Saxophone Quartet - Binaural Recording with five second set-up.

[NorseHorse]

Recorded a symphonic band two weeks ago. At the end of the show, I went to take down the Crown SASS mic I used, and the director says, "Did you want to get this?" I look over and see that four of the saxophone players had come out from the stage and into the audience with stands. They were going to play a little "bonus" piece and were about to start at any moment. Luckily, the recorder was still rolling and I had slack in the mic lines. I told them to hold for a moment, quickly grabbed the mic stand, moved it to a spot in front of the quartet where I thought it would sound good, and gave 'em the thumbs up. You can hear a girl ask "What is that thing?" as the recording starts!

I'm sharing this recording because a) there's been discussion about binaural techniques recently, b) there's only one other saxophone quartet sample on GearSlutz, and c) I only had about five seconds to observe, decide, and place the mic, which is rather unusual. Anybody else have experiences like that?

Enjoy! (Don't forget to listen on headphones...)

[boojum]

That's got it. That's binaural. Thanks. I was afraid my hearing was whacked after listening to some other stuff.


Cheers

[Michaelrophone]

Very nice!

[ajfarber]

Ah, the "Liberty Bell March"!

[Shureman]

Really enjoyed the recording! I do a lot of impromptu recording too. Your story sounded very familiar!

What file sharing outfit do you use? I use divshare. Very quirky at times.

[smart1222]

not bad.

[newyorkbrass]

In-Tune Saxophones! (or is that the mic?)

Sounds awesome.

[Mark A. Jay]

I'm probably going to be derided for some of my comments, but such is life. Having said that, here we go...

To start with, I think this sounds very good, and very natural; I think your approach worked exceptionally well in this space / with the SASS used at the distance it was used from the performers. I have been in your shoes as well (near-zero lead / adaptation-time, as we probably all have) and it's always a delight when the results 'just work' out to be very good. Such is the case here, and in my opinion, you're to be commended for your work.

However - and my point here pertains only to the technical aspects of binaural and not to the aesthetic that was achieved in your sample recording (as I think it is very good sounding) - it's not technically accurate to call a recording made with an SASS microphone binaural. It's a stereo recording.

In my opinion, calling a SASS-derived recording binaural is not unlike calling an an ORTF configuration an A-B configuration. While it's true that both techniques result in a two-channel recording and that both use two microphones, and that both feature microphones spaced some distance apart, they are different enough technically to warrant each of them having a distinct name. True, they have traits that are similar to one another, and in certain circumstances, the recordings made using each technique will sound similar, but it's not technically accurate to call one by the other's name. This is likewise the case between a SASS recording and a binaural recording.

Indeed, even Crown proper say (in their literature) that it is a stereo microphone (the product name is an acronym for Stereo Ambient Sampling System) and never in their technical literature - at least anywhere that I have seen - do they use the word binaural - always "stereo".

On the other hand, the SASS has a lot in common with other stereo approaches that use some form of barrier or disc between the microphone elements to create a shadowing effect similar to that of binaural (and notably absent in something like an A-B or an ORTF). Also, the SASS uses a microphone element spacing similar to that of a binaural mannequin microphone to yield arrival time differences similar to those that can be found in a binaural mannequin microphone - clearly Crown recognized the importance of shadowing and arrival time in terms of how we localize sound.

So yes, the SASS has elements that approximate a true binaural microphone, but since it's an approximation it's really a stereo microphone. Again, for a microphone to truly qualify as a binaural microphone it has to have three things (as has been discussed in the "Binaural vs. ORTF" thread): a) a head that approximates the dimensions and shape of a human head, b) artificial ears on the mannequin head, and c) microphones in the ear canals of the artificial ears.

I too have an SASS microphone in my tool box which I have used successfully on numerous occasions, However, I prefer to use the Neumann KU-100 because it is truly a binaural microphone while the SASS is a stereo variant, and indeed, if true binaural is desired then I use the KU-100.

Still, from an aesthetic perspective, I think what you posted is a great-sounding recording and I am not taking anything away from the quality of your work as I think it does sound very good. I simply want to be as technically accurate as possible, thus my comments about the differences between the two approaches.

[Shureman]

Thanks for drawing the distinctions between stereo and binaural.
Precision is helpful.

