• Register
  • Help
Results 1 to 10 of 10

Topic: Sampling a piano: moving the mics?

Share/Bookmark
  1. #1

    Sampling a piano: moving the mics?

    A quick question:

    1. When sampling a piano, have people experimented with moving the mics to just above each string being sampled (and then panning the notes later for an accurate reproduction of each note's location)? My thought is that this would improve the recording of ppp and pp layers, since less noise would enter.

    Getting ready to sample a piano, as you might have guessed, and would appreciate the benefit of other peoples' experiences.

  2. #2

    Re: Sampling a piano: moving the mics?

    That works to a certain degree. I sampled a piano once using this method (Stockhausen does his piano recordings like this).

    It only works in mono, which could be fine. But if you record in stereo and move the mics around all the time and later re-adjust the position of each note in the keyboard stereo image you will find there are all sorts of weird stereo imaging problems caused by the continous moving stereo ambience, no matter how soft that would be.

    If I may make a suggestion; You could experiment with a recording where the mics are making fysical contact with the soundboard or the bridge. This way you get no background noise or ambience. You could use reverb plug-ins upon playback or -better- a large speaker mounted on a big surface upon playback to correct the recording position.
    Best regards,
    Michiel Post


  3. #3

    Re: Sampling a piano: moving the mics?

    Michiel, expliquez s'il vouz plait.

    Are you talking about using a piezo contact mic?

    And then you play back the recording through the speaker mounted on the large surface and re-record it?

    Is that how you recorded your harpsichord, by the way?

  4. #4

    Re: Sampling a piano: moving the mics?

    Quote Originally Posted by Nick Batzdorf
    Michiel, expliquez s'il vouz plait.

    Are you talking about using a piezo contact mic?

    And then you play back the recording through the speaker mounted on the large surface and re-record it?

    Is that how you recorded your harpsichord, by the way?
    Well all right then. The idea to record a piano without any room noise or distance isn’t new and it’s a very appealing idea to have a `pure` recording of the singing of the strings and not the acoustics of the recording space. One way would be to record the piano in an anechoic room (like the Kawai grand was recorded for the GRAND) which leaves out all elements of the room. This results in a clean recording but still has a considerable distance from the source.

    Traditional miking techniques would always suggest to mike a piano from at least a meter and better 2 or more for a natural sound. This is because the resonating of the strings has to mingle with a space (the air and all the reflections the room) in order to get the right tonal balance. Recording from such a distance is a tour de force for microphones. Maybe in 10 years the pick-up elements of microphones have gone through a complete metamorphosis but at this moment it is a lost fight to pick up the sound of strings (plus body resonance plus room reflections) from 2 meters. The microphones noise floor is playing parts here. And so is the microphones sensitivity to very small amplitude changes and frequency bands. Most likely most microphones can still pick up frequencies above 10 k from such a distance but the sensitivity is getting lower and lower when you move up the spectrum. By the time you need to capture the frequencies outside our hearing (ABOVE 16 K) microphones usually leave a very bad taste in your mouth (or ears), while these frequencies are very important (they contain the higher overtones series for each note). To make things worse the sum frequencies of the basic note fundamental frequency and their full overtone scale are not captured at all.

    This is why an acoustic source sounds totally different than the well recorded amplification of that sound.

    Now if you go back to the source and start at the source (the singing strings) and start your sound reproduction process from there you can overcome some of the major obstacles.

    Starting at the source means going as close as you can to the vibration. This would be at the bridge or on the sound board. The signal is so strong there that you will need special pressure gradient microphones that resist the force of the vibration. Luckily these exist and are relatively cheap. A pressure gradient microphone (or pick up) can capture frequencies up to 40 kHz: in combination with a 96 kHz audio converter you can captured more of the piano sound than with ultra professional Omni’s from a few feet distance. You can adjust the position of the pick up element during the recording to stay close to the resonating strings.

