Friday 20 May 2016

Updated Room MEASUREMENTS & MUSINGS on Importance of Sonic "Accuracy" and the Audiophile.

Chapter I: Another Round of Room Measurements...

Recently, I acquired some more LP's, got a few more IKEA Kallax storage units for said LP's and put up some art as well. Here are front and rear shots of the sound room the other day...


With the physical changes and reviewing my post from last year regarding the use of (((acourate))) along with suggestions made by Dr. Uli Brueggemann and Mitch Barnett (aka Mitchco), I decided that it was time to re-do the room measurements and see if I can incorporate the important suggestions - 48kHz sampling rate for the measurements, sweep range of 10-24kHz, and incorporation of the subsonic filtering at 15Hz. If you're wondering, the dimensions of the room are ~20' x 15' x 8' with a slight frontal tapering as shown; not a huge room by any means and not ideal dimensions but at least it's not squarish/cubish (here's a study on room dimensions and acoustics). It's in the basement of my house, built as a family media room with extra thick walls to reduce sound leakage when the kids are asleep at night (I pretty well can run the sub at reference levels with kids asleep upstairs at night so long as I close the door to the basement). The main speakers - Paradigm Signature S8v3's - are 7.5' apart up front, and the sweet spot is approximately 9' from the speakers arranged like an isosceles triangle. I make sure there is decent amount of space between the couch and the rear LP storage (>3'). [As you can see, I have the ~15-year old rear full-range Paradigm Studio 80v3's back there.]

Before getting to any measurements, remember we're playing the "long game" here :-) as in we want to protect hearing for a lifetime. Test tones are not euphonic and you'll likely be going through multiple trials with tones playing at reference levels of ~80dB. My set of ear protectors:
Customized ear plugs... Tools of the audiophile "trade" :-).
I didn't mention this with my previous post, but remember awhile back when I showed measurements using the Behringer DEQ2496, I also have the Paradigm PBK-1 kit which allows me to program the DSP inside the Paradigm Signature SUB 1 subwoofer to better integrate into the room.

I agree with Mitch when he suggests clearing up the room and removing sonic obstructions (to make it more reflection-free) to get a better room measurement. Even though putting the couch back and having the coffee table might be necessities in the room when I'm listening, I suspect the brain has some ability to anticipate some of the sonic changes. Plus if I want "the best" sound from my room, I would be removing the coffee table anyways, so might as well measure the room acoustics optimally. Here's the measurement setup with furniture out of the way as best I could as a one-man job, speakers angled slightly toed-in with the approximate equilateral triangle shape and the measurement microphone at the apex over where I'd sit for the "sweet spot" at ear level. Notice the Paradigm-supplied USB calibrated microphone in place:
Although I know some people measure with the microphone positioned vertically, I actually found the direct horizontal "on-axis" orientation produced more consistent results for me both with the Paradigm and Behringer ECM8000 mic used later on.

Paradigm Perfect Bass Kit setting for the subwoofer.
All this does is "shape" the frequency response of the sub to something more consistent with a flat response. Notice the null at 60Hz which was not completely corrected by the software. We'll keep an eye on this!

After this, I now do the usual subwoofer integration step of finding good settings for amplitude, crossover, and phase of the sub. This is also when I typically will move the sub around and maybe try some placement changes to the listening position keeping in mind esthetics and practicality. I will refer you to this excellent write-up by Avanti1960 on the Steve Hoffman Music Forums using pink noise to find reasonable settings. I used the freely available WaveSpectra software to get the job done using my Behringer ECM8000 mic with the E-MU 0404USB interface as ADC. (I bought the ECM8000 with calibration file years ago, these days I'd probably just get the inexpensive Dayton EMM-6 with downloadable calibration file using the serial number off their website.)


The main difference of course is the added low end down to 20Hz here with the sub and tweaking allowed me to fill in the dip from 40-50Hz a little better while fine tuning the phase knob. Again, we see that unfortunate 60Hz null in the FFT which I wasn't able to flatten out fully with the current sub position and seating spot.