I really think a lot personal preferences for liking binaural, ortif, MS, and on and on is a function of the way individuals percieve things, and not that any one method is vastly superior in depicting a stereo illusion. There are certain aspects of each miking method which are more likely to "trick" people into percieving a sense of audio 3D. What works for some does not work for others.

I listened to the sax quartet, and I listened to several of the binaural recordings on Mark's website (all with headphones). Both methods sounded equally as spatial to me. I guess I'm lucky. On the other hand, ortif mic configurations, X/Y.... don't. In that, I'm unlucky since so much stuff out there uses these methods. I attribute this to whatever cues my ears and brain look for to be fooled into thinking there is a symphony orchestra in my listening environment aren't present, but they obviously are present for many, if not most others. A good MS matrix fools me everytime!

[NorseHorse]

Originally Posted by Binaural_Mark

So yes, the SASS has elements that approximate a true binaural microphone, but since it's an approximation it's really a binaural microphone. Again, for a microphone to truly qualify as a binaural microphone it has to have three things (as has been discussed in the "Binaural vs. ORTF" thread): a) a head that approximates the dimensions and shape of a human head, b) artificial ears on the mannequin head, and c) microphones in the ear canals of the artificial ears.

While true binaural seems to have been clearly defined, that is not ALL binaural. After all, "true binaural" only approximates a real human head, approximates real ears, and approximates ear canals. As you mentioned, all of this is artificial.

Therefore, the core purpose of binaural stereophony is to closely approximate the sonic cues achieved by human hearing through the use of spacing and baffling techniques which mimic the human head, for the purpose of playback over headphones. To that effect, the SASS is binaural. If a piece of well designed plastic baffling is more effective at reproducing what we hear than something that looks like "pinnae", then I say it is more truly binaural in it's approximation of the human head and ears. After all, even so-called "true binaural" is fake in nature.

[Shureman]

Originally Posted by NorseHorse

While true binaural seems to have been clearly defined, that is not ALL binaural. After all, "true binaural" only approximates a real human head, approximates real ears, and approximates ear canals. As you mentioned, all of this is artificial.

Therefore, the core purpose of binaural stereophony is to closely approximate the sonic cues achieved by human hearing through the use of spacing and baffling techniques which mimic the human head, for the purpose of playback over headphones. To that effect, the SASS is binaural. If a piece of well designed plastic baffling is more effective at reproducing what we hear than something that looks like "pinnae", then I say it is more truly binaural in it's approximation of the human head and ears. After all, even so-called "true binaural" is fake in nature.

All good points. There is likely no exact point where a binaural capture device can be deemed perfect - to many variables - down to such yucky things as how the sound waves interact with hair on and in the ear for example. But as NorseHorse suggests, a good approximation, which in vivo has that you are there quality, including front to back cues, should be deemed binaural, don't you think?

Mark, you are far more versed in this topic than I. Has anyone designed an adjustable binaural recording head - one which can be easily manipulated at the recording session to continuously change mic spacing, "ear" shape, head shape, etc... while someone is monitering the capture offsite till the focus is as good as possible? To clarify, by "continuously" I mean making changes, uninterrupted, while the music is in progress with automation, not music, listen, stop, adjust, music, listen, stop, adjust....

[Mark A. Jay]

Originally Posted by NorseHorse

While true binaural seems to have been clearly defined, that is not ALL binaural. After all, "true binaural" only approximates a real human head, approximates real ears, and approximates ear canals. As you mentioned, all of this is artificial.

Therefore, the core purpose of binaural stereophony is to closely approximate the sonic cues achieved by human hearing through the use of spacing and baffling techniques which mimic the human head, for the purpose of playback over headphones. To that effect, the SASS is binaural. If a piece of well designed plastic baffling is more effective at reproducing what we hear than something that looks like "pinnae", then I say it is more truly binaural in it's approximation of the human head and ears. After all, even so-called "true binaural" is fake in nature.

While I agree with much of what you have said (and indeed, from an aesthetic perspective, I have no issues with the SASS approach - I even own one - as I have ststed in this thread and (I believe), others), there are some things with which I do take issue.

