Topic: Using cardiod microphones to get the best stereo imaging

As I continue to try to make the most realistic simulated player sound for my upright piano equipped with a MIDI strip, I went from binaural back to the microphones, making sure that there is only one microphone per output channel whith no blending, chich seems to hurt the accuracy of stereo imaging (though perhaps a little bit may help the spread of sound coming off of the soundboard in a real piano, but that hurts the precision of the sound that you hear with your ears imaged off the string that is struck). This time, I placed the microphones at roughly the player position, close together simulating two ears, and just about 90° off, using cardioid microphones so that they are more directional. I seem to like the U87 best, for its clarity, although it actually sounds more precise in some ways than my piano does (as an acoustic piano).   The CMC6mk4's are pretty nice as well.  Curiously, when I had the microphones centered the way that a player would be facing the keyboard, the vast majority of the sound appeared to be my right. This is not the stereo imaging that I wanted, especially when playing middle C and notes surrounding C also sounded as if they were coming from the right side of the keyboard.  (and I do not have a control on my system that creates left-right balance, either correctly or, in this case, incorrectly.)

I first tried to reduce this by reducing the keyboard width. My speakers are outboard of the piano, so I've traditionally run a keyboard width equaling 0.40 for most of the presets that I have used and downloaded. In this case, I reduce the width to 0.10 which seem to help, but still had keys sounding as if they were too far to the right.

Next, I shifted the two microphones about one third to the right, which helped make the sound appeared to be centered on my piano. Then, I thought, maybe it would be more realistic if I just took the microphones and rotated them to the left while leaving them centered. This is what the uploaded preset displays: two U87 – cardioid microphones, placed 90° off, with the right one facing almost directly into the piano, and left one about 90° left of that. It gives me reasonably accurate stereo imaging, but it does reduce my bass response very slightly (which on my piano is not a big deal, as I have a little more bass than I need reflecting in the "box" beneath my keyboard, the floor, and the speakers each to the left and to the right, as well as my subwoofer).

http://www.forum-pianoteq.com/uploads.p...81%29.mfxp

The differences that I find between this and the binaural are the following: 1 – The stereo imaging with this microphone set up is not quite as precise as that of the binaural; 2 – The binaural is a bit more dull – There seems to be less sparkle with the binaural headphone microphone, which also leaves the sound to "finish" too quickly, almost as if somebody's hand is on the piano harp keeping the metal harp from ringing for several seconds after the note is done.

How is the binaural headphone microphone created? Is it a formula based on some of the existing cardioid microphones? Is there a way to change its frequency response and make it a bit livelier?  Has anyone had any better success in stereo imaging for a more precise Player setup?

- David

Re: Using cardiod microphones to get the best stereo imaging

Interesting. Could you post an mp3 or two of your results? (I saw that you posted the fxp, but you may be wanting a specific sound for certain types of music, so it can be hard to judge the intended results from an fxp alone.)

Re: Using cardiod microphones to get the best stereo imaging

Yes, it's too bad that I can't post an MP3 that was recorded from inside my ears!

I assume that you mean for me to make an MP3 via pianoteq with the different presets, rather than recorded with an acoustic microphone. I will do what I can when I get a chance in the next day or two, including showing the presets with the microphone shifted to one side versus centered, Etc. Thank you!

Last edited by dklein (17-11-2017 02:01)
- David

Re: Using cardiod microphones to get the best stereo imaging

OK, Jake, here's what I worked out:

I first made up a MIDI file using the same phrase at different points in the keyboard. I chose a preset that I had created for Steinway Model B with some modifications, such as a slightly worn condition and in microphone width of 0.10 (rather than my usual 0.40) - on my system [ with the speakers to the sides of the piano, and the reverb as the knock.wav reverb file, adjusted some, to sound like the case-and-harp ring to the piano that goes on for a pretty long time after hammerstrike ], this preset sounds quite a bit like the Native Instruments "Alicia's Keys" VST.  I then created MP3s using this through four separate microphone positions: a binaural headphone microphone, a centered U87 cardioid microphone pair, a slightly shifted-to-the-right U87 cardioid microphone pair, and a centered-but-rotated-90° to the left U87 cardioid microphone pair.