Okay, it is now time to "get it going on" with AudioVero (((acourate)))! Here's the room with the calibrated Behringer ECM8000 microphone in place before starting my room measurements:
Behringer mic upright... As I mentioned above, I actually found the on-axis positioning pointing straight ahead provided more consistent results. I did make measurements with both orientations.
I'll refer you to the detailed step-by-step post late last year for the rest of the procedure... Like I said above, my intent is to incorporate some of the suggestions as an update to my room correction filter setting to create a more "accurate" listening experience. Suffice it then to just show the results of this most recent round of measurements:
48kHz sampling rate, 10-24kHz range, peak recording level -7dB, IACC10 ~74% (pretty good for home environment).
Target curve creation - flat to 1kHz then -6.4dB to 24kHz which equates to about -6dB at 20kHz.
Filter creation. Note subsonic filtering below 15Hz. A little trial and error was used to set the "Excessphase window" parameters.
Test convolution. Some improvement in IACC.
Uncorrected frequency response.
Corrected frequency response. Note that the Y-axis scale is the same 25dB range as the one above! And I'm not smoothing over the peaks and troughs too much either.
Clearly the differences between the uncorrected and corrected versions are marked in the frequency domain. Notice though that the 60Hz dip is still there but less severe overall than prior to correction. The left channel is obviously more problematic (about -6dB left, -3dB on the right compared to something like -15dB without the correction filter). Remember, ultimately we can't beat physics; I'm going to have to look into further speaker/listening positioning adjustments or invest in physical bass traps to do something more about that dip.

But what's more remarkable with good DSP room correction are the improvements in the time domain:

Speaker step response over 5ms: Blue/Brown = Uncorrected, Red/Green = Corrected with filter. Notice the 3 drive units are now in lock-step to create that corrected plot. Both right and left channels are well aligned in time and amplitude. Remember that such esteemed companies as Vandersteen and Quad place an emphasis on this time-domain alignment (see recent Vandersteen Model Seven Mk II measurements).
Speaker impulse response uncorrected. (Nice time alignment of speakers even before the correction.)
Speaker impulse response corrected - precise bilateral minimum phase response now!
All with minimal phase deviation throughout the audible frequency.
I have spoken of room reverb time in the past. Acourate has a very nice RT calculator screen to analyze this... Top right you see settings for tolerance levels using various standards like the DIN 18041 for live musical performances, and EBU 3276 for formal listening rooms like studios. As you can see from the tolerance levels (black), my room is a bit reverberant for studio purposes but would easily satisfy DIN 18041.
I won't bore you with the details around verification of the filter with REW to prove to myself that the Acourate filter works (again, this was covered in the previous post).

As I described before, the music with JRiver convolution DSP using the Acourate filter sounds excellent. Clearly improved frequency response resulting in natural sounding tonality. And with the improved time domain speaker step response and precise impulse response, soundstage and imaging are superb... Lately I've been enjoying Johnny Frigo's well recorded albums Debut of a Legend (1994) and Live from Studio A in New York City (with Bucky and John Pizzarelli, 1989 Chesky - SACD version sounds fantastic) - great violin work and fantastic sound staging highlighted with the DSP activated. The clocks and chimes at the start of the audiophile favourite "Time" (Dark Side Of The Moon) should sound "surround" and simply palpable. Another clear effect to listen for is how the speakers just "disappear"; for example a modern pop recording like Ed Sheeran's "Photograph" from X (2014) has his vocals literally front-and-center sounding like he's sitting with his guitar right between the speakers (although the added reverb is obvious and points to the studio work). It's just one of those things you're going to have to listen for yourself; preferably having the option to turn the effect off and on quickly to experience the difference time and frequency domain corrections make. This is what more accurate sounds like, when the drivers of a speaker system are temporally aligned and consideration given for room effects.

Remember though that there is a price to pay for this level of accuracy. You need computing power for real-time processing (not a big deal these days), and there is some timing latency to be aware of - gapless playback is a problem currently when streaming through DLNA/UPnP from JRiver 21 like what I have been doing with the ODROID-C2/Volumio (direct JRiver playback seems to be fine as the program can compensate, JRiver is aware of this and hopefully can come out with a solution soon). Also, remember that the higher the quality of the speakers / hi-fi system / room, the better the ability to fix frequency and time-domain issues... You obviously cannot expect a DSP to make a boombox sound like a true hi-fi system! It could try getting a little closer though :-).

