth in your mix.

If you are using only the Decca Tree signal, you do not need to worry about this. The Decca Tree has a fixed distance to all the instruments in the orchestra. So, the instruments further away from it will have all the aspects mentioned above naturally captured. A French horn at the back of the stage will have less high frequency content than the strings, which are much closer to the Decca Tree.

However, if you add some close mic signal to give the French horn more definition, you now have two conflicting perspectives of depth in your mix. One from the Decca tree, telling your ears that the French horn is far away. And one from the spot mic, which tells you that the French horn is close. To increase the sense of depth, try narrowing the stereo field of the close mic and EQing out some of the high frequency content. I would not add too much reverb to the close mic, as this would defeat the purpose of adding clarity and definition. But keep an ear out for how much close mic you can add before it starts to confuse the sense of depth. Add just enough to add clarity without overpowering the sense of depth. Finding this balance is your job as the mixing engineer. Blending in close mic signal will always bring the instrument or section closer.

Blending in close mic signal

As described above, if you need to use close mics, add just enough to get the detail and punch you want, while still maintaining a sense of depth. Be aware that often the close mics are not needed at all, or only at certain points in the track. If you do need the close mics, it is perfectly fine to automate their level subtly throughout the track. If an instrument has a solo, bring the close mics up for it if necessary, and then fade them down again. Avoid sudden large changes in gain. If you raise the close mic by 6db all at once, the instrument will suddenly sound much closer. This sounds unnatural. Instead, increase the signal slowly so that you have the amount you need at the start of the solo. If you can hear a close mic as such in your mix, the close mic level is too high. Bring it down.


Panning the close mics

Close mic signals are not naturally panned. The microphone is placed in front of the instrument, about 1m to 1.5m away. The resulting sound will naturally be centred. To compensate for this, you need to pan the close mic signal. People often start by panning according to an orchestra seating plan. This is a sure-fire way of blurring the positioning of your instruments. Do this instead:

Listen to where your instrument is placed in your main mic signal (for example, the Decca tree). Slowly bring in some of the close mic signal. You will hear the image move towards the centre, as the close mic signal is not panned. Now pan the close mic signal towards the position in your main mic. Bring down the close mic signal. Slowly raise it up again. Does the image still move towards the center? Pan more. Does it move away from center instead? Dial back the panning. Repeat these steps until you don't hear any movement in the instrument position when you raise the close mic signal. Your close mic is now panned correctly for this mix.

This is quick and easy on a mixing console where you have one hand on the fader and the other on the panpot. With a mouse, it takes a little longer because you can only change one parameter at a time. But the results are the same.

Note how the correct panning can change depending on what your main microphone signal is. If your main signal is the Decca Tree you will need different panning than if it is the Outriggers. Using an orchestra seating chart will lead you to believe that one panning is correct for every situation in every mix. This will lead to fuzzy positioning and unfocused sound. Instead, use your ears and the method above to maintain the natural positioning of your main mic signal.


On to the next stages

If you have gone through all these steps, you have decided on the main mic signal for your track. You have decided if you need any extra mics (Outriggers/Wide/Sides to add width, Hall/Room/Ambient to add spaciousness). And you have decided if and where to use the close mics. Your close mics will be panned correctly if required.

From here, you can proceed to the other steps of mixing, such as volume automation, adding reverb, EQ, compression, etc.

What is your workflow regarding microphone positions? Any corrections, additions, other tips and tricks?

[GR Baumann]

Thanks Linos. Food for thought!

[Linos]

Thank you Georg. I am adjusting my template to these principles at the moment. One thing I noticed is that sample library developers often pan the close mics wider than what's in their tree/main mic. Berlin Strings are a good example, and if Inremember correctly, Spitfire Appassionata strings too. It seems to be common practice, with many libraries panning the close mics wide out of the box.

I will have to make comparisons between out of the box panning and the method outlined above. Out of the box you get close mic sounds spanning the whole stereo width. The main signal only covers about 75% of the stereo signal. That blurries the position of the instruments, and the side signal should sound drier. The method above should sound more natural, with well defined positioning. But it's narrower. Unless you widen the main mic signal

I have a lot of experimenting to do in the coming days and weeks.

[GR Baumann]

Your post is inspiring. Berlin Strings and the likes are not in my bucket, way too expensive and overkill for me, at the moment, but I think the option to Merge the Mix once you found the balance you are looking for is fantastic.

Yeah, very interesting and somewhat logical to narrow the stereo width of close mics.

