Topic: Core Sound TETRA Questions for LEN

[boojum]

Interesting.  Thanks for the posted link.

[Church-Audio]

I will be selling 4 matched capsules and the tube necessary to build these mics if anyone wants to build there own I can supply it with a 4 ch battery box and 4 matching capsules  for $300 cabling will be extra if needed. The tube will be precut lengths that you can jerryrig into a array. I dont want to get into building the arrays but I will supply the parts to do so. You will need software to convert the program material.

[hugh]

Hi folks.  To follow up on Gordon's link with a bit more detail of my chain and that particular recording:

This was recorded at Zoe's, a tapas bar in Beverly Farms, MA, which has really nice acoustics -- their ceiling is curved and finished with acoustic treatment.  The band is bass, drums, guitar (small amp on the floor), saxophone (not mic'd), and vocal.  The mic was suspended upside-down from a boom, quite close to the band.  Recording chain is: TetraMic, to a Daking Mic-Pre IV (my favorite toy!), then FireFace 800 for A/D conversion, then a laptop running AudioMulch (48kHz, with a four-channel file writer at 24 bits, and the RN Digital Inspector VST spectrum analyzer and peak meter).  The original recording is A-format, which I then adjust to pre-measured gains of my mic pre's four channels, then converted to B-format using the VVMic software supplied with TetraMic, then mixed down to stereo.  (The A-format calibration and B-to-stereo downmix software are my own: http://cabezal.com/~inguz/stuff/AFormat.zip).

I usually mix things down to crossed-hypercardioids for stereo.  But this set, the mic was really too close to the band: sax way over the left side, and vocal PA almost 180 degrees the opposite side; crossed-mic mixes sounded really weird and unbalanced, and phasey when the sax player moved around.  So I played around for a long while trying to get the right mix: rotating the microphones, trying different directivities from cardioid through figure-eight, even tilting up, down and sideways.  And finally settled on the UHJ mix, which keeps a really nice balance of reverb and ambience, but brings the sides in toward the front.  Downside of UHJ is that it usually sounds quite "distant", even for classical orchestra.

So that's one amazing thing about recording with TetraMic.  You really can throw up a mic just about anywhere that sounds good -- even upside-down, as here, or pointed straight at the door, or whatever -- and play around with complete freedom in rotating and mixing to stereo later.  And I think the results are good: incredibly stable and solid stereo image.  (There is a little "wandering" on the Zoe's tapes, but that's actually due to using Audacity's default compressor instead of SC4.  But these tracks really aren't much compressed at all).

Hugh

[F0CKER]

^ one of the reasons I keep fighting through my Soundfield issues.

The flexibility is awesome. 

[illconditioned]

I will be selling 4 matched capsules and the tube necessary to build these mics if anyone wants to build there own I can supply it with a 4 ch battery box and 4 matching capsules  for $300 cabling will be extra if needed. The tube will be precut lengths that you can jerryrig into a array. I dont want to get into building the arrays but I will supply the parts to do so. You will need software to convert the program material.

I'm interested in hearing more about this, Chris.

Can you post a picture of the "tube" that you refer to?

  Richard

[Church-Audio]

I will be selling 4 matched capsules and the tube necessary to build these mics if anyone wants to build there own I can supply it with a 4 ch battery box and 4 matching capsules  for $300 cabling will be extra if needed. The tube will be precut lengths that you can jerryrig into a array. I dont want to get into building the arrays but I will supply the parts to do so. You will need software to convert the program material.

I'm interested in hearing more about this, Chris.

Can you post a picture of the "tube" that you refer to?

  Richard

Its just 4 pcs of brass tube that can be joined or welded into a tetrahedral mount. Some machining would be required.

Chris

[live2496]

The original recording is A-format, which I then adjust to pre-measured gains of my mic pre's four channels, then converted to B-format using the VVMic software supplied with TetraMic, then mixed down to stereo.  (The A-format calibration and B-to-stereo downmix software are my own: http://cabezal.com/~inguz/stuff/AFormat.zip).

Hugh,
Thanks for providing the details. Can you tell us more about how you go about setting your levels?

Gordon

[hugh]

I generally look at the RNDigital plugin during a soundcheck -- it has nice resettable peak displays -- and aim to be around -9dBFS at peaks.  Real performance always peaks at least 3dB louder than soundcheck in my experience!  Occasionally of course I miscalculate and need to pull the gains down during a track break.  With 24-bit recording I feel comfortable recording some distance away from peak; the noise floor is pretty low.

The pre I use doesn't have ganged gain across the channels (an advantage of units like the MOTU Traveler) but it does have stepped gain controls instead of potentiometers, and I've measured them fairly well: channels are matched within about 0.3dB (and I calibrate afterward to adjust for that), but channel gain matching doesn't vary significantly at all with the actual gain setting.

[SparkE!]

I will be selling 4 matched capsules and the tube necessary to build these mics if anyone wants to build there own I can supply it with a 4 ch battery box and 4 matching capsules  for $300 cabling will be extra if needed. The tube will be precut lengths that you can jerryrig into a array. I dont want to get into building the arrays but I will supply the parts to do so. You will need software to convert the program material.

I'm interested in hearing more about this, Chris.

Can you post a picture of the "tube" that you refer to?

  Richard

Its just 4 pcs of brass tube that can be joined or welded into a tetrahedral mount. Some machining would be required.

Chris

The angle between any two adjacent mics would be about 109 degrees.  So, to make the tubes meet correctly, you'd make 3 marks, equally spaced circumferentially around each tube and use those marks to help you align 3 cuts, each approximately 109/2=54.5 degrees to the axis of the tube.  It's probably easier to avoid the machining of the tubes and simply drill 4 holes of the correct diameter into the center of each face of a tetrahedral block and then just glue the tubes into the holes.  Of course, then you have to machine a tetrahedral block.  Adjacent faces of the block are about 71 degrees from each other.  On second thought... This doesn't look too easy to me.