Before ending off this illustration of my activities, here's a "flatter" target frequency response curve for listening when I'm in the mood (slight roll off from 10kHz down to -3dB by 20kHz) and the resultant speaker frequency response overlaid:


Basically within seconds, just by changing the DSP setting in JRiver, I can change the sound of the speakers; like getting new boxes in the room with basically just as good time domain qualities but with different tonality depending on what I choose to listen to that day... Isn't technological progress wonderful?!

BTW, I really like well recorded, "natural" sounding albums reproduced with this "flat" setting. For example the Hugh Masekela album Hope (1994, recorded live at Blues Alley in Washington DC, look for the Analogue Productions SACD) has an enveloping soundstage around the listener, with a nice tight bass that's just great with a full range speaker system. And for lovers of classical Chinese music with the erhu, Yu Hong Mei's (于红梅Erhu Chant (闲居吟) from Channel Classics sounds fantastic; tremendously expressive and emotional. Again if you can find the SACD, it's a great "true" high-resolution recording with a multi-channel layer as well. Of course, SACD/DSD material needs to be converted to PCM in order for complex DSP processing. I'm very happy with the result of DSD to PCM 24/88 conversions (tests last year here, and here).


Chapter II: Meditations on "Accuracy"

Above, I spoke about creating a "more accurate" listening experience. Is the concept of "accuracy" meaningless (a mirage? a delusion?)? Is it ever possible to describe such a state?

No delusion... And of course it is possible! Because of one simple fact... Music production studios aspire to "achieve optimum acoustic properties (acoustic isolation or diffusion or absorption of reflected sound that could otherwise interfere with the sound heard by the listener)" (as per Wiki). Whether it's a top notch commercial studio like Abbey Road, or Columbia Studio, or smaller audiophile labels like 2L, or AIX, or the home recording guy who still wants to up his game in the home studio; there is a certain standard of acoustic control and music production that needs to be understood and verified in making sure that the best recording can be made to the intent of the artist. I'm sure audiophiles appreciate that doing this costs money and professionals who do this work need specialized training and must appreciate the science behind the whole endeavor. Just look at the Studio Creations website for some of the factors from architectural design to statutory regulations to air conditioning that must be accounted for to create a desired, "accurate" recording studio. I also expect that an audiophile would accept that measurement devices would be utilized professionally in setting up the optimal acoustic space and in choosing and tuning the best speakers, microphones, control console, ADC/DAC, etc. for the job! It would be preposterous for studios to be only subjectively "tuned by ear".

When I speak of "accuracy", I am referring to an appreciation of the above. I am not talking about some kind of idealistic reproduction of "live music" though it would of course be nice if recordings were made to sound like this when reproduced accurately. Or some vague subjective notion of "sounding good" and "making the listener happy" though of course again, I hope artists and sound engineers are successful in creating such a product. As "music lovers", we all wish for a powerful emotional response as a higher-level goal of the art form in general. Appreciating subjectively "good sound" and feeling "happy" with the music isn't difficult and doesn't require fancy expensive gear; I'm sure we've all been doing it since childhood through AM radios, audio consoles, cheap boomboxes, poor quality earbuds, scratchy damaged vinyl... The "music lover hobby" is massive, inclusive, consists of all races and creeds, men and women. I will pick up the occasional Rolling Stone, browse the newest MOJO or read Pitchfork still for the "music lover hobby" which includes broad appreciation of the artists and exploration of the artistry of the lyrics, music and rhythm.

But when I write this blog or make the rounds to see what's new in the audiophile press, the goal of accurate music reproduction through the hardware is inextricably linked with the "hardware audiophile hobby" in my mind. Any one of us can take out the credit card and pay big money to buy expensive DACs and speakers if the desire and credit is available. But to me that's neither exciting nor educational in the service of a hobby that hopefully has an enduring trajectory. As a hobby, my perspective and hope is that "hardware audiophilia" satisfies an internal curiosity not only of what to buy or what to do, but also why there is value in the technology or equipment. In doing so, hopefully this promotes growth in understanding what makes a good sound system, anticipates the path of good sound reproduction as technology evolves, and with understanding and wisdom, reaching out to others to provide useful advice having understood the principles and experienced what makes a difference. In sum, I believe there are two separate hobbies that become intertwined for any one of us. You broadly see subgroups out there - those who mainly love the music but care less about having the best gear, those who mainly love the hardware but spend relatively little on the music collection, and of course those who try for a balance between the music collection plus the best hardware and room set-up with the hope for well mastered albums and accurate gear that allows for extraction of as much detail as possible for the listener. I would of course like to see myself in that last group.