The fact that companies / institutions such as Bruel & Kjaer, Head Acoustics, Cortex, Neumann, M.I.T. (vis-a-vis their research on binaural via the KEMAR), University of Aachen (who have worked closely with Klaus Gennuit of Head Acoustics - a rather learned individual) - as well as countless resarch universities around the world - have been working on very carefully defining the HRTFs and few (if any) have suggested that anything but a mannequin head with pinnae (and ear canal-located transducers) was the best means by which the HRTFs should be implemented. Yes, they argue about which ear shape is 'correct' and I concede that there is no universally-correct ear - it is an approximation. However, an approximation that is based on a mountain of reasearch, conducted at some very well-equipped universities and laborotories over the years. I also believe that not unlike a lot of things that I see in this forum, it comes down to what I call 'intellectual bragging rights' as to whose ear is chosen as the 'correct' anthromorphy.

Indeed, as I have mentioned before, there is actually an ISO committe whose job it is to homolugate a 'standard' ear shape. The size and geometry of (most) mannequin heads is based, by the way, on other anthromorphy data as to the 50th percentile cranium.

Indeed, when you look at the work done by M.I.T., University of Aachen, Bruel & Kjaer, Head Acoustics over the past 30 years or more, the one thing that they have in common for their research and development is a facsimilie of a human head as the de-facto standard for binaural (and not the SASS, not the Jeklin Disk, not ORTF, not XY, not A/B). I myself have been working in the field of perception for the past 20 years, and as I have mentioned in other posts, the world of research acknowledges that a binaural mannequin head is the transducer of choice for accurate binaural signal acquisition, especially when perceptual studies (playback of the binaurally recorded sounds) are to be carried out whith listeners using headphones. This is not my opinion, it is the fact of the matter and this method is used at every manufacturer with whom I have worked who is interested in the qualitative aspects of sounds in a controlled approach.

On the other hand, I have also acknowledged that the crown SASS approach approximates binaural (and "binaural" to me means 'mannequin head-based HRTFs) as it gets the concept right but not the details. I feel like I have to say it yet again - I like the sound and image presented by the SASS, I truly do. My issue is with technical accuracy and NOT (as I have stated before) the aesthetic.

With all due respect, I believe in mathematics and physics; these have been the cornerstone disciplines of the research work done at the various companies and organizations that I have cited, as well as countless others. To say that the SASS approach is the equivalent (or as you imply, better than) a mannequin with pinnae, and ear canal-located microphones is to turn a jaundiced if not wholly blind eye to an overwhelming body of evidence to the contrary.

We like what we like, and that's as it should be. Math and physics, however, don't lie.

[Mark A. Jay]

Originally Posted by Shureman

All good points. There is likely no exact point where a binaural capture device can be deemed perfect - to many variables - down to such yucky things as how the sound waves interact with hair on and in the ear for example. But as NorseHorse suggests, a good approximation, which in vivo has that you are there quality, including front to back cues, should be deemed binaural, don't you think?

Mark, you are far more versed in this topic than I. Has anyone designed an adjustable binaural recording head - one which can be easily manipulated at the recording session to continuously change mic spacing, "ear" shape, head shape, etc... while someone is monitering the capture offsite till the focus is as good as possible? To clarify, by "continuously" I mean making changes, uninterrupted, while the music is in progress with automation, not music, listen, stop, adjust, music, listen, stop, adjust....

Yes, I also agree that the SASS approach is a good approximation. I personally like the sound of that microphone, and have myself generated some very good recordings with it.

Wow....your other question...I doubt that anyone has designed the transducer you have described. If for no other reason, the sheer number of combinations. Think about it - as a minimum, here are the variables (the ones that just come to mind):

a) Head geometry (itself comprising volume, x,y,z values, etc)
b) ear location (translation) in x, y, and z
c) pinnae shape *

*and here is where it gets very complicated. However, what effect the shape of the pinnae have on the HRTF alone or in concert with the other variables is wholly beyond me. I say this in all sincerity - I believe that the pinna shape as it pertains to perception is probably if not a lifetime's research work, the it's pretty close to that.

I suppose some DOE methodology could help minimize the size of the experiment in order to determine which variable have the greatest impact. Alas, that implies that there are data or some means of generating said data.

A lot of very good questions - and I wish I had a lot of really good answers for you. What I can give you are some thoughts and observations - but I doubt they will satisfy your questions fully. I think some of those questions are better left to reasearchers with the proper measurement spaces (anechoic, hemi-anechoic, diffuse field chambers etc) as well as credentials. However, here are my thoughts:

Were I to venture a guess, I think there will be one primary factor about the pinna shape and geometry. That is, there may be one aspect of it (i.e. curvature vs. area or some other parameter - no, I have no idea off the top of my head) that is the primary factor in affecting the directivity / the role that the pinna plays in the overall HRTF. However this is only a guess.