I have no template, mainly due to hardware limitations, but I am going to fiddle around in HOOPUS too now.

[Linos]

Last touches to my template are done. All the close mics are muted by default now, but ready to be brought in for solos and the like. I think I am going to mix my next project twice, once in the old style and once with the new one. It will be interesting to hear the differences. It's already clear that I will be using less reverb. With the close mics gone from the main signal, there is less need for it.

[GR Baumann]

That should be interesting. I am not convinced yet if it is the right strategy to mute all close mics. I intend to see it rather like a balance of close, decca and outriggers. Wish I had a couple of different libraries to test that too, but well, I have you to do that, so I cannot complain.

[Lawrence]

Very interesting stuff, Linos. I’m looking forward to hearing your final conclusions.

[Linos]

Thank you, Larry. I will keep you both posted. I myself am curious to see how the mixes will differ. It should be an enlightening experiment.

[wst3]

It will absolutely be enlightening, and I would be willing to wager some small amount that most will like the new approach

While not following quite the same path, my approach can be stated as simply "use as few microphones as possible".

This started when I did, and I did not have a lot of choices for microphones. I would put up one (maybe two if I was feeling rich) room microphone and then add whatever was left to highlight different instruments. Almost exactly what you are doing, just a different motivation.

When I finally had access to more microphones (and 16 - count them 16 - tracks) I used them all. I can remember sessions where I had more than 16 microphones on the drum kit alone. Needless to say, those tracks sounded awful. A friend who was a very accomplished engineer took one listen and diagnosed the problem - too many microphones.

Now I'm not recording a lot of live instruments these days, and most libraries have already addressed the phasing problems between microphones, and yet I still start out small.

A drum kit, for example, gets one or two overheads (still haven't decided which I like better) and a few accent microphones, most often just two - one on the kick and one between the hat and snare.

For orchestral instruments I'll audition all the stereo pairs and pick the one I like best. I add in the room microphones a tiny bit for space, and I sometimes use close microphones as accents. Sounds awfully familiar eh?

Both of us seem to arrived at the same place, just widely different paths !!!

[Linos]

Interesting story, Bill. It seems that your practical approach has led you to the same point that my theoretical approach has led me to. My machine is powerful enough to handle a large template with multiple mics. Not having to compromise on mic selection made me stop and think. What do I really want to load into my template? This led me down the path of researching how classical and film music is recorded and mixed.

I'm now working on a mockup and will create two mixes. One with my old template and one with the new one. I am curious to see how different they sound and which one I prefer.

[wst3]

Add me to the list of the curious!

[Linos]

Here is the promised comparison with a real-life example. I'll call the two mixes 'Decca Mix' (adheres to the principles in the opening post) and 'Cinema Mix' (with close mics on every instrument/section).

The differences are subtle, but audible. To my ears the 'Decca Mix' has more depth, and it sounds warmer and rounder. To me it sounds more natural and in line with how I know classical recordings.

The 'Cinema Mix' has a brighter, more modern sound. It has more punch, but sounds flatter to me in comparison.


Reflection for Orchestra

Decca Mix

https://app.box.com/s/tops64e2347tot865kju8gdbjuufxeg0


Cinema Mix

https://app.box.com/s/xwi228d2bv0fxjhnbiss746l6bwdg497

(Sorry for not imbedding the audio files here. When I try to upload the mp3 files I get a 'HTTP error' in return)

The two mixes are not miles apart. Still, I prefer the Decca Mix in this case. The decca technique saves quite some resources too. There are many more mic signals in the 'Cinema Mix'. So that one uses much more ram.

You can hear the more vivid depth better without the strings.


Decca Mix Woodwinds and Brass only

https://app.box.com/s/nk5bbpku70f026jbu3guqqcru76hcbjo

Cinema Mix Woodwinds and Brass only

https://app.box.com/s/91hixcbhkfd6py777cfkj0asu8xuu8q6

[Edu]

Here is the promised comparison with a real-life example. I'll call the two mixes 'Decca Mix' (adheres to the principles in the opening post) and 'Cinema Mix' (with close mics on every instrument/section).

The differences are subtle, but audible. To my ears the 'Decca Mix' has more depth, and it sounds warmer and rounder. To me it sounds more natural and in line with how I know classical recordings.

The 'Cinema Mix' has a brighter, more modern sound. It has more punch, but sounds flatter to me in comparison.


