Mixing real time acoustic recordings with MIDI sequencer ones -how to achieve ambient unity

[Kristofer Emerig]

I posted this to Simon:

On your encouragement, I'm working on a Harsichord concerto. Here's the problem; I don't have a real string ensemble. I have some SF2 files for the (2) violins, (2) violas, cello and double bass, and an application called "Music Write Proffessional 2000", which I usually just use to get a legible score - yes I still write on staff paper first; It makes me feel comfortable. The SoundFonts I have sound alright, but they have a pretty spacious feel, even without additional reverb. How can I incorporate a real harpsichord, played in a 12 x 12 foot room, with that spacious auditorium environment of the string section in a seamless and credible way? Suggestions? I've got my material together and I'm ready to do this, and eager to get some feed back after I upload it.

He suggested placing this question here for feedback.

[Patrik Ohlsson]

I've experimented with this a little in Cubase.
When using recorded audio with vst I try only to use closemic'ed samples since they have little or no room ambience included.
I try to record the instrument in a similar way, with little or no room ambience. Then I set up a FX track (or two) with an Altiverb instance. And then I play the VST's and recordings through the FX track and this usually creates a seemless result.

[Kristofer Emerig]

I get what you're saying Patrick, but the fundamental problem is that I'm starting with strings with a good amount of "reverb" built into the Soundfont itself. You're right, it'd be better to start with no environmental effects as a starting point, but my technical knowledge and resources are minimal. I've heard Jimi Hendrix recorded some of the drum tracks in a bathroon- that's the only way he could achieve the sound he wanted. I think he had a bigger bathroom than me though. Probably real marble fixtures for optimum response.

[Patrik Ohlsson]

A bit of work dragging a harpsichord into the bathroom:)
So you can't edit and remove the reverb on the soundfont file? Haven't worked much with soundfonts..
You could try to imitate the reverb of the soundfont and record a natural audiotrack with as little ambience as possible and then use a reverb plugin to try to imitate the soundfont. Don't know of any other solutions to this problem.
Sometimes it could sound nice with a closesampled solo instrument and background instruments with a bit of ambience. However if the harpsichord is only meant to be a part of the orchestra then it's a bit harder.

[Kristofer Emerig]

Yeah Patrick, you hit the proverbial nail on the head. The problem is that as the soloist, I want the harpsichord to be somewhat centered and up front sounding, and yet the harpsichord must also blend in convincingly enough to function as continuo. I've noticed that even on professional quality, big label productions, achieving that balance is difficult, and many (harpsichord & ensemble) recordings simply fall short. I think Pinnock and the English Concert recordings of JSB's Harpsichord concerti on the Archiv label are about the best I've heard, but that's a fully integrated acoustic performance recorded in real time.

[Patrik Ohlsson]

Okay.
Personally I use the EWQL harpsichord, but it's hard to blend it with the rest of the orchestra, although all instruments are sampled in the same hall and not at close range. Vienna instruments are probably one of the bests at sampling classical instruments. They probably have a more crisp sounding harpsichord than EWQL. I'll check out the English Concert recordings.

[steve urwin]

It might help to use some dry midi samples alongside the wet SF2 's, and the level of the wets right down..

[Doug Lauber]

The harpsi must be mic'd close and surrounded by suspended blankets, to capture a dry sound and eliminate the room reverb. Next you must try to duplicate the wet acoustics of the other instruments, without going too far. When it seems about right, I'd back off on the amount of simulated room ambience/reverb. It doesn't necessarily need to be perfectly matched. The brain can compensate.

[Kristofer Emerig]

I've never tried the blankets. I have tried nearly every other conceivable recording configuration on my harpsichords and clavichords; from all distances and positional relationships, ie, mic in the case, behind the spine against a wall, even under the case itself. I've also experimented with damping choirs of string by different methods. On my Italian virginal, I get the best lute stop effect suprisingly with a pair of sweat socks. They seemingly have just about the right mass and surface area. Different damping techniques can yield astonishing effects on overtone propogation, approaching synth-like qualities under the right recording conditions. I'll try the blanket technique though. I can achieve good and widely varying sound from all of my instruments. Where I always fall short is in the technical ability to capture those sounds in the recording process. I'm not a techie, & have no toys to speak of, nor a knowledge of their utilization if I had them. I recording directly into Vienna Wave Studio with a Soundblaster 5.1 card, I think. All of the Soundfonts I have are freeware downloads. Yodel is recorded with all preinstalled patches on the Soundblaster card, for instance.