Nice Comparison of Recording Techniques

Creating a home studio for recording the classical guitar. Equipment, software and recording techniques. Amplification for live performance.
User avatar
rojarosguitar
Posts: 4178
Joined: Sat Sep 19, 2009 12:24 pm
Location: near Freiburg, Germany

Nice Comparison of Recording Techniques

Post by rojarosguitar » Fri Sep 08, 2017 9:13 pm

There is a nice Blog explaining different stereo recording techniques for the classical guitar, underpinned by actual recorded examples.
Just search 'How To Record Classical Guitar' by Uros Baric.

I tried to post a comment on it, but the blog is old and I'm not sure if it still will be moderated. Just not to have the thoughts lost, for the record:

Hi Uros Baric,

Many thanks for your instructive experimentations with recording techniques. There are not many attempts to explain it that clearly and underlined with sound samples of any consistency.

Also congrats for the very consistent repetition of the same piece!

Now, the only thing we all will agree on is that people do not agree on what is the ‚best’ sound, and that’s why it’s so important to know all of these techniques and also to know whether they’re going to work for the intended purpose or not.

Because I respect your work very much, only a few general remarks (from my own experience with classical guitar):

– Omnis and everything involving omnis (like MS with omni middle channel) will work best in a very good sounding. In theory one could bring the mics closer to the player with omnis because there’s no proximity effect, but recording classical guitar in near field is not at all desirable sound-wise IMHO. Bringing them further away from the guitar increases the proportion of the room sound in the pickup of the mics and here’s where all the problem start with not so great rooms.

Your room is apparently very well treated so that your omni recordings all sound great.

– MS is very critical on the mid to side ratio. You really have to tweak that until the localization and spaciousness are just right (guitar doesn’t need a ‚spread 🙂 ). Also MS in its proper sense depends heavily on use of a real figure eight side mic. If you don’t do that, you could just as well use a stereo pair for the side and a third mic for the mid. Then you don’t need to go through the pain of matrixing it, don’t need to buy a real fig 8 mic (expensive), have more spaciousness (because now you can go near coincident like e.g. an ORTF scheme plus a mid mic mixed to taste). I have done that a lot for classical recordings and it works very well!

And if you really go MS, it’s worth to spend some money on the side mic, they can make a whole ot of difference. They should be also very quiet!

– depending on the mics you use (how much off axis coloration they have) ORTF (which has been developed to record whole orchestras) may work or not, because you are sitting well inside the middle axis of the ORTF array and quite close, so you get what the mics get at this not so useful angle. This may actually be a bonus with very bright cardioids, off axis they sound softer, but maybe not too great.

– ORTF and some other near-coincident techniques (like NOS etc.) absolutely depend on usage of directional mics (cardioids mostly). If you do ‚ORTF’ with omnis, it’s ORTF only by name, because in reality it is AB with 17cm spread and strongly off axis mics. The same goes for all other near coincident arrays that always use both, a run-time and loudness differences. Yu don’t have loudness differences just by angling an omni.

– If you need mono compatibility, from the whole lot only XY and real MS will do it.

– (this is mainly for all the people listening to your samples) For critical listening to classical guitar recordings listening through headphones (may they be as good as it goes) isn’t enough, you really have to listen through a pair of good loudspeakers. Only then you will know how naturally the guitar sits in the room and whether there are phase problems etc etc…

Well, that’s a lot of stuff, but I hope it will help also people who are trying to find out the best way to record classical guitar.

As to my own preferences:

I quite like most of your closer spaced AB. Whether you use cardioids or omnis of course depends on what you have and like, but in a half-ways decent room and with some experimentation as to the players position and mic placement it’s almost certain to get good results. With larger spacing of AB there is always some kind of lacking focus, so I would avoid it.

Your ORTF seems to work very well in the room you have.

XY is also very nice, though a bit too compact to my taste. Maybe a spatial shuffler plug in would help that.

MS with cardioid is great. MS with omni seems to exhibit some phase problems in my ears and also a slight instability of the image on loudspeakers.

That’s it so far … I really admire the work you have done. Keep going!

Music is a big continent with different landscapes and corners. Some of them I do visit frequently, some from time to time and some I know from hearsay only ...

SteveL123
Posts: 423
Joined: Thu Feb 09, 2017 5:05 pm

Re: Nice Comparison of Recording Techniques

Post by SteveL123 » Sat Sep 09, 2017 2:07 pm

Can you spell out the abbreviations MS ORTF etc so those of us who are not recording engineers but interested in learning can at least look them up?

User avatar
rojarosguitar
Posts: 4178
Joined: Sat Sep 19, 2009 12:24 pm
Location: near Freiburg, Germany

Re: Nice Comparison of Recording Techniques

Post by rojarosguitar » Sat Sep 09, 2017 4:31 pm

:D Yes, of course. The post was first of all intended as a record. But I'm willing to answer any question...

We are talking here about the different microphone arrays used in stereo recording, usually consisting of two microphones, usually of same making.

First of all the directional characteristics:
  • 'omni' is an abbreviation of 'omnidirectional', meaning that it doesn't prefer any direction in space for pickup (which of course is a big simplification). Pics up a lot of the room, so not a great option in bad sounding rooms
  • cardioid is a microphone with one preferred direction (again simplified).
  • figure 8 is a microphone of two preferred directions lying opposite each other - literally like an eight
Now as to the different arrays usually deployed in stere recording with 2 microphones (there are techniques using more than 2 mics as well):
  • AB - mostly used with a pair of 'omnis' one of them being called A and the other B - belongs to the family of so called 'spaced' arrays. It uses mostly time of arrival cues for our hearing something as Stereophonic.
  • XY - mostly used with a pair of cardioids, one called X and the other Y, (and as I pointed out above depends absolutely on the mics used being directional, because here only the volume differences in different directions are providing the stereo cues) - belongs to the family of so called coincident arrays, because the diaphragms of both mics are brought together as close as possible (which has obvious limitations with real world microphones)
  • MS - abbreviation of Mid-Side recording, is another 'coincident' technique deploying a pair of microphones, one of which (S - Side) by definition has to be a true Figure 8 microphone, whereas the other (M - Middle) could be of any kind. Mostly cardioids or hyper-cardioids (slightly different kind of directionality).
  • Blumlein - another coincident technique deploying two figure 8 microphones crossed at 90° and with the 45° axis pointing to the center. Pics up a lot of the room sound, not so great for bad sounding rooms. Also needs a very careful placement and two relatively expensive Figure 8 microphones.
  • ORTF - one of the several 'near-coincident techniques combining the advantages of coincident and spaced arrays. ORTF was developed by the French Radio and is a precisely defined arrangement of two cardioids (please do some research if you want to have more details). NOS, developed by the Dutch Radio is another one of that sort. These are easiest to realize arrays with relatively cheap microphones (though the better you can go ...)
.

Hope that clarifies at least the terms being used.

Here some short assessments of the different techniques.

Generally spaced and near-coincident techniques seem to deliver a more spacious and alive sound at the price of the focus and localizability of the instruments. Also the 'spread' of instruments can be tighter than real or more exaggerated (on loudspeakers).
Coincident techniques are felt as narrower, less spacious and less alive, with their advantage of having a rock solid localization and being mono compatible. There are some tricks to make coincident recordings wider and more alive, but that's already a bit of studio magic.

MS is a different animal entirely. If you just record MS to two tracks, you can't really listen to it directly, because the signal first has to undergo a transformation before it becomes a stereo signal. Some modern digital recorders have the trick on board.
The great advantage of MS is that you can change the spaciousness from fully mono to totally exaggerated space (an unpleasant sound anyway).
The mixing of MS recordings should be done on loudspeakers; it's quite difficult to properly judge it through headphones, be they as good as they may. One big disadvantage of MS is that it requires either a dedicated MS-Stereo microphone, which combined with the need for a very quiet microphone for guitar makes it expensive, or, it requires a Figure 8 microphone for the side channel, an they are very expensive if they are any good. But if you have a possibility to work with Figure 8 of decent quality for guitar recordings, that could be very satisfying.