    Next step is to configure how to treat the sound from there. Obviously we are not used to hearing a piano from the position of the bridge and we need to compensate for this weird recording position. So what to do with the signal once you have it recorded? This is where art comes into play. You can do the most obvious. Mount the reproducing speaker to a sound board, if possible to a piano sound board. This will come very close to the real sound of a grand piano. It’s not that simple to set up this kind of installation just to hear a well recorded piano sample play back but there are other ways. If you can describe the behaviour of the sound board and "copy and paste" it to the recorded signals you may just play the samples back over a computer-based effect processor that "pastes" the soundboard-to-room transition behaviour to the sample. This is possible with convolution. You will need a sweep tone, a speaker that is small enough to be positioned very close to the soundboard and a set of microphones to pick up the sound of the amplified sweep from a typical recording distance. If you convolve the sweep with the captured material you get an ir that (in scientific terms) has all the information that was generated by the transition from the soundboard to the room.

    Simply load the IR in a convolution machine and you can play back your piano samples trough the IR and play the result over normal speakers.



    I worked with a scientist who did the above experiment for one velocity layer of a small piano and it worked very well. The brilliance and presence of the piano was very convincing and the typical room response was also captured. Noise was totally missing and the dynamic decay of the tone was very close to the original. This convoluted piano was recorded in mono (playback was pan-potted like a real piano) and you had more or less the idea of hearing a real piano when you played those samples. I say more or less because this technique was limited to one layer so playing this wasn’t realistic at all.



    Best regards,
    Michiel Post


  5. #5

    Re: Sampling a piano: moving the mics?

    Thanks for the detailed reply. I may have misphrased my question--when I mentioned placing the mic "directly over the strings," I didn't necessarily mean extremely close to the strings. I only meant above each string, instead of just leaving two mics in one position for the entire sampling project.

    I wonder, now, though, if one solution to the problem with soft layers might be to move the mic each time one samples, and push the mic closer to the string for the softest notes. (And pull the mic back gradually for louder layers.) Would this work? Or is this what's already being done with the White Grand and others that have a lot of layers?

    Thanks for any insights.

  6. #6

    Re: Sampling a piano: moving the mics?

    Quote Originally Posted by Jake Johnson
    I wonder, now, though, if one solution to the problem with soft layers might be to move the mic each time one samples, and push the mic closer to the string for the softest notes. (And pull the mic back gradually for louder layers.) Would this work? Or is this what's already being done with the White Grand and others that have a lot of layers?

    .
    One the White Grand (and the upcoming Black Grand) the microphones stayed in the same position all the time.
    In my opinion, moving the mikes around would be a catastrophe unless your after a special effect.
    Worra
    SampleTekk

    Arf, arf, arf...

  7. #7

    Re: Sampling a piano: moving the mics?

    This technique sounds interesting, Michiel! Now that convolution will be a standard feature in samplers of the near future, are you planning to use such a technique for new samples?

    It seems to me, such samples will be better for live performance than the more ambient samples of today.

    Already two jazz pianists in Amsterdam hearing me with my laptop with rhodes samples showed interests in working with a laptop. I'm sure they would also be very interested in good piano samples for live gigs.

  8. #8

    Re: Sampling a piano: moving the mics?

    Worra:

    My thought was just that moving the mics to each note and changing the distance from the string might reduce the amount of ambient noise. Why does it sound like such a bad idea to you?

    Thanks.

  9. #9

    Re: Sampling a piano: moving the mics?

    Quote Originally Posted by Jake Johnson
    Worra:

    My thought was just that moving the mics to each note and changing the distance from the string might reduce the amount of ambient noise. Why does it sound like such a bad idea to you?

    Thanks.
    By moving mikes you are asking for phasing problems. I also have a gut feeling that the finished samples would sound a bit strange as a unit, but then again, I might be wrong!
    Nothing wrong in trying out new stuff. You could always find something good.
    I read about when Beatles recorded "Good Morning" from "Sgt Pepper". The brass are miked in a un orthodox way, they simply put the mikes right into the bell of the instruments. Not your usual standard way, but it sounds great!
    Worra
    SampleTekk

    Arf, arf, arf...

  10. #10

    Re: Sampling a piano: moving the mics?

    It will be another month before I'll have time to sample this piano. I'll post a few samples once I do.

Go Back to forum

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •