Understanding the State of the Art of Digital Room Correction

[mitchco]

that's the whole point ; once again, this quote from Dr Toole : "The stated or implied sales pitch is: give me any loudspeaker in any room and my process will make it "perfect". A moment of thought tells you that this cannot be true." Attached are my in room L & R responses above transition without any eQ : why the heck would I let an algorithm mess above the transition in the first place? And that's where I've lost precious time thanks to @mitchco and his (IMHO purposely) misguidance via promotion of full range targets. If you limit eQ below transition you don't mess with FDW, phase and all the sophistication of his so called "SOTA DSP" ; if you think those are good for you then you have a poor system to start with. I'll be very happy to support @mitchco if his claim was simple and honest, something like : for the price of a fancy cable I can make your limited/compromised system sound better than with a change of fancy cable. That I'm sure he can do, but SOTA is much too much hype then and with my SOTA active speakers it was a wrong way.

Dear cucumber, you are the guy that has been repeatedly banned on another forum for harassing me. You were banned again for the 6th time just last week, lol. Now you are here repeating the same nonsense. Goodbye @Le Concombre

 

[tecnogadget]

I consider myself an objectivist above all else, but I think this particular thread is an excellent example of over-objectivity embroidering toxic scepticism. I am frankly surprised, dumbfounded, and disappointed by the lack of welcome and support for this thread, but what strikes me the most is the ungratefulness I've found.

I thought this kind of information would be welcome and appreciated at this site, especially in the face of the uncountable mentions in almost any long thread about how DSP is the future of sound and almost mandatory for reference reproduction.

The way you all have been screaming your heads off about the debate between single or multi-spot metering, or raw vs averaging, you remind me of the inquisition and witch hunts or poor Galileo Galilei. The problem is that being able to define everything in a single video is a difficult thing to do, and perhaps not too much emphasis has been placed on it, but it seems that everyone has overlooked the concept of "envelop". It is not about simply smoothing the measurement response but about "enveloping" in a clever way through FDW and getting a smoothed response that sits on top of the comb filter region and avoids overcorrection of the dips. This makes all the sense in the world because you shouldn't boost on very steep dips, it would be a way to filter out the "noise" (what we are not interested in) from the measurement, in the end, what the OP proposes is to work with this new frequency response which can be considered as a higher envelope of the original magnitudes. Using As an example, Acourate calculates a psychoacoustic frequency response that takes into account the transient behaviour of music signals.

And finally, the OP never closed his mind to a single measurement, he commented that according to his analysis Focus Fidelity did a good job for multiple measurements. The OP simply wanted to share the power of single measurement by empirical real-life verification over a wide area, is that so difficult to digest?

Like most I came here and wanted to learn about the "state of the art" in room correction. I know what I know – although some think it is not a lot So I ask questions. If in return there's just hand-waving because "it's all very complex and complicated" and "proprietary" (aka a secret) then I get a little frustrated.

Really? The truth is that you and many others have been going against the OP from the start, it was not a friendly exchange of views or questions. Honestly, I am surprised if he finds the will or energy to answer you

that's the whole point ; once again, this quote from Dr Toole : "The stated or implied sales pitch is: give me any loudspeaker in any room and my process will make it "perfect". A moment of thought tells you that this cannot be true." Attached are my in room L & R responses above transition without any eQ : why the heck would I let an algorithm mess above the transition in the first place? And that's where I've lost precious time thanks to @mitchco and his (IMHO purposely) misguidance via promotion of full range targets. If you limit eQ below transition you don't mess with FDW, phase and all the sophistication of his so called "SOTA DSP" ; if you think those are good for you then you have a poor system to start with. I'll be very happy to support @mitchco if his claim was simple and honest, something like : for the price of a fancy cable I can make your limited/compromised system sound better than with a change of fancy cable. That I'm sure he can do, but SOTA is much too much hype then and with my SOTA active speakers it was a wrong way.

I think you are being harsh and unfair to the OP. Based on this and your previous posts you are calling him a scammer to his face. The OP indeed offers a professional service, but the amount of information he has shared for free either in this video, his multiple articles on different forums, and helpful replies, are more than enough for many people to experience this DSP approach on their own, and therefore I think he has earned some respect. And by the way...your In-Room response looks quite good above transition...but you still have a difference of 6dB in the response between 2kHz and 10kHz...that can clearly be improved by DSP...

mitchco Thank you very much for taking the time to make this video, and to continue to support the questions of those who are interested in "State of the Art of Digital Room Correction", I hope we can continue to count on your active participation. It is much easier to destroy than to build, I applaud you for the latter.

Kind regards, Stefano.

 

[mitchco]

@tecnogadget Thank you Stefano!

 

[JP]

I thought this kind of information would be welcome and appreciated at this site, especially in the face of the uncountable mentions in almost any long thread about how DSP is the future of sound and almost mandatory for reference reproduction.

For every person posting there are dozens following along, and a lot of those have constructive questions but haven’t formulated their thoughts enough to ask. Yet.

