Keith_W
Major Contributor
After reading a bit about BACCH ORC, which uses binaural mics to capture the HRTF, it struck me that there are many different ways you could take measurements for room correction. The difference is how many environmental factors you want to incorporate in your measurement. In a nutshell:
- SPS (single point sweep with omni mic) with no obstructions: Move the sofa, coffee table, and anything obstructing line of sight to the speakers away, and place the mic at MLP and do a sweep. Use that to generate room correction. This is the usual recommendation for Acourate and Audiolense, and is also recommended by @mitchco in his book.
- SPS with sofa: Leave the sofa and all the usual furniture in situ, place the mic at MLP, and do a sweep. The idea is to do a correction with the normal configuration of furniture when listening.
- MMM: this time you sit at the MLP and move the mic around your head. This not only incorporates the furniture, but also your body in the measurement.
- Binaural: This is at the other extreme. The effect of the sofa, body, and HRTF are incorporated into the measurement.
As you can see, strategies for taking measurements for room correction range from one extreme to the other. At one extreme, environmental factors are minimized, and only the room is corrected. At the other extreme, the effect of room, furniture, body, and head, and pinna are incorporated into the measurement and therefore corrected. Yes, I know that MMM and Binaural can not capture phase and timing information. Let us ignore that for the moment and focus only on FR correction.
Now obviously, accounting for body/head/pinna means shorter wavelengths are also incorporated. Because of the small size of the head compared to the wavelength, the head does not even begin to affect the measurement until 1kHz and up, and the pinnas from 5-6kHz and up. Here is an image I found online of the HRTF of three individuals:
What you can see is that everything <1kHz is broadly the same. Above this, there is quite marked variation. So I can see that depending on whether you believe that full frequency correction vs. correction below Schroder only should be done, as we discussed in another thread I started ("should we correct to Schroder or full range"), it will influence the way you take your measurement. If you believe in correcting below Schroder only, there is no benefit of incorporating the HRTF into your measurement. So for the purpose of this discussion, let us assume that you are correcting full range.
I do not own a binaural mic, but I have done experiments comparing sweeps of sofa vs. no sofa vs. MMM. About the only thing I haven't done is duct tape the mic to my balding head although I am considering doing that as an experiment! As you would expect, the results are markedly different. I can't find my MMM measurement, but here is a comparison of the FR of my system with/without sofa, before DRC:
You can see that the sofa boosts the bass frequencies and especially the high frequencies. If I were to make a correction, the result would be markedly different. The MMM looked different again, from my recollection there was even more of a treble boost.
It has also occurred to me that headphone guys want to incorporate as many environmental factors as possible. The importance of the HRTF is emphasized when generating headphone corrections, with some (like Griesinger) advocating tailoring the FR to the actual subjective response in your brain. No crappy binaural mics here, it's either deep insert mics, or better still, using test tones. This works by adjusting the volume of a test tone until it subjectively matches the reference tone, and this generates an extremely personalized transfer function which incorporates ear canal, the tension of the eardrum, movement of the ossicles, fluid pressure in the Organ of Corti, hearing loss, and so on. I have tried this, and it generates a horrible sounding correction. I am unsure if it is the method which is at fault, or my inability to volume match the tones. But that is a separate discussion.
Given that different authorities advocate different methods of obtaining a raw curve for room correction, some advocating for minimal environmental factors, and some others advocating for incorporating everything including the head and pinna, the only way for me to make a decision was to try them all. I took measurements with/without sofa and MMM, used those measurements to make filters, and listen for what I subjectively preferred the best. For me, I liked the sound of "SPS without sofa" the best. I felt that the other two methods knocked back the bass and treble too much and gave the music less impact and clarity.
I would like to opinions on:
- whether you believe in maximizing or minimizing the effect of environmental factors when designing your room correction,
- using binaural mics for room correction,
- which sounds best to you
- your reasoning behind your decision.
I also welcome any criticism of any flaws in my methods and reasoning.
