Stereo microphone technique calculator: ORTF, NOS, Blumlein, A–B, X–Y
Pick a technique, or set the angle between the microphones, the distance between the capsules and the polar pattern yourself, then move the source. The calculator shows how much earlier the sound reaches one microphone and how much louder it is there, and what each difference alone does to the source's place in the stereo image. It is a geometric model with ideal patterns, not a prediction of a room or a real microphone.
Time difference
+0.25 ms
earlier on the right · alone moves the image by: 27 %
Positive: the sound reaches the right microphone first.
Level difference
+5.3 dB
louder on the right · alone moves the image by: 41 %
Positive: the right microphone is louder. The pattern and the distance difference (1/r) both count.
estimate: Model with ideal polar patterns and a point source, no reflections. The percentages come from a 1992 listening test (25 students) as read from Sengpiel's compiled table and interpolated linearly by us; time and level are not added into one position, because the trading rule between them is not checked here.
The ideas behind the controls
- Time, level or both
- Coincident pairs (X–Y, Blumlein) give level differences only; spaced pairs (A–B) give time differences only; ORTF and NOS have both, which is why they are called time-and-level pairs.
- The polar pattern
- Every first-order pattern is a + (1 − a)·cos φ: omni a = 1, cardioid 0.5, hypercardioid 0.25, figure-of-eight zero. The smaller a is, the more directional the microphone.
- Distance to the source
- A close source adds a level difference from distance alone (1/r), even for two omni microphones. Far away only the pattern counts.
Common mistakes
- Taking the percentage of “shift” as a law: it comes from one test with particular music and speech and a particular audience.
- Adding milliseconds to decibels into one number: they are not added here.
- Using the values for a hall: reflections and the real microphone change the image a lot.
What this tool cannot tell you
- Ideal first-order patterns: a real microphone deviates with frequency and angle.
- A single point source in the horizontal plane, no room.
- The ORTF (110°, 170 mm) and NOS (90°, 300 mm) presets are the usual settings; their standard was not consulted.
- M–S is not included: it is not a spaced pair of capsules but a sum-and-difference matrix.
Where the formulas come from
- Sengpiel: phantom source position — Table compiled from a 1992 test (25 students, speech and string quartet): 25 % at 3 dB or 0.23 ms, 50 % at 6.5 dB or 0.48 ms, 75 % at 11 dB or 0.81 ms, 100 % at 18 dB or 1.5 ms. The page was read; the original test was not.
- First-order polar patterns — The definition a + (1 − a)·cos φ is the usual one; the values of a (0.366 and 0.25) put the nulls at 125.3° and 109.5°, checked in the tests.
Frequently asked questions
Why does Blumlein show “—” for level at some angles?
With two figures-of-eight at 90°, when the source is on one microphone's axis the other is at its null (zero sensitivity), so the ratio between them is undefined.
What is “alone moves the image by”?
How far toward the loudspeaker the source would be heard if only that difference existed (0 % centre, 100 % fully in the speaker), from the table of the 1992 test.