I do know that a lot of research has been done into how the HRTF changes as a function of proximity (source to ear distance, azimuth, etc). I think the folks at M.I.T. did a lot of work in this area if memory serves (and it may be on-going).

So what are the alternatives to a 'standard' ear and head shape? Well, many use omni's mounted to eyeglasses (which achieves the shadowing effect, and probably more faithfully than does the SASS or otehr approaches). Yet, this is specific to the user - so it's not standard. That doesn't make it 'bad' it just makes it unique.

In-ear types are the next alternative, and indeed, this approach makes sense (in my opinion). The issue with in-ear binaural is that it can be rather difficult to position the microphone element in the same location in the ear canal every time, which means that there will be variations (from recording to recording). However, if you had the luxury of a diffuse field space, I can see a) inserting the mics into one's ears, and b) walking into the diffuse space with a known noise source, and c) acquiring a miniute or so of data that could be used as a sort of correction (and if you did this prior to each recording, you could generate a correction for any recording to any recording - it's just a question of what vectors you choose to compute the FRF. That might be enough to allow normalization of one to another, though I suppose that there could be some minor time delay differences (due to how far into / out of the ear canals the mic elements were), but my gut tells me that these would probably be minor compared to the differences in spectra borne of dissimilar locations (run to run / recording to recording) in the ear canals.

Remember that when making acoustic measurements (which is effectively what a recording is), the near-field is generally to be avoided because small changes in source-to-mic position translate into very large changes in measured response (or rather, usually do) - which leads to poor reproducability. That's why a speaker measurement (for example) done at 1 m distance is pretty 'forgiving' (that is, if the microphone is at 0.99 m versus 1m from the speaker, the spectra measured there (with a given excitation) will be nearly identical), but a speaker measurement made in the near field is best done with a jig that precisely positions the microphone in as close to the identical position as before.

Still, the near-field approach has some great advantages (i.e. measuring speakers etc) because you can really gain signal to noise ratio in a speace that would (at the desired measurement location) be otherwise too noisy.

Back to the in-ear approach... if you consider the size of the microphone element relative to the size of the ear canal, it can be significant. As such, when it is in the ear the acoustic impedance of the ear canal has (in my opinion) got to be altered by the presence of the microphone if it is 'dimensionally / volumetrically significant' relative to the volume in which it is placed. I think this makes sense if you think about it - think of an 'amplified' example - like how the boundary conditions would change when you had a 2-liter soda bottle, and then how it would differ as you moved a fist / arm into the soda bottle. On the otehr hand, something the size of a straw or a small dowel would have little effect.

This is why there exist (in the audiology / instrumentation world) probe microphones that are very small (dimensionally), much smaller than most omni elements that go on glasses or that people put in or on their ears. Indeed, that's a whole 'nother real of acoustics / psychacoustics and perception with which I am only vaguely familiar.

Anyway...back to the original question: Who would make such a device as you have suggested? It really would be a company or a university or some otehr public trust with enough intellectual and financial resources to pull off such a Herculean research feat. Having worked in industry for 20 years now, I can tell you that only when economic times are good do people on the R&D staffs of companies / universities breathe a little easier.

If someone were going to make a pile of money from putting the research into building and using the device you have suggested, it would get funded overnight - if not sooner.

Truly, what's required is a wealthy philanthropist who has a burning desire to find out about these issues and who could aford to drop ma few million in research over several years. I don't know of any, but if I did, I would hit said person up with a research proposal, quit my job, and then hire people way, way, way smarter than me to implement the research.

[Mark A. Jay]

Originally Posted by Shureman

... Mark, you are far more versed in this topic than I. Has anyone designed an adjustable binaural recording head - one which can be easily manipulated at the recording session to continuously change mic spacing, "ear" shape, head shape, etc... while someone is monitering the capture offsite till the focus is as good as possible? To clarify, by "continuously" I mean making changes, uninterrupted, while the music is in progress with automation, not music, listen, stop, adjust, music, listen, stop, adjust....

Shure: I meant to touch on this, but my digits were weary.