Consider MQA as a real-life contemporary example of how "accuracy" is part of the evolution of the audiophile world. Whether one is skeptical about the claims of MQA, realize that a core claim by Meridian is that it incorporates a "de-blurring" algorithm to improve time domain accuracy of the recording from the studio to the home DAC (and "authenticated" as such). In my opinion, if the audiophile world embraces MQA as an exciting new technology that makes audio "better", then it must embrace the concept of "accuracy" as a desirable goal! No matter whether MQA achieves its "de-blur" effect as intended, the excitement is over the potential for a higher fidelity, more "accurate" rendering of the musical signal (with the touted ability to compressed file size for streaming as well). In my opinion, DSP filters like Acourate (and other similar software like MathAudio, Dirac, DRC/Designer, Audiolens...) likely will have significantly more impact because it actually will analyze and customize the sound with your speakers and in your room. Again, if we actually accept MQA's claims of time-domain accuracy, the promise of MQA would only truly be realized on a system with time-domain correction on the analogue side - the part where the sound actually hits the ears at the listening sweet spot! What's the point of the MQA technology aligning microscopic digital details like pre-ringing and minimizing "temporal blur" when multiple millisecond time-domain anomalies are not corrected due to speaker and room limitations? This reminds us that maintenance of accuracy through the whole playback chain is an essential component of what the "hardware audiophile hobby" strives for. It is also what truly beneficial new technology seeks to improve on.

[Needless to say, I noted that the "music lover hobby" attracts broadly including men and women in large numbers. Notice the "hardware audiophile hobby" is primarily the domain of men. This is an important distinction to recognize and I think accept; probably due to inherent differences between the sexes which is unlikely to change.]


Chapter III: "Accuracy" and the Media

If we go back to the foundations of the hardware side of our hobby (for example, reading "The History of High End Audio"), it's clear that from the beginning, the goal was "high fidelity". Whether this meant a low-distortion McIntosh amplifier that drove difficult speaker loads, noise suppression of clicks, pops, rumble of early turntables and vinyl, achieving low noise in a preamplifier, or speakers capable of full frequency response. There were legions of music lovers throughout time but those that really cared about high quality artificial reproduction of recorded music began with engineers and hobbyists who wanted fidelity. They were the pioneers...

I take issue with Steve Guttenberg's recent Stereophile editorial "The Fallacy of Accuracy". It seems like he provocatively titled the article to raise a few eyebrows doesn't it? And then there's his pithy conclusion of:
"My take: Park your faith in accuracy and measurements on the back burner, and go for speakers that make a big chunk of your collection of recordings - the good, the bad, and the ugly sounding - shine."
You see, a big reason that we can perhaps if we choose "park (our) faith in accuracy and measurements on the back burner" is because in the span of history, so many generations of gear have come and gone that on average, engineering is at a high level already. The quality of sound we achieve these days with almost any reputable speaker is already objectively very good! In my opinion it is perverse for extreme "subjectivists" to dissuade or detract from objectivism and measurements because it is objectively good engineering that allows us all to enjoy what we have today. Nobody says that the sound must be exactly the same for everyone. There is such a thing as subjective preference just like which target curve I choose with room correction. But the basic foundation of performance can be quantified to a reasonable range and factors such as the time-domain accuracy can most likely be "tightened up" in any particular system. Let's not be romantic though and think that somehow speaker designers like Andrew Jones and Paul Barton don't use objective measurements or scientific understanding to assess accuracy to do the "magic" they perform. Seriously, does anyone actually believe that Andrew Jones didn't bother to measure the frequency response of his superb TAD Evolution One? Or that Paul Barton didn't have a multi-decade relationship with speaker measurements done at Canada's National Research Council (NRC) labs? Check out this choice quote about Paul Barton and measurements from 2007 in the preceding link:
"The lab facility is very much Paul's home away from home. He has spent far longer in the chamber and lab than any other speaker designer, and in the 25-year process since he began going to the NRC has gained enormous sophistication about acoustics and the sound of speakers."
Right. I'm not sure in what way there is a "fallacy" when it comes to "accuracy" and objective measurements other than Mr. Guttenberg reading a little too much into a small portion of the thought process with choice quotes from respected audio designers... Like I said, the article seems to be nothing more than just trying to be provocative and "selling" a relativistic perspective as if opinion rather than fact is all that matters in this hobby. An apparent ridiculous attempt to dilute the essence of actual science in these engineered products. Let's also not forget that there's fantastic work being done by folks like Floyd Toole (great book Sound Reproduction: Loudspeakers and Rooms) and further articulated by Sean Olive; both having worked with the NRC in speaker design. Yes, folks... Within the universe of possibilities evoked by the word "magic", there still is scientific truth and empirical objectivity worth studying and aspiring to even if the last bit of fine tuning for the design of the speaker requires the ears of an "artistic" engineer (and visual esthetic of a good industrial designer).

