Become a Site Supporter and Never see Ads again!

Author Topic: The World's Finest Blumlein Array  (Read 11879 times)

0 Members and 2 Guests are viewing this topic.

Offline voltronic

  • Trade Count: (40)
  • Needs to get out more...
  • *****
  • Posts: 4104
Re: The World's Finest Blumlein Array
« Reply #15 on: July 28, 2020, 03:24:29 PM »
Looking forward to listening to all of these.

Volt- I've been long curious about Harpex's virtual patterns and stereo output derivations but have never gotten around to demoing it using any of my own recordings.  Especially curious about it's synthesized near-spaced and A-B configurations, which I presume use the plane wave decomposition function Len mentions, yet I remain skeptical about how well real-world phase/spacing interactions can be simulated using sample data from only a single point in space.

The bold part is what has me baffled.  I don't understand how you can take a coincident mic array where there are effectively no time-arrival differences between mics, and make it sound like it was actually in two physical spaces without altering the sound significantly due to new phase interactions between the individual tracks.  If it can be done convincingly, I will happily eat my words.

For now, it is pretty incredible what can be done with post-processing of ambisonic and other coincident arrays.
I am hitting my head against the walls, but the walls are giving way.
- Gustav Mahler

Acoustic Recording Techniques
Team Classical
Team Line Audio
Team DPA

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (16)
  • Needs to get out more...
  • *****
  • Posts: 15700
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: The World's Finest Blumlein Array
« Reply #16 on: July 31, 2020, 10:57:51 AM »
Len, thanks again for the links to the various virtual microphone renderings of the Angelica Women's Chamber Choir recording.  They are very helpful.

It's very interesting to hear those including a center microphone at different levels, as a stronger center emphasis is something I frequently find myself desiring in Blumlein recordings (some, not all).  Generally I've achieved this myself by setting up for a recording using a somewhat more narrow inclusive angle than 90 degrees, or by making a Mid/Side ratio adjustment afterward, both of which increase center level at the loss of some overall width.  The use of second order patterns may allow one to set things up so at to retain a bit more width.  Note to those listening: I find the biggest difference to be the difference in reverberant balance between the Blumlein and all three variants using the 2nd order hypercardioids, which is to be expected as all three hypercardioid variants share less collective sensitivity to the rear.  The difference in level through the center I find to be the second most noticeable difference, and the difference in width impression somewhat more subtle.

If I were searching for the most optimal stereo output for that recording, I'd probably first try Blumlein along with some center-pointed 2nd order hypercardioid added to it, and compare that against the three 2nd order hypers with the inclusion of a rear-facing pair.. both in search of an optimal reverberant balance without overly compromising the other aspects.

The reason I'm posting about this is that I'm realizing it reflects what I've come to recognize as the core aspect of my own approach to live music taping, which is gaining a useful degree of increased control over the directional sensitivity of the recording array in an overall sense.  Not simply at the level of single microphones and stereo-pairs but across the entire horizontal plane.  Blumlein is fantastic when the situation allows for it to work well, but as tapers we are rarely presented with the perfect situation for using it.  I find it is far more useful to have a useful ability to adjust and balance the contribution from each direction individually, collectively producing an overall balance which is otherwise unachievable.

This ability to optimally balance the contribution from all directions is incredibly valuable in the odd world of concert taping where the freedoms of the recordist are so greatly constrained, and it is this aspect in particular I want to emphasize to other tapers as being of special interest with regards to OctoMic.. even more than its especially well behaved Blumlein capability.  It is what I see as the biggest difference between the situations in which tapers find themselves in comparison with most every other form of music recording.

I currently achieve such functionality by either using four baffled omnis facing in the cardinal directions, or five supercardioids (center one a M/S pair) in combination with two wide omnis.  A good ambisonic coincident mic offers similar horizontal plane directional balance flexibility, and a higher order one has potential for increased performance that may serve to offset the absence of spacing between elements.  Although not as daunting as once was, running eight channels is still a lot for tapers, yet a number of tapers do that now.  I do with far greater setup complexity than is required for OctoMic.

This post is an strong endorsement of the value of multi-directional pickup in general and OctoMic represents a compact, simple to setup, single unit way of achieving that. I credit TetraMic with being one of the especially useful tools which helped me develop a deep, intuitive understanding of coincident mic'ing, and helped further the deep recognition of what I now think of as a fundamentally important aspect of control over differentiated pickup across all directions.