Reflection for Orchestra

Decca Mix

https://app.box.com/s/tops64e2347tot865kju8gdbjuufxeg0


Cinema Mix

https://app.box.com/s/xwi228d2bv0fxjhnbiss746l6bwdg497

(Sorry for not imbedding the audio files here. When I try to upload the mp3 files I get a 'HTTP error' in return)

The two mixes are not miles apart. Still, I prefer the Decca Mix in this case. The decca technique saves quite some resources too. There are many more mic signals in the 'Cinema Mix'. So that one uses much more ram.

You can hear the more vivid depth better without the strings.


Decca Mix Woodwinds and Brass only

https://app.box.com/s/nk5bbpku70f026jbu3guqqcru76hcbjo

Cinema Mix Woodwinds and Brass only

https://app.box.com/s/91hixcbhkfd6py777cfkj0asu8xuu8q6

Hey Linos, great examples! I joined thesoundboard just to comment on this

I've been exploring and testing on mixing with samples a lot too, trying to achieve how recorded Decca Tree sounds, but I have not managed to have as clear 'image' as you using just the Tree mics.

Would you mind sharing which libraries did you use for these demos and how did you do the 'placement'? Did you change anything in their panning/balance?

For ex. with Violin I, usually when I use just Tree mics, be it from Spitfire, OT, EW, etc, it sounds either too 'full' (covering too much space left to right). But in yours it sounds still 'Decca Tree' but with just the right balance to be on the left but still 'sit in a space'.

I agree with Katherine's video premise but disagree that one can just let the 'tree mic' be. Usually if just using tree mics (even if using just single library collections), it will sound nothing like if a Decca tree had recorded those instruments in space. (Something which you did manage to achieve beautifully in your demos).

[Linos]

Hi Edu, welcome to the forum.

The strings are all Cinematic Studio Strings, with Cinematic Studio Solo Strings mixed in at low volume. The woodwinds are Vienna Symphonic VI Woodwinds, Berlin Soloists Oboe, Spitfire BBCSO and Spitfire Symphonic Orchestra. The brass is Sample Modelling French Horn, Cinematic Studio Brass, Spitfire BBCSO and Majestic Horn.
It's quite a selection of libraries. I have a big template with all the instruments purged on disabled channels. I can activate them at the touch of a button. This allows me to try out all the different libraries I have to see which one I like best for a particular part.

You might think that this would cause a lot of problems in the mix. But I have actually found that with the top-down approach, there is not that much to do at the mixing stage. Of course, you have to bring the dry libraries into a room with the more ambient libraries. I did this here for the VSL and sample modelling French horn using SPAT Revolution Essential. Spat acts as my stage placement tool. I used a room size of 2500m3. After that I added a bit of tail. I am not sure if I used VSS3 for this. But any decent reverb will do. I set the amount roughly to match the Spitfire Symphonic Orchestra instruments. Again, this is a step where people spend a lot of time and energy trying to get it just right. I think that energy is misplaced. The reverb doesn't have to be perfect. If the decay time and the colour of the tail are in the same ballpark, it will work just fine in the mix.

For the strings, I used the Cinematic Studio Strings 'Room mic' as my main pickup. This was to approximate the ambience of the Spitfire Symphony Orchestra. I added the 'Main Mic' at -6db to maintain depth. With the room mic alone, the strings would have been too far away. Adding the main mic brought them forward to the correct position. I added some reverb to the CSS bus. Again, to roughly match the ambience of the SSO. And that's all I did! No panning, no stereo trickery.

One thing I have learnt over time is that simplicity is often the key to mixing. Don't add processing just because you have the tools or because other people are doing it. Don't add tape saturation or EQ because you think you have to. Instead, only act when you hear a specific problem. Then the first step should always be to try to fix it at source. In the case of sample libraries, this means asking yourself if you can solve the problem by changing the mic selection or by adjusting the mix of the microphones you are using. If you can't, ask yourself if you're using the right library for the job, or if another library might be a better choice. If you are working with the right library and the right mic array/mic mix, then target the problem you are hearing. For example, if your violas sound a bit boxy, try an EQ cut on the viola signal around 1kHz or so.

Only add processing to solve specific problems. Frequency build-ups are quite common when working with samples. Targeting these with EQ will improve the clarity of your mix. But be aware that the sample libraries we use have been recorded by experienced engineers with excellent equipment. Trying to 'improve' their sound quality is usually doomed to failure. They already sound good if used well in the mockup.

After all this text, I'm not even sure I've really answered what you wanted to know. If you have any more questions, feel free to ask.