[Church-Audio]

I will be selling 4 matched capsules and the tube necessary to build these mics if anyone wants to build there own I can supply it with a 4 ch battery box and 4 matching capsules  for $300 cabling will be extra if needed. The tube will be precut lengths that you can jerryrig into a array. I dont want to get into building the arrays but I will supply the parts to do so. You will need software to convert the program material.

I'm interested in hearing more about this, Chris.

Can you post a picture of the "tube" that you refer to?

  Richard

Its just 4 pcs of brass tube that can be joined or welded into a tetrahedral mount. Some machining would be required.

Chris

The angle between any two adjacent mics would be about 109 degrees.  So, to make the tubes meet correctly, you'd make 3 marks, equally spaced circumferentially around each tube and use those marks to help you align 3 cuts, each approximately 109/2=54.5 degrees to the axis of the tube.  It's probably easier to avoid the machining of the tubes and simply drill 4 holes of the correct diameter into the center of each face of a tetrahedral block and then just glue the tubes into the holes.  Of course, then you have to machine a tetrahedral block.  Adjacent faces of the block are about 71 degrees from each other.  On second thought... This doesn't look too easy to me.

I dont remember saying it was easy but it can be done.

[Gutbucket]

One simple method that comes to mind if you have access to a standing belt-sander machine:
Make a simple jig to hold the brass tube square to the sanding surface and mark the rotation of the tube. Set the sander table angle to make a 54.5deg with the belt.  Sand one angle, rotate tube 120deg, sand second angle, rotate tube 120deg sand third angle.  Repeat with the other three tubes.

Might be easier to epoxy the angled tube ends together.  It's mighty tricky to hold those pieces in place while the solder or braze flows.  If you do one and then try to add a second, the first joint will melt - a big part of what makes Albert's tube soldering work so impressive.  You really need to hold them all in place and do them all at once.

[live2496]

I generally look at the RNDigital plugin during a soundcheck -- it has nice resettable peak displays -- and aim to be around -9dBFS at peaks.  Real performance always peaks at least 3dB louder than soundcheck in my experience!  Occasionally of course I miscalculate and need to pull the gains down during a track break.  With 24-bit recording I feel comfortable recording some distance away from peak; the noise floor is pretty low.

The pre I use doesn't have ganged gain across the channels (an advantage of units like the MOTU Traveler) but it does have stepped gain controls instead of potentiometers, and I've measured them fairly well: channels are matched within about 0.3dB (and I calibrate afterward to adjust for that), but channel gain matching doesn't vary significantly at all with the actual gain setting.

Taking a look at the console programs...
So if I had 4 unmatched levels, I could record a test tone and the software will adjust the gain?
Could you describe the recording of the test tones? Do you record at 4 compass points, or what steps are involved in this process?
And is this 4 mono wav files or 1 interleaved file?

C:\Hugh Pyle>aformat
A-Format file analysis and normalization.

Assuming several channels of audio, each with a calibration
tone of several seconds near the start or end of the track.
Scans the file for 1-second regions where the signal is
greater than -40dB and very constant (lowest standard
deviation of the Hilbert envelope).  Typical deviations
with well-recorded calibration tones will be less than 0.05dB.

Reports the relative gains of these calibration signals and
optionally normalizes the file(s) by amplifying each channel
to the gain of the loudest calibration signal (producing 32f WAV).

Usage:
AFormat <infile> [/output <outfile>] [/gain n] [/cal n] [/test n] [/testmin n]

  <infile> (input filename, or * for all WAV files in the current folder)

  Optional:
  /output <outfile> (output filename)
    (use '-' for stdout)
    (if not specified, only gain coefficients are calculated)
    (if input is '*', this is the 'base' output filename)
  /gain <n> (dB; defaults to 0)
  /cal <n> (seconds to scan for calibration signals; defaults to 60)
    (negative values seek from the end of the file)
    (or use a filename if calibration signals are separate)
  /test <n> (seconds of cal signal to look for, defaults to 1)
  /threshold <n> (dB, defaults to -40)

[hugh]

The test tones are to take the same input, through each channel of the pre/convertor, to measure those channels in turn.  For my pre, I did this acoustically: with a single microphone near a loudspeaker playing a sine tone, plugging the mic into each channel one at a time while recording the outputs.  (Then for convenience I aligned the channels in Audacity, since the whitespace isn't useful).  The result for my pre is here: http://halfbricking.com/Samples/daking.wav

Then: AFormat /input <aformat_recording.wav> /output <aformat_calibrated.wav> /cal daking.wav

[hugh]

Hey, another stereo sample: http://halfbricking.com/Samples/sample_WhenWeCome_1644.wav

(Full surround version here: http://www.ambisonia.com/Members/hughpyle/ambisonicfile.2008-10-20.9949056803/)

[live2496]

The test tones are to take the same input, through each channel of the pre/convertor, to measure those channels in turn.  For my pre, I did this acoustically: with a single microphone near a loudspeaker playing a sine tone, plugging the mic into each channel one at a time while recording the outputs.  (Then for convenience I aligned the channels in Audacity, since the whitespace isn't useful).  The result for my pre is here: http://halfbricking.com/Samples/daking.wav

Then: AFormat /input <aformat_recording.wav> /output <aformat_calibrated.wav> /cal daking.wav

I see. Then you are recording the same source with the gain/trim in the same relative position. And the software can tell you how much to adjust up or down to get them all the same. (This adjusts for the small differences between inputs).