You are right in that any reverberation processor will always add to whatever coloring the sound already has gathered during the recording process. No matter what we do, it will always end up as a series of compromises. Sitting inside a nice concert hall, in a good position, is a nice experience - but I wouldn't want a recording from that position. The sensation of sound is completely different when you are immersed in a 3-dimensional soundfield - listening to a dummyhead recording in headphones can be very nice, but playing that recording back on speakers ruins the experience. You always have to consider the final step of reproduction, and take actions to accomodate for it.

I myself prefer recordings which have a reasonable level of support/spot micing. It is for sure a "non-purist" approach, but for me - it simply does the trick. The requirements of certain filmscores today can also require spot mics, since it can emphasize weak instruments which in a purist approach would be very hard to accomodate. Naturally we have to develop our products to be flexible, to cater for the variety of tastes out there.

Experimenting with placing the instruments closer to the microphone and running them more wet sounds like a good idea though. It all comes down to your preference. If you prefer that sound, go for it. If you prefer a more distant placement with direct sound (our virtual spot mic) then go for that.

This is a interesting topic! If we were going to be really purist with spot mics then we'd also have to accomodate for mic bleed. I know engineers hate bleed, but in some cases _to my ears_ it sounds just great. Especially when recording big-bands or other acoustic ensembles. I've recorded big-bands in many different ways and the only way to get it to sound truly authentic, for me, is to have everyone in the same room. I'm convinced it's because of the spatial information that the bleed gives off. It also changes the way I mix because, since I can't get rid of the bleed, I have to factor in the room sound present in the mics. What ends up happening is I "mix the room". Of course it helps if the room also sounds great :-) Not sure if it's worth simulating the bleed in some way. It's a funny thing...in a perfect world there would be no bleed, why the heck would I want to add some ;-) 

MS, yes you are right.  I agree completely and in fact do exactly what you are saying normally, I am just thinking about the background of convolution.  An example of spot mics and basically the recording approach you are talking about that I love is the 1980s London Records FFFR recordings of Mahler.  Like the Solti Mahler 8 - maybe the greatest recording ever made.  These are done with many mics all over the place.

mohurwitzmusic,

what you said reminds me of something Dietz was saying - that the reverb or "bleed" is actually another performer in the room.  Very important idea!

Another thing I was thinking was - since convolution is created by an impulse that is then removed, and the changes made by the room retained, then the "dry signal" in effect replaces that removal.  Though it is somewhat artificial as that dry signal in reality is not really dry the way samples are - it is totally mixed with the hall no matter how close you are - unless you are like two feet from the source.

Hi William,

as it seems, Martin Saleteg has already answered most of your questions. I just want to add a short summary:

All IRs used by Vienna MIR Pro underwent a meticulous post-production process. Based on the laser-aided measurements we conducted in the recorded Venues, all remnants of the direct signal are deliberately removed from the actual impulse response - for several reasons:

- The loudspeaker-emitted direct signal will be of inferior sonic quality than the original dry signal (opposed to microphones, loudspeakers are quite imperfect transducers).

- The loudspeaker-emitted direct signal will _never_ be in phase with the actual dry signal. Mixing these two components can lead to severe phasing issues.

- All input signals are encoded to Ambisonics in MIR Pro. This means that the MIR engine is able replace the original dry signal of any IR not just tonally, but also spatially (in fact with a resolution of a single centimeter).

... all of this means that MIR will increase the sound quality by replacing the recorded direct signal with the dry input signal. In addition, the perceived placement of the signal in the virtual room will be even better than by using the direct signal components themselves (which will always by a bit blurry and diffuse, for several reasons). MIR Pro will also allow for much more artistic freedom when balancing the direct/dry and wet signal components in a way that doesn't suffer from phasing or runtime-delays.

Kind regards,

@William said:

since convolution is created by an impulse that is then removed, and the changes made by the room retained, then the "dry signal" in effect replaces that removal.

So that is correct.  As I thought.  Though MS did not mention it, since what he talks about is normal procedure for using any reverb.   For example, on the old Lexicon reverb I have - still a very good 24 bit algorhythmic unit - you must use dry signal or it is completely artificial sounding, though all of that has nothing to to do with convolution.  Though what Dietz says clears up what I was wondering about in replacing the original loudspeaker-created impulse with a new dry signal.

One other question I have - what happens when a signal in MIR is totally dry?  I haven't yet tried this (mainly because I am too lazy).  Is there still spatial localization?  How is that achieved if so?  Since the replaced impulse's reverb image would no longer be present...

@MS said:

Sitting inside a nice concert hall, in a good position, is a nice experience - but I wouldn't want a recording from that position.

I would - in fact, this is one of my main ideals, derived from the great classical recordings by Capitol, Columbia, Mercury Living Stereo, of the 1950s-60s, which sought to place just one, two or three microphones in the ideal position in the ideal concert venue to record every nuance of sound and still rank among the best recordings of classical music ever done.  Not easy to achieve, but an ideal for recordists to attempt to emulate since it has the great advantage of doing exactly what the human brain does every time one hears any sound.

@Another User said:

I would - in fact, this is one of my main ideals, derived from the great classical recordings by Capitol, Columbia, Mercury Living Stereo, of the 1950s-60s, which sought to place just one, two or three microphones in the ideal position in the ideal concert venue to record every nuance of sound and still rank among the best recordings of classical music ever done. Not easy to achieve, but an ideal for recordists to attempt to emulate since it has the great advantage of doing exactly what the human brain does every time one hears any sound.

For me personally, I can enjoy such recordings through good headphones, or through really good speakers in an acoustically treated room. It starts falling apart when additional acoustics and coloring is added by the listening environment (IE, TV surround systems or boomboxes). I feel that mixes which have a "closer" sound with more support mics translate better in such situations.

@Another User said:

- All input signals are encoded to Ambisonics in MIR Pro. This means that the MIR engine is able replace the original dry signal of any IR not just tonally, but also spatially (in fact with a resolution of a single centimeter).

IOW: With MIR Pro you get a unique and extremely versatile high-precision panning engine for free. ... as a matter of fact we could actually throw out the convolution engine and sell the remaining parts as Vienna Power-Pan Pro. ;-D

@MS said:

Sitting inside a nice concert hall, in a good position, is a nice experience - but I wouldn't want a recording from that position.

I would - in fact, this is one of my main ideals, derived from the great classical recordings by Capitol, Columbia, Mercury Living Stereo, of the 1950s-60s, which sought to place just one, two or three microphones in the ideal position in the ideal concert venue to record every nuance of sound and still rank among the best recordings of classical music ever done.  Not easy to achieve, but an ideal for recordists to attempt to emulate since it has the great advantage of doing exactly what the human brain does every time one hears any sound.

It is _way_ beyond the scope of a little forum message to discuss this in depth (the topic is about as old as the art of recording itself). But I think it is important to point out that a recording can't and won't be a "pure" representation of a real acoustic event any time soon. Even those highly reputable efforts form the early days of stereo are anything but a faithful representation of the actual audience's perception. 😊

There are several well-understood technical and a handful of psycho-acoustic reasons. The main reason for _me_ is that a "pure" recording (assuming that it could exist) is about as interesting as filming a theatrical performance with one camera from one perspective only, without edits and specific lighting. It might have documental values, but there's a reason that the creation of a _movie_ (which uses the possibilities of the medium to the fullest extent) needs much more effort and of course different technical approaches than that. 😉

In other words: Concert music and recorded music have to be understood as the equivalents of theatre and cinema: There might be convergences and overlaps, but they are not the same art form. They adhere to very different artistical, aesthetical and technical premisses.

Kind regards,

That is fascinating info from the experts MS and Dietz - I was trying to goad you guys into something like that and succeeded! 

Also, very helpful in to hear that about the dry signal.  But no matter how it is done MIR is the best sounding reverb I've ever heard.