Len, big thanks for developing these tools and making them available to the audio community.
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline Len Moskowitz (Core Sound)

  • Site Supporter
  • Trade Count: (3)
  • Taperssection Member
  • *
  • Posts: 381
    • Core Sound
Re: The World's Finest Blumlein Array
« Reply #17 on: July 31, 2020, 01:14:53 PM »
You're very welcome.

> If I were searching for the most optimal stereo output for that recording, I'd probably first try Blumlein along with some center-pointed 2nd order hypercardioid added to it, and compare that against the three 2nd order hypers with the inclusion of a rear-facing pair.. both in search of an optimal reverberant balance without overly compromising the other aspects.

With OctoMic and the SPARTA beamformer plug-in, that's ridiculously easy. And all done in post.

Len Moskowitz
Core Sound
www.core-sound.com

Offline kuba e

  • Trade Count: (1)
  • Taperssection Member
  • ***
  • Posts: 492
  • Gender: Male
Re: The World's Finest Blumlein Array
« Reply #18 on: August 02, 2020, 02:34:16 PM »
Thank you all for very interesting posts. I know very little about ambisonic recording. And I would like to ask a few questions.

Please, is it possible to decompose a directional pattern of the 1st order to an omni and a fig. eight component? For example:
1st order Cardiod =  0.5 x (1st order omni + 1st order fig. eight)?

Directional patterns of 2nd order look complicated. Is there an easy way to decompose it to the components?

I was playing a little bit with Len's recording and Sparta plugin. There is an audible difference between the 1st and 2nd order. I also tested the 2nd order beams at +-45. When I changed the beam type from Card to HyperCard, I got a wider stereo image, a less reverberant enviroment, but also less low frequencies. I did not notice this decrease in low frequencies in the 1st order when switching between cards and hyper. I don't know if I setup the plugin properly. I would like to ask if an ambisonic mic generally behaves as an ordinary mic - the more directional pattern, the less sensitivity for bass? Has an ambisonic mic a proximity effect when recording close to a source and decoding to the directional patterns?

And the last question. I'm curious about the artificial simulation of a spacing between the microphones. Len is sharing the recordings of his ambisonic mic and a pair of a regular mics in ortf. I wanted to generate ortf from Len's ambisonic mic and compare it with regular pair. Len is recommending Compass plugin. Unfortunately, I was not able to set it up correctly. Has anyone done it? (Ha ha, if this works, it would completely change our hobby. Maybe we will be pleasantly surprised one day)
« Last Edit: August 02, 2020, 02:41:12 PM by kuba e »

Offline Len Moskowitz (Core Sound)

  • Site Supporter
  • Trade Count: (3)
  • Taperssection Member
  • *
  • Posts: 381
    • Core Sound
Re: The World's Finest Blumlein Array
« Reply #19 on: August 02, 2020, 06:31:37 PM »
Please, is it possible to decompose a directional pattern of the 1st order to an omni and a fig. eight component? For example:
1st order Cardiod =  0.5 x (1st order omni + 1st order fig. eight)?

Yes, very easily.

Three of the four first-order B-format channels (X, Y and Z) are figure-8s. One of them (W) is an omni.

And with a beamformer plugin, you can create any combination of them and output any first-order microphone directivity pattern, including sub-cardioid, cardioid, super-cardioid and hyper-cardioid. As many you want, in any combination. Pointed in any directions that you want. Simultaneously.

And all of them are equal to or better than any of the world-class mono mics.

Quote
Directional patterns of 2nd order look complicated. Is there an easy way to decompose it to the components?

They are not decomposable - they are the second-order spherical harmonics, just as the first-order channels are the first-order spherical harmonics.

The second-order B-format channels are combinable with each other and with the first order channels to produce patterns like the second-order cardioid and hyper-cardioid directivity patterns you can see on our web site: www.core-sound.com/OctoMic/2.php

And there are many more, including ones with very, very low side- and back-lobe levels, and very tight directivity.

Quote
I would like to ask if an ambisonic mic generally behaves as an ordinary mic - the more directional pattern, the less sensitivity for bass?

No. What you're hearing may be an artifact of the plugin. Or an interaction of the room's bass response and the directivity pattern.

OctoMic goes down to around 30 Hz @ -2 dB. That's for everything OctoMic records.

Quote
Has an ambisonic mic a proximity effect when recording close to a source and decoding to the directional patterns?