To my ears, the microphone pair that is shifted to the right has the best stereo imaging when I am sitting at the keyboard and actually playing. The one that is rotated to the left isn't too bad either, but has more of an alteration of the bass notes. But, on playing these back to myself via the MP3s, as I created them to share with you, I realize that some of the stereo imaging is not as good when I am just sitting here listening, as compared to when I am sitting here and essentially looking at the location of my fingers when playing notes on the keyboard – that action of playing seems to help my brain perceive more accurate stereo imaging!

So far, no great logic that I have learned. I know that the microphone pair, rather than the binaural headset, has more of an open sound to it, and, for some reason I don't understand, shifting the paired microphones to the right on the keyboard seems to improve the accuracy of its stereo imaging, which really doesn't make much sense to me.

Please tell me what you think. The files are noted below:

MP3s:
Binaural Headphone:
http://www.forum-pianoteq.com/uploads.p...281%29.mp3
U87 Cardioid pair, 90 degrees off, centered:
http://www.forum-pianoteq.com/uploads.p...20Mics.mp3
U87 Cardioid pair, 90 degrees off, shifted to the right:
http://www.forum-pianoteq.com/uploads.p...20Mics.mp3
U87 Cardioid pair, 90 degrees off, centered but rotated to the left:
http://www.forum-pianoteq.com/uploads.p...20Mics.mp3

FXP (with binaural earphones, for example):
http://www.forum-pianoteq.com/file/4n1x60u

Test Wandering MIDI file:
http://www.forum-pianoteq.com/uploads.p...20MIDI.mid

Microphone preset files:
http://www.forum-pianoteq.com/uploads.p...eight.mfxp
http://www.forum-pianoteq.com/uploads.p...tered.mfxp
http://www.forum-pianoteq.com/uploads.p...shift.mfxp
http://www.forum-pianoteq.com/uploads.p...tated.mfxp

Last edited by dklein (20-11-2017 05:10)
- David

Re: Using cardiod microphones to get the best stereo imaging

Yikes. It will take me a few days to absorb all of this. Thanks for posting these.

Re: Using cardiod microphones to get the best stereo imaging

I think you have to take into account the lid of Model B that, when fully open, shifts the sound toward the right by its reflections. With your centered mfxp, some of the sound to the left is lost, which is not the case if you use the centered mfxp with an upright like U4. Closing the lid corrects this impression a bit but also makes the sound less lively. So, to get a better player impression your right-shifted mfxp is correct for using a grand model while playing on an upright in my opinion.

Re: Using cardiod microphones to get the best stereo imaging

Thanks, Gilles.

I have to say that I forget that even when the microphones are placed out from under the lid and closer to the player, that the overall sound still gets directed more to the right when the lid is up. (Perhaps this is also part of my bias, living at home with my upright Steinway piano instead of a grand.)

Additionally, I wish that I could modify the binaural head/speaker to make it a bit more airy overall. Otherwise, I will likely keep fighting with the other microphones and their orientations to produce the most realistic stereo effect.

- David

Re: Using cardiod microphones to get the best stereo imaging

...and I just realized that when I tried the rotated microphones to correct the right-shifted. Imaging problem, I rotated the microphones to the left, which seemed to make things better, although I now realize that that was backwards. This morning I have uploaded a more correct set of "rotated" files, with the microphones rotated to the right (which effectively "centers" the right-shifted stereo image onto the soundfield. I have uploaded both the microphone preset as well as the MP3:

http://www.forum-pianoteq.com/uploads.p...20Mics.mp3

http://www.forum-pianoteq.com/uploads.p...right.mfxp

Once again, the stereo imaging is more accurate to my ears when I am looking at the keyboard and playing the notes that it is while looking straight ahead and listening as I sit at the keyboard. Additionally, this rotation to the right, essentially the opposite rotation of the prior one that was to the left, places the image of the piano "deeper" into the piano than the one that was rotated to the left.

For that reason, I then moved the microphone closer to the keyboard, but still right-rotated and centered:

http://www.forum-pianoteq.com/uploads.p...20Mics.mp3

http://www.forum-pianoteq.com/uploads.p...loser.mfxp

Last edited by dklein (21-11-2017 10:07)
- David