Assuming that someone could in fact do this (let's assume that they can), then it would probably only b eth start of anotehr long road of research. You are right that the various combination should be listened to, but certainly, every run should be recorded (or as we say in the test and instrumentation world, 'acquired'). In this way, subsequent work could commence, namely, double-blind juried listening tests. Yes, these could be generic (i.e. all salient demographics), or could be demographic specific. For instance, double blind juried tests are a great way to ascertain whether (for example) males 35-45 have the same preference of the sounds as did females of that same age group.

Now, the bigger issues creep in. For the juried test to be valid, the data being presented in the acquisition space (and thus incident upon the ears of the mannequin) should be identical run to run. This lets a live performance out, because no matter how precise the musician(s), no two performances will be exactly alike. I know, they could be very close, but from a scientific standpoint, the stimuli presented to the mannequin head would have to be identical.

One way to do this would be to (for instance) play back a recording of a specific cut of music over the speakers, then acquire the first data set (the first recording) with the head in 'position number 0'. Then, position number 1 would be set in the mannequin, and the track played back. This would be repeated ad infinitum (well, almost). Now...would the acoustics of the room play a role? Most assuredly. Would the location of the mannequin relative to the speakers play a role? Most assuredly. How about this for a variable - the genre of music. What choice should be taken? Who decides what are the 'right' speakers to use?

How about this for another variable - would the listening tests be carried out at calibrated levels? In other words, would the recordings be played back that they were the same loudness as was imparted to the ears of the mannequin? For the test to be valid, they would have to be presented this way (fortunately, juried study listening software packages allow for a proper loop calibration to be performed making this not only possible, but the standard for these tests).

My guess is that it would be somewhat 'cleaner' to obtain a vast array of binaural impulse responses (for the mannequin in its various configurations) and use those to convolve a signal to represent the mannequin configuration - and those signals would be the ones used for the listening studies / juried evaluations.

You can see how something that on the surface isn't all that complicated can quickly become immensely complicated.

Why would it be necessary to be so precise? Because the validity of the experiment would hinge in no small way on the accuracy of the data, especially if these data were then to be played to jurors in a double-bling listening study. So yes, you would have to be uber-precise.

Anyway, I know this sounds crazy or even Orwellian (but in a good way - a little 'double-think' now and again might help...), but if one knew their own ears' impulse responses, then I suppose that it would be possible to normalize any binaural response to the binaural response of the particular individual.

It's late...and I need to work yet some more, but I wanted to throw that out there.

[John Brook]

Love the recording, but something really strange. Listening with headphones, it sounds like the instrumentalists are behind me, or that I'm facing the audience. How is that possible?

And what's with that lady yapping all through the performance????

[Mark A. Jay]

Originally Posted by John Brook

Love the recording, but something really strange. Listening with headphones, it sounds like the instrumentalists are behind me, or that I'm facing the audience. How is that possible?

And what's with that lady yapping all through the performance????

I'm not sure if you're commenting about my KU-100 samples (in the "what's you favorite binaural recording?" thread) or the S.A.S.S. one posted by Norse in this thread as your post wasn't addressed to anyone specifically, however, I'll assume you're talking about his posted sample - and commiserate.

I don't know if it's a commentary on social manners and decorum, or if it's that people assume that (despite posted signs that there is a recording taking place) the microphones in use won't pick them up while talking. I've run into this myself - a binaural mannequin head can be pretty merciless as it will record - for better or for worse - what a person would hear at the location of the head. People either assume what we do is less important than their need to talk about whatever, and they just don't care, or they simply assume they won't be heard in the recording. You can't blame people for that.

I know a gent who travels the world recording various recitals, arias, symphonies and the like, and was once comissioned to record a symphony (or similar). Mind you, the organizing body wanted a full house for the recording and had posted placards on the doors, in the W/C, the lobby - pretty much everywhere asking the patrons to remain as quiet as possible due to the recording. It worked flawlessly - the head had been positioned about 20 rows back from stage, and everything was going wonderfully...when suddenly, during one of the very quiet passages in the piece, a gentleman a few rows in back of the mannequin had fallen asleep, and yes, was snoring. Loudly. Since it was a classical piece that movement turned out to be useless because of that one incident. If memory serves, they used the movement from the following night - absent snoring patrons - and no one was the wiser.