- SPS (single point sweep with omni mic) with no obstructions: Move the sofa, coffee table, and anything obstructing line of sight to the speakers away, and place the mic at MLP and do a sweep. Use that to generate room correction. This is the usual recommendation for Acourate and Audiolense, and is also recommended by @mitchco in his book.
- SPS with sofa: Leave the sofa and all the usual furniture in situ, place the mic at MLP, and do a sweep. The idea is to do a correction with the normal configuration of furniture when listening.
- MMM: this time you sit at the MLP and move the mic around your head. This not only incorporates the furniture, but also your body in the measurement.
- Binaural: This is at the other extreme. The effect of the sofa, body, and HRTF are incorporated into the measurement.
As you can see, strategies for taking measurements for room correction range from one extreme to the other. At one extreme, environmental factors are minimized, and only the room is corrected. At the other extreme, the effect of room, furniture, body, and head, and pinna are incorporated into the measurement and therefore corrected. Yes, I know that MMM and Binaural can not capture phase and timing information. Let us ignore that for the moment and focus only on FR correction.
Now obviously, accounting for body/head/pinna means shorter wavelengths are also incorporated. Because of the small size of the head compared to the wavelength, the head does not even begin to affect the measurement until 1kHz and up, and the pinnas from 5-6kHz and up. Here is an image I found online of the HRTF of three individuals:
What you can see is that everything <1kHz is broadly the same. Above this, there is quite marked variation. So I can see that depending on whether you believe that full frequency correction vs. correction below Schroder only should be done, as we discussed in another thread I started ("should we correct to Schroder or full range"), it will influence the way you take your measurement. If you believe in correcting below Schroder only, there is no benefit of incorporating the HRTF into your measurement. So for the purpose of this discussion, let us assume that you are correcting full range.
I do not own a binaural mic, but I have done experiments comparing sweeps of sofa vs. no sofa vs. MMM. About the only thing I haven't done is duct tape the mic to my balding head although I am considering doing that as an experiment! As you would expect, the results are markedly different. I can't find my MMM measurement, but here is a comparison of the FR of my system with/without sofa, before DRC:
You can see that the sofa boosts the bass frequencies and especially the high frequencies. If I were to make a correction, the result would be markedly different. The MMM looked different again, from my recollection there was even more of a treble boost.
It has also occurred to me that headphone guys want to incorporate as many environmental factors as possible. The importance of the HRTF is emphasized when generating headphone corrections, with some (like Griesinger) advocating tailoring the FR to the actual subjective response in your brain. No crappy binaural mics here, it's either deep insert mics, or better still, using test tones. This works by adjusting the volume of a test tone until it subjectively matches the reference tone, and this generates an extremely personalized transfer function which incorporates ear canal, the tension of the eardrum, movement of the ossicles, fluid pressure in the Organ of Corti, hearing loss, and so on. I have tried this, and it generates a horrible sounding correction. I am unsure if it is the method which is at fault, or my inability to volume match the tones. But that is a separate discussion.
Given that different authorities advocate different methods of obtaining a raw curve for room correction, some advocating for minimal environmental factors, and some others advocating for incorporating everything including the head and pinna, the only way for me to make a decision was to try them all. I took measurements with/without sofa and MMM, used those measurements to make filters, and listen for what I subjectively preferred the best. For me, I liked the sound of "SPS without sofa" the best. I felt that the other two methods knocked back the bass and treble too much and gave the music less impact and clarity.
I would like to opinions on:
- whether you believe in maximizing or minimizing the effect of environmental factors when designing your room correction,
- using binaural mics for room correction,
- which sounds best to you
- your reasoning behind your decision.
I also welcome any criticism of any flaws in my methods and reasoning.
Last edited:
, I usually apply MMM while removing or covering with absorbers if possible objects close to the LP like chair/sofa as their close reflections alter the sound field above bass, a recommendation I actually got from Dr. Ulrich Brüggemann from Acourate. If correcting only bass they can stay but on the other hand the differences are not high and in the end what which result personally sounds best to us matters.