And as you can see in the post today and in previous instances, when an article focuses on speakers (as Guttenberg does), the idiosyncratic interactions between speaker and room is so massive that of course there should be a plethora of speakers with variations in frequency responses, power handling, time domain parameters that could "work better" in any given situation. It's no surprise if a "good sounding" speaker fails when matched to inappropriate gear or in a room that doesn't synergize.

As I noted in the opening paragraph of this section, "high fidelity" began with a purpose. To be faithful (or "accurate") in the reproduction of what is on the "record" be it vinyl, CD, SACD, DVD-A, Blu-Ray or digital download. How we exactly define adequate "high fidelity" is something we can spend volumes and many blog posts on. It depends on the person (ie. we all have varying hearing abilities and preferences), the situation (eg. my desktop computer speakers don't have to be the same fidelity/accuracy as the speaker in the main sound room), and the purpose (eg. the Sennheiser HD800 would be terrible headphones for jogging, Bluetooth speakers perfect while camping). But a science based engineering approach allows us to understand how the equipment works, and know when not to waste time/money on the frivolous. Objective knowledge does not take away from choice but can help inform better decision making.

Ultimately, although most of us are no longer engineers trying out new designs or putting together DIY kits, we can still learn to understand for ourselves and find satisfaction in so doing. Not just what we like on the software consumption side (the "music lover hobby"), but as educated consumers with insight into worthwhile hardware (the "hardware audiophile hobby"). Sure, we can keep an open mind, but let's face it, in the audiophile world of 2016, it's obvious that not everything is worthwhile. As the old saying goes, keep an open mind, "but not so open that your brains fall out" (Walter Kotschnig, 1940).

Chapter IV: Conclusion

As I close off today, I want to suggest something which I feel strongly about having explored and thought about audiophilia over the decades. The concept of "accuracy" isn't a peripheral consideration. It is in fact one of the pillars of this hobby from my perspective. Without some internal definition of what "accuracy" means, the "hardware audiophile hobby" is without a compass unifying hobbyists. It becomes a consumeristic affair where the goal becomes that of moving product whether there are technological merits or not (eg. expensive cables, unproven USB doohickeys...). There does not appear to be a willingness or ability for many mainstream audiophile pundits to weed out bizarre claims and call out functionally anomalous measurements for what they are. So often, the subjective-only reviewer will wax poetic about some piece of equipment, yet the performance suggested by objective analysis are clearly such that no reputable studio or acoustician would likely endorse such products. It's in fact rather bizarre the dissociation between the ideas and values held by professional music engineers and the audiophile world around sound quality and what is "good" equipment. Even more embarrassing are the times when writers accept claims for products that obviously cannot work as advertised simply based on common-sense and basic education (just consider $$$ cables!).

Finally, I hope that in these blog pages and in articles like this, the audiophile reader finds it empowering to consider that we are not simply consumers. Knowledge, curiosity, and ultimately experience allows us to grow beyond the ubiquitous hype and hyperbole of manufacturers, or what some magazine writer, forum poster, or blogs might claim. Of course I do not exclude myself from this potential as well; to err and have subjective biases is human... But what is important is to have tried to verify impressions and understood truths where they may be found. Exploring ways of optimizing what you have whether it's $200 worth of gear or $200+K is a great start.

------------------------------------

Have a great week and hope you're all enjoying the music; with unabashedly the most accurate sound you can optimize from your system as you see fit.

Further Reading: Over the last couple of months, I have had the great pleasure of reviewing and providing suggestions to Mitch Barnett (aka Mitchco) towards an eBook he has been writing about the use of (((Acourate))). As you probably know, Mitch's blog posts on Computer Audiophile including the ones on the use of Acourate have raised the "standard" when it comes to introducing many of us to DSP-based speaker/room optimization (and was the inspiration for my own post last year). In this book, Mitch goes into much greater detail drawing from his experiences in professional sound recording, discusses technical aspects of sound quality (and "accuracy" of course!), explores advanced techniques in multiway crossover digital filter design (including driver linearization and time alignment) as well as measurement techniques like beamforming to achieve quasi-anechoic results. Topics beyond the already excellent free articles. He also has a detailed chapter on acoustic measurements and ways to validate the filter performance. All written in a practical and no-nonsense step-by-step presentation aimed at getting results. I don't think it's hyperbole to say that this is the essential reference/manual for an extremely powerful piece of software which can be overwhelming for the uninitiated. At the cost of less than 3 caffè lattes at Starbucks ($9.99), I think this is a very reasonable price for the amount of work involved. You can check out his eBook on Amazon.

16 comments:

  1. Absolutely wonderful article. Touches so many subjects i've been trying to get thru: that in order to compare two sounds, the switch needs to be almost instant. The importance and magnitude of room, matching speaker to the room, using measurements to do so and not equipment lottery. Especially the need to KNOW what the system does in order to improve it, that measurements do not take away from anything, you don't have to do what they say as the final sound is still personal.

    Possibly the only thing i would add is the importance of having a neutral reference even if the end goal, "My Sound" is not flat, to design so that one can always switch between the two fast. Since it is the "flat configuration" that is going to be used when doing room corrections, acoustic treatment, measuring new speakers etc. Also that if you put 30k on your system and do not own a measurement mic it's like having a Ferrari without dashboard, you need to feel the speed and guess how much fuel you have left..

    ReplyDelete
    Replies
    1. Thanks for the note Squid. Good idea to always have a "flat" reference on hand.

      Nice analogy - $30K into a system with no measurement mic would certainly be unfortunate!

      Delete
  2. I absoluted detested that editorial by Guttenberg.

    My desire is to replicate the sound of classical chamber music (or live jazz) as best I can given my financial constraints.

    To claim that there is no benchmark I can use for that replication is crazy. I just go to the concert hall to refresh my memory about what live music sounds like.

    ReplyDelete
    Replies
    1. Correct.

      For music intended to sound like live jazz, chamber music, classical orchestra, absolutely! Go to the concert hall and compare that to well recorded acoustic music on your system. So long as one knows that the recording is supposed to replicate a natural "space" then by all means make the comparison!

      Delete
  3. Regarding the null at 60 Hz, you could try using 2 subwoofers to try and null it (sorry couldn't resist). There are various configurations like the Welti 2 sub setup (rear sub out of phase) to more subwoofers along the lines what Geddes recommends. There is a good thread at Audio Science Review forums discussing it.

    About the Dark Side of the Moon when you are talking about the surround effect on Time, is this using the SACD or blu-ray 5.1/4.0 respectively or is it stereo playback? If it's the latter I would be interested to know what it correlates to in the measurements that allow such effects to happen. Excellent time/phase response, controlled directivity, etc?

    ReplyDelete
    Replies
    1. Hi Modal,

      Yeah, I'll look into a second sub :-). I suppose I could stick it in the rear and although the SUB 1 isn't huge, it would be noticeable...

      The DSOTM I'm talking about is stereo playback. I haven't even begun doing 5.1 room correction filters! I enjoy surround sound but let's face it, the majority of music I have is still stereo. I've been bugging Mitch to do some work on 5.1 :-).

      What I can say is that the time/phase accuracy adds significantly to the sense of "space". I guess it's like what MQA purports to be doing with their claim of "deblurring". The clock/chimes at the beginning of "Time" sounds like they occupy a well circumscribed space not just in front but also to the side and rear when sitting in the sweet spot. It sounds good before the correction but the sense of reality "jumps up a notch" and sounds more natural because of the frequency response as well.