There's no proximity effect if the target directivity is a B-format channel. There will be a proximity effect if the target is a combination of the channels. In that way it's pretty much the same as mono mics.

Quote
And the last question. I'm curious about the artificial simulation of a spacing between the microphones. Len is sharing the recordings of his ambisonic mic and a pair of a regular mics in ortf. I wanted to generate ortf from Len's ambisonic mic and compare it with regular pair. Len is recommending Compass plugin. Unfortunately, I was not able to set it up correctly. Has anyone done it? (Ha ha, if this works, it would completely change our hobby. Maybe we will be pleasantly surprised one day)

Spaced arrays with first-order directivity patterns can be decoded from a single ambisonic mic - not very well with a first-order ambisonic mic, and somewhat better with the first-order components of a high-quality second-order ambisonic mic. The process uses plane wave decomposition, which - as I understand it - works only at first-order.

The Harpex plug-in provides for it, but I'm not satisfied with its results.

That said, it's easy to do all of the ORTF arrays (and other spaced arrays) with OctoMic. In fact, you get the world's best behaved (and IMO sounding) ORTF, ORTF-surround and ORTF-3D arrays. How? Use multiple OctoMics.

For the ORTF-surround and -3D arrays each of the OctoMics plays the role of between two and six mono mics. All of the pointing angles are set in post. And the OctoMics give better results than the mono mics.

See the photos below.
« Last Edit: August 02, 2020, 10:37:53 PM by Len Moskowitz (Core Sound) »
Len Moskowitz
Core Sound
www.core-sound.com

Offline kuba e

  • Trade Count: (1)
  • Taperssection Member
  • ***
  • Posts: 492
  • Gender: Male
Re: The World's Finest Blumlein Array
« Reply #20 on: August 03, 2020, 06:04:49 AM »
Len, thank you very much for the explanation.

It is great that your mic goes around 30 Hz, -2dB. I am also starting to understand that when we are decoding from B-format to virtual polar pattern, there is no proximity effect. (I imagine that the virtual directional pattern is based on the differences of all microphones and therefore the microphone has no proximity effect and has a flat characteristic up to 30 Hz.) - this is not correct

Please, I have one more question about decomposing virtual directional pattern. Sparta beamformer plugin can make only Card and HyperCard pattern. So for example, If I make two 2nd order Cards at 90° and -90° and subtract them, do I get 2nd order Fig 8?

Yes, multiple OctoMics must give a great freedom when recording. It is a lot of channels for hobby tapers now. But this can change in a near future. I can imagine that one OctoMic will be connected with one connector and we will be able to record several of them with simple multi channel recorder. And one OctoMic would be also great in a Gutbucket's Oddball mics array. It's also nice for me that with OctoMic it's easy to learn the basics of recording. With one click, you can change the pattern or angle and hear the result immediately.

I appreciate how much work you've done in developing Ambisonic mics. Making development in a multinational company is easier in terms of money, time and capacity. You are a smaller company and it is even more valuable.
« Last Edit: August 06, 2020, 02:12:57 PM by kuba e »

Offline kuba e

  • Trade Count: (1)
  • Taperssection Member
  • ***
  • Posts: 492
  • Gender: Male
Re: The World's Finest Blumlein Array
« Reply #21 on: August 04, 2020, 04:07:08 AM »
I noticed now that the polar patterns are only up to 10kHz. I found in the specification that a calibrated OctoMic works in the range of 30Hz to 18kHz + -2dB. Just out of curiosity, what's going on above 10 kHz?

Offline Len Moskowitz (Core Sound)

  • Site Supporter
  • Trade Count: (3)
  • Taperssection Member
  • *
  • Posts: 381
    • Core Sound
Re: The World's Finest Blumlein Array
« Reply #22 on: August 04, 2020, 11:20:55 AM »
Still a very flat frequency response on-axis, but the patterns are not as clean.
« Last Edit: August 05, 2020, 05:25:23 PM by Len Moskowitz (Core Sound) »
Len Moskowitz
Core Sound
www.core-sound.com

Offline Len Moskowitz (Core Sound)

  • Site Supporter
  • Trade Count: (3)
  • Taperssection Member
  • *
  • Posts: 381
    • Core Sound
Re: The World's Finest Blumlein Array
« Reply #23 on: August 04, 2020, 02:49:19 PM »
It is great that your mic goes around 30 Hz, -2dB. I am also starting to understand that when we are decoding from B-format to virtual polar pattern, there is no proximity effect. (I imagine that the virtual directional pattern is based on the differences of all microphones and therefore the microphone has no proximity effect and has a flat characteristic up to 30 Hz.)

Close, but not quite correct.

When using the native B-format patterns, there is no proximity effect.

Patterns that derive from the B-format patterns - like cardioid, derived from omni and dipole - will have a proximity effect.
Len Moskowitz
Core Sound
www.core-sound.com

Offline kuba e

  • Trade Count: (1)
  • Taperssection Member
  • ***
  • Posts: 492
  • Gender: Male
Re: The World's Finest Blumlein Array
« Reply #24 on: August 05, 2020, 07:56:19 AM »
Thank you Len for the answers, it is nice from you. I like the theory about recording a lot. But I am hobby taper so some of my questions may not have been relevant.

The proximity effect of virtual patterns is not easy to understand intuitively. I have to read more about ambisonic mics.

But I guess I'm starting to understand the ambisonic mic frequency response and polar patterns. The ambisonic mics have very tight polar patterns up to 10 kHz. So the frequency response up to 10 kHz is measured on the microphone axis in a free field. Above 10 kHz, the comb filtering is starting to be dominant (because there is a distance between capsules) and the polar patterns are becoming beam polar patterns. And for this reason, the frequency response above 10 kHz is measured only in a diffusion field.
« Last Edit: August 05, 2020, 07:58:38 AM by kuba e »

Online DSatz

  • Site Supporter
  • Trade Count: (35)
  • Needs to get out more...
  • *
  • Posts: 3347
  • Gender: Male
Re: The World's Finest Blumlein Array
« Reply #25 on: August 05, 2020, 10:02:18 AM »
Len, hi. I just don't see how any pressure gradient transducer can fail to have proximity effect. Once the capsules have delivered their signals, with components from distant and close-up sources combined, how can any form of processing separate them out again?

Actually as I'm sure you know, it's the shapes of the wavefronts that matter; a plane wave won't provoke a bass boost if you manage to generate it an inch away from the capsule somehow. But any curved wavefront--such as one that emanates from a point-ish source in the near-ish field of a pressure gradient capsule--will be "read" by that capsule like a plane wave with a bunch more low-frequency energy in it. And once the cream has already been stirred into the coffee, [simplification here, but both experientially and experimentally valid:] no amount of further stirring will separate it back out.

--best regards

edited later to add: The same thing applies to breath noise (popping) and handling noise/solid-borne sound. If there's any way to process those things out of a microphone's signal without affecting the rest of the program material, I'll carry you on my own aging shoulders down to the Patent Office.
« Last Edit: August 05, 2020, 10:10:18 AM by DSatz »
music > microphones > a recorder of some sort

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (16)
  • Needs to get out more...
  • *****
  • Posts: 15700
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: The World's Finest Blumlein Array
« Reply #26 on: August 05, 2020, 11:07:09 AM »
Thanks D, was going to question that myself.  Kuba e, the proximity effect of virtual patterns shouldn't be overly mysterious in practice, having the same behavior as non-virtual patterns. The mathmatics of basic ambisonics assumes plane waves, but real world microphones used for ambisonics do have proximity effect, and it is adapted for in various ways as necessary in playback decoding.

As for high frequency pattern behavior, the degree to which single-point coincidence is achievable in real-world ambisonic microphones is determined by the size of the microphone elements and how closely they can be positioned to each other while maintaining the appropriate geometric arrangement.  The closer together they can be positioned, the higher the frequency the virtual polar patterns can be maintained, resulting from various combinations of  the individual microphone elements. Above a certain frequency determined by the relationship between the wavelength at that frequency and spacing between microphone elements, spatial aliasing begins to occur.  The patterns start to loose their smooth curves and intended shape and grow increasingly "spikey" in terms of small scale deviations resulting from adjacent regions of constructive and destructive interference (forming 3-dimensional comb-filter moire patterns) while the average spatial sensitivity tends toward omnidirectional overall.  The pattern increasingly degrades, while the overall response may or may not.

Such behavior is unavoidable but can be made non-problematic for the intended applications of the microphone by appropriate design. As long as that spatial aliasing transition can be pushed high enough in frequency, it doesn't pose an audible problem for music recording intended for human listeners.  No traditional ambisonic microphone is appropriate for recording ultrasonic sources such as bats and insects, etc.

This is the same acoustic effect one can sometimes hear at somewhat lower frequencies where it is audible and objectionable in some older and less well aligned PA setups, particularly earlier less well-designed line array PAs, when the sound gets audibly "swishy" with rapid regions of response variation as one listens while walking around in relation to the speaker array, while not being especially noticeable if standing still or sitting.  In better aligned and well-designed systems the effect is similarly pushed high enough in frequency to become unnoticeable, yet still occurs at frequencies above general human perception.
« Last Edit: August 05, 2020, 11:08:52 AM by Gutbucket »
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline Len Moskowitz (Core Sound)

  • Site Supporter
  • Trade Count: (3)
  • Taperssection Member
  • *
  • Posts: 381
    • Core Sound
Re: The World's Finest Blumlein Array
« Reply #27 on: August 05, 2020, 12:00:54 PM »
It's definitely counterintuitive, but according to our DSP gurus, it's really so.

Take, for example, the B-format W-channel. It's omni, and for OctoMic it's a near-perfect omni across frequency.

It's really, truly an omni and has no proximity effect, despite the capsules in the array being cardioids.

I just ran a quick & dirty test that confirmed it.
« Last Edit: August 05, 2020, 12:41:29 PM by Len Moskowitz (Core Sound) »
Len Moskowitz
Core Sound
www.core-sound.com

Offline Gutbucket

  • record > listen > revise technique
  • Trade Count: (16)
  • Needs to get out more...
  • *****
  • Posts: 15700
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: The World's Finest Blumlein Array
« Reply #28 on: August 05, 2020, 01:19:12 PM »
I think that would be expected given the pressure omni component in each of the A-format cardioid capsules which are collectively summed to produce the W-channel, while the bi-directional component of each A-format cardioid in which proximity effect would be expected to manifest is differentially cancelled out.

Can you similarly test X, Y or Z channel?


This would seem like it might be related to Near-Field / Distance Compensation in ambisonic decoding schemes, so I did a quick search and found the paper linked below. The last lines in the foot note quoted below about a properly aligned ambisonic microphone "encoding distance by virtue of accurate transduction of the incident wavefronts" may be indicative of the behavior Len is describing.

From: https://pdfs.semanticscholar.org/e101/9d172e9263296925f6b528959bfa1236ad4c.pdf

Excerpt of note from that paper:
"The exact decoder matrix recreates the pressure and velocity at the central position under the assumption that the wavefronts are planar, i.e., sources and loudspeakers at an infinite distance. Sources and loudspeakers at finite distances produce wavefronts with a “reactive” (or imaginary) component, which is perpendicular to the direction of propagation, in addition to the “real” component, which is parallel to the direction of propagation. This results in the wellknown bass-boosting proximity effect in directional microphones. It is important to understand that this is an actual physical effect, not a design flaw in the microphone or loudspeaker.4"

footnote 4-
"The implication for B-format signal encoding is that the X, Y, and Z signals must have a low-frequency boost and phase shift relative to the W signal, the amount of which is a function of the source distance. For natural acoustic sources, a properly aligned Soundfield-type microphone does this by virtue of accurate transduction of the incident wavefronts, and thereby encodes distance. For synthetic sources, this must be included in the encoding equations. Further discussion of this is in AppendixB."
« Last Edit: August 05, 2020, 01:22:26 PM by Gutbucket »
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Online DSatz

  • Site Supporter
  • Trade Count: (35)
  • Needs to get out more...
  • *
  • Posts: 3347
  • Gender: Male
Re: The World's Finest Blumlein Array
« Reply #29 on: August 05, 2020, 01:27:49 PM »
Len, when you synthesize an omni pattern, it's no surprise that proximity effect will be minimal. The same is true of conventional, dual-diaphragm studio microphones when set to "omni" even though their output is then the sum of two back-to-back cardioids. Each cardioid is a ~50/50 mix of pressure and pressure gradient, and in the omni setting the pressure gradient components are summed in opposite polarity, so they mostly cancel. Not completely, though, because they're not at the same point in space. So even in the omni setting there's still some small amount of proximity effect which is measurable--but of course the lower you go in frequency, the smaller the fraction of a wavelength that's represented by the distance between the two "half-capsules."

Would you please try the same experiment with the Blumlein pair that this thread is about and let us know what you hear and see?

--best regards
« Last Edit: August 05, 2020, 01:34:54 PM by DSatz »
music > microphones > a recorder of some sort

 

RSS | Mobile
Page created in 0.075 seconds with 44 queries.
© 2002-2024 Taperssection.com
Powered by SMF