I would say, if I had only to channels of stereo to my disposal and I would have to record a whole ensemble, I'd resort to ORTF with the two best stereo pair of cardioids I could find in my mic locker.

Recording a solo classical guitar is a very special case, because it is only one instrument, but it excites different vibrations in the whole space, so recording it just mono renders a dead and boring sound (IMHO). On the other hand you don't have the issue of representing the distances of different instruments properly in your soundstage between the loudspeakers. That's why the classical stereophonic recording arrays are not the necessarily best answer here. An ORTF pair or an AB pair could render the sound very nicely but with a spaciousness that looses focus and localization and also possibly spreads the guitar over the sound stage a bit, which feels a bit unreal.

That's why my approach would be: using an ORTF kind of an array with a third center microphone mixed to taste, making it a kind of a crossover between ORTF and MS. One also could use an AB approach with narrower spacing. Here is a lot of room for experimentation. Or do MS (see above for the disadvantage: high price).

Several tips here:
  • If you record with two microphones, always pan them dead left and right, otherwise you'll have more or less phase cancellation problems resulting in unpleasant colorations of sound. If you find the array sounds too spacious, bring the microphones more together and/or turn them more inward (if cardioids).
  • For XY it's almost imperative to have a matched cardioid pair
  • put some time in experimenting with the positioning - the results can be different as night and day. Good idea is to find a good spot where you like to sit and play, put on a headphone and let a second person walk very very slowly in tiny increments with the microphones on a stand while you play and listen. You may find a sweet spot!
  • Use the best components in your recording chain you can afford without breaking a bank - nothing is more sad than great playing drowned in the hiss and harshness of bad mics and preamps. The quality standard of even cheap equipment is not bad these days, please don't use the absolutely cheapest if you haven't to
  • If you happen to record in MS technique, it's IMHO not the best idea to commit the matrixed stereo signal to 'tape'. Use the matrix only for listening back, but record the mid and the side channels on separate tracks so that you can make your spaciousness decisions later while listening to it in a relaxed mood.
Please feel free to ask more questions if you need. A lot of this stuff can be found here and there, but it's nice to have it in one spot for reference.

best wishes for your recording
Robert
Music is a big continent with different landscapes and corners. Some of them I do visit frequently, some from time to time and some I know from hearsay only ...

SteveL123
Posts: 423
Joined: Thu Feb 09, 2017 5:05 pm

Re: Nice Comparison of Recording Techniques

Post by SteveL123 » Sun Sep 10, 2017 3:33 pm

Thanks for the detailed reply! :D Will be useful for future reference. I looked up "near field" and most of it was on monitor speaker placements in a recording studio control room during playback, where "near field" is 2 to 4 ft, "far field" is 10 to 12 ft from the listener. What about during recording? Are near and far distances the same for mics (from the guitarist) during recording as for speakers during playback?

User avatar
rojarosguitar
Posts: 4178
Joined: Sat Sep 19, 2009 12:24 pm
Location: near Freiburg, Germany

Re: Nice Comparison of Recording Techniques

Post by rojarosguitar » Sun Sep 10, 2017 8:07 pm

The term 'near-field' is different for microphones and loudspeakers. Different people probably will have a bit different ideas, but for me anything that is significantly closer than 0.5m is near field. Middle field is where you get the most natural sound with a good balanced ratio of direct/room sound. It may be somewhere between 0.7 and 1.5m. Far field is behind that. (that's how I use the terms at least).

The point is that a guitar doesn't radiate its sound equally in all directions. If you come too close you will get quite unbalanced sound depending on the notes you are playing. It sounds quite different from how a guitar sounds for a listener sitting even in the first row of a concert room.

Now, recording is different from listening with our ears because we have all these psycho-acoustic issues a microphone does not have, that's why I would come closer than a good listening distance, but not too close. It depends on many factors, like the quality of the room, the kind of microphones and the quality of the player and his guitar, that is, how well the tone is projected (not necessarily just the volume).
Music is a big continent with different landscapes and corners. Some of them I do visit frequently, some from time to time and some I know from hearsay only ...

Return to “Classical guitar recording and amplification”