      Delete
    2. Cool, something for me to look into. Do you think you'll attempt Acourate correction with the surround setup as well?

      Delete
  4. I love what you write and your mission, but I don't really endorse the use of convolution filters in the manner you are doing.

    By all means use it up to about 250-300 Hz, but above that you are causing problems. To illustrate this, assuming you can drive your drivers individually, take close-up, on-axis (between driver and listener) gated FR measurements of your midrange and high range drivers while running through your correction filter. You will find that their direct output is not flat in terms of frequency response. Your target should be to get them flat, according to what I have learned about targets and psychoacoustics. This is a more sophisticated approach than the target curve design you described on 27 Nov 2015. (I *really wish* that the approach you used was adequate, because it would make the correction process so clean and simple, and limit the uncertainty to 'which curve is the best curve for me'.)

    all the best,

    ReplyDelete
    Replies
    1. Targs - I seem to be under the influence of other ideas about targets and psychoacoustics. But, I do not think a "flat" target or measured response is at all desirable above the bass transition frequency. What is almost universally recommended is a smoothly downward sloping response with increasing frequency measured near the seating position. That is what is important for listeners.

      Up close, near field measurements of individual driver responses would only be of interest to speaker designers.

      Delete
    2. Thanks for the note Tnargs. Of course there can be different opinions on what should be done such as whether we should be trying to fix frequency responses above 300Hz...

      I just have to note that Acourate does a great job and the basic technique works for me. Also have a look at this AES presentation (thanks for the link Mitch):
      http://www.aes-media.org/sections/pnw/pnwrecaps/2008/jj_jan08/

      Use frequency dependent windowing to correct direct arrival of the high frequencies for spatial cues, equalize out large peaks in the lows, and try not to overcorrect deep nulls. That's basically what I'm trying to do with Acourate here...

      Delete
    3. Hi guys,
      @Fitzcaraldo, you will find that the 'universally recommended' downward sloping targets are for ungated or 'summed sound' measurements. My flat target is for the direct sound only.

      @Archimago, I didn't realise you were doing frequency dependent windowing. That's good to know. But it only means the B&W curve (also the 'Mitch/Katz' curve) is less appropriate as a target than ever. Thanks for the link to the JJ presentation: I used one diagram from it in my presentation to the AES Adelaide Section on the same topic in 2012. I could share or discuss it if you approach me via email. cheers

      Delete
  5. I am currently reading into Sonarworks Reference 3 HD, which should be more easy to use. I was told by support, the HD version can measure speakers up to 4m apart accurately and also is suited to include multiple subwoofers. Nevertheless, I had no chance to try it out so far, but I see they are making measurements on many(!) spots around a listening position, to enhance accuracy in a bigger area.

    Can this be done with Acourate, as well, or do you always measure only one spot? I guess it would be beneficial, e.g., when listening with friends on a couch like yours to have an average out of multiple positions.

    Regarding your 60Hz dip: will you smoothen it via bass traps and a second subwoofer; or only one of these possibilities?

    Regards

    ReplyDelete
    Replies
    1. Hi Balduin,
      With Acourate, it's a single spot measurement and I know Mitch has demonstrated that despite this, the quality of sound around the central sweet-spot remains quite high.

      Thanks for bringing up Sonarworks! I only read about it recently and it looks like quite a capable piece of software as well. I'll have to give it a try with the headphones especially...

      Delete
    2. Thanks, this would also have been my question. :)

      Regarding headphones, Sonarworks do a good job. Of course, I can only subjectively judge, since I do not have the means to measure the result - but my HE-500 (with FocusPads) sound much more realistic after their calibration. They provide you with "Perceived Acoustic Power Frequency Response" + THD at 80 and 90 dB graphs of your headphones. Correction on their plugin follows the B&K 1974 curve; only slight adaptations are possible but unfortunately not to your target.

      See link, for what it does on my headphones: http://up.picr.de/25775711bz.png

      Btw, regarding headphones, there is some interesting stuff going on at InnerFidelity: http://www.innerfidelity.com/content/first-peak-head-measurements-harman

      Best regards!

      Delete
  6. This comment has been removed by the author.

    ReplyDelete