Anyway, therein lies the danger of in-situ binaural or similar techniques and a very real advantage for ORTF or some close miking techniques. Realism versus control. So, if your comments and questions about the spatial aspects (of the sample) are about Norse's recordings, I'll leave him to answer that. If they are about mine, post in the 'favorite binaural recordings' thread and I'll be happy to try and answer them there.

[John Brook]

You guess correctly - I was referring to the S.A.S.S. recording of the saxophones.

Well, snoring's a first, but I have a problem with one conductor who gulps are at the beginning of every composition and every major downbeat. It comes through sounding like he's snorting. I can cut some of them, but not all of t hem. Argh.

[dallas simpson]

Just to chip in about the 'apparently coming from behind' phenomenon. This is a common aspect of binaural especially when there is a large mismatch between the HRTF (head related transfer function) of the recording system and the listener's head / ear combination. It can also occur with 'budget quality' headphone / playback systems where there is a phase inaccuracy between the channels due to poor transducer / channel matching.

With computers phase inaccuracy can also occur when there is an error in the software / soundcard area and a sample or more gets 'dropped' somewhere along the line on one channel on playback, effectively causing a 'phase shift' of the audio, so this can be more of a problem with computer listening.

I made a mini-disk recording some years ago where this apparently happened in the MD on a section of a binaural recording and it was very difficult to restore the binaural image back to normal!

Front rear reversal can also be considered as a fraility due to the bi-lateral symmetry of the human head/ear/body morphology. The localisation of centre front sound images, and distinguishing from rear localisation, relies on minute discrimination of the shadowing and phase location due to the pinnae (external ears) as the time differences between left and right ears will be the same for front and back centre localisation. Small errors of mis match between the recording system and the listener can therefore cause front rear perceptual reversal.

The converse is true - binaural images off centre and to the sides and even more so, moving images, - are much more stable between differing head / ear morphology of listeners.

[John Brook]

Thanks, Dallas, for that explanation. I'll go with my laptop, as the Sony's are pretty decent (well, if price is an indicator).

[Mark A. Jay]

Here's an excerpt of a binaural recording (from a rehearsal) that I made some time ago. The microphone utilized was a Neumann type KU-100.

I don't remember exactly where I placed the mic, but I seem to recall it being at a height of about 4 feet (a little over 1 meter) from the floor, roughly center-stage. It 'faced' the orchestra (the violins at left, celli at right, etc).

As this is a binaural recording, you'll need headphones to get the full and proper effect.

Let me know what you think of it. I hope you like it.

[Shureman]

A nice sounding recording, full and warm, but I'm afraid the binaural effect got totally lost in the journey from your source to my computer. I also noticed a more prominent reverb trail in one of the channels. Room's acoustics?

Anyway, I wore headphones, right and left cups properly situated. The binaural effect was stunning for the virtual haircut, not at all for this excerpt.

My compliments to the orchestra as well. Can you share who they are.
I do the Louisiana Philharmonic.

[Mark A. Jay]

Originally Posted by Shureman

A nice sounding recording, full and warm, but I'm afraid the binaural effect got totally lost in the journey from your source to my computer. I also noticed a more prominent reverb trail in one of the channels. Room's acoustics?

Anyway, I wore headphones, right and left cups properly situated. The binaural effect was stunning for the virtual haircut, not at all for this excerpt.

My compliments to the orchestra as well. Can you share who they are.
I do the Louisiana Philharmonic.

I'm not sure how the effect was "lost" as you say. I can offer no good explanation to you. At the same time, others who have listened to it (the same excerpt written at the same bitrate in mp3 format) thought the binaural effect was dead-on (citing placement of the strings, celli, horns, tympani, etc).

As far as the reverberation disparity, as I said, the mic was essentially on center-line of the stage during rehearsal, and the auditorium was set up in a pretty symmetric manner (with respect to the radiused diffusers overhead and so forth). There were no patrons in the house at the time of rehearsal, so from a boundary condition perspective, there's not much difference, so there should be no real delta in the house, per se.

As far as the orchestra goes, it's a community (volunteer) orchestra here in Michigan, and I think this was the 2nd or 3rd rehearsal before the actual concert.

Thanks again for the comments though - it is very important to hear what people think of my work, whether good, bad, or in-between.

[Shureman]

Please extend my compliments to the orchestra if you have the opportunity. That's one superb sounding community orchestra!

I'll give the recording another listen.