And I think many plainly understand that when a @j_j answers a question with a question, they’re pointing you to the answer while suggesting you do some homework.

Hopefully the people pushing disingenuous conversation don’t spoil it for all.

 

[gnarly]

Lot's of good discussion in this thread...thanks everybody.
And of course thank you Mitch for the opening post/presentation. Has a well laid out logical sequence, is quite instructive, and with supporting links....nice work

I have not yet tried any form of automated room correction, so seeing examples of the Acourate and Audiolense workflow has been very helpful.

I do have substantial experience with FIR, having used it to tune dozens of DIY speaker builds.
The tunings have been for speaker-only with no consideration of what room they might go in; outdoor quasi-anechoic measurements, spinorama type, on and off axis optimizations.
Most of the builds have been 3 or 4-ways for use with a sub, and have all used a FIR-tuned amplification channel for each 'way'. So lot's of FIR files to say the least....
Usually linear phase tunings, but sometimes IIR when latency doesn't work (live sound applications).

It's been interesting to contrast the "digital room correction" procedure presented, to the "digital speaker correction" procedure I've come to use.
Without a room to consider, there's been no need to to divide the frequency spectrum into the Normal Modes, Diffusion, and Absorption ranges that each use varying DSP approaches. (Lot's for me to learn here)
I simply work to get mag and phase as smooth and flat throughout the spectrum as possible; with smooth directivity curves.

There is also contrast as to what degree of correction to apply in terms of the smoothing level to correct (at least in my mind
My approach has been to correct any magnitude ripples that respond well to minimum phase correction, and that hold up both on and off axis. Very clean builds that show relatively few multiple-arrival paths to the mic, often hold up well even to corrections made with just 1/24th smoothing. Whereas builds that show more multiple-arrival paths, only hold up to say 1/3 to 1/6th smoothing.

Whether to tune to a single measurement, or multiple measurements, is an open issue outdoors for me still and has depended build by build. Although obviously not nearly the size of the issue indoors

I call this SOMA...(State of MY Art)...........a term i wished everyone would use in place of SOTA, (State of THE Art),
Hearing SOTA always puts me on get-wary alert. (My only humble criticism with your preso, Mitch.)

Anyway, I can totally understand the back and forth in this thread about how much correction, and how to go about it, once inside a room.

My approach to date has been to build the best speaker i can quasi-anechoically.
Then, if i don't like the way it sounds in a room; first take care of the room's RT60 curve, and then play with speaker placement along with direct reflection/diffusion treatments based on placement.
Only EQ use has been to curb sub modes, or make a down sloping house-curve preference (with i do with global shelving on the input signal).
So an old school approach other than the (mainly linear phase) FIR tuning.

Here's one puzzler i have about what's been shown with the room correction procedure.....
It seems it relies on a speaker passing a certain level of anechoic correctness, before beginning.

One of the points made several times was the reliance on Psych smoothing for DRC (which makes total sense to me, knowing the dangers of overcorrection to a spot.)
If i just used Pych smoothing for making outdoor tunings/correction, I'm leaving a lot of speaker goodness on the table.
Finer quasi-anechoic mag and phase performance is easily accomplished, before bringing it indoors. I can hear this outdoors, comparing differnt degrees of mag and phase correction.

What do you consider the signs that a speaker itself needs more correction before beginning your room integration process?

And a second question if i may...which DRC software would best accomplish both driver-by-driver, DIY speaker-builder corrections, along with subsequent room corrections? It appeared that may be Audiolense??

thx, mark

 

[BrokenEnglishGuy]

This speaks for itself!

View attachment 161562

Do I only need the UMIK 1 for this?
Or for the time alignment and these kind of stuff i need the mini dsp?
I'll use my PC as source for ever with my speakers, no problem

 

[AdamG]

This thread is being closely monitored and future Moderator actions are being considered. Keep this conversation Civil, respectful and constructive or you will be shown the door. We recognize and support Constructive criticism and debate/dialogue as the foundation of Science and advancement of shared knowledge. Be careful in the choice of your words and remember why we are here.

Please and thank you!

 

[markus]

Lol, I provided answers in the very post that you quoted but you chose to ignore. Just like you have been ignoring @j_j 's explanations. Repeatedly to the point of trolling. I believe the problem lies with you. Others seem to have no issue understanding my explanations:

Hi Mitch

Wow, this is great. Last week I was too busy to watch your video and therefore my delayed response. Now, I have watched it from start to finish and I must say, that you prepared this SOTA DRC extraordinary well. It is very thoughtful and this all the way through so I am enormous impressed by this work. You explanations are very clear, and all with technically foundation and without any voodoo. So this is really great. Thank you and all the best.

Juergen

Goodbye markus.

Now that I posted actual data we could work with on an objective level you want to stop the conversation?

 

[mitchco]

Thanks for posting your very informative video.

I’ve had Acourate for many years and agree with your assesment of the advantages of pc based DRC as opposed to hardware based DRC. However, a PC based system is not always family friendly so hardware based DRC, e.g. miniDSP running Dirac, still has its place.

A couple of questions re. the presentation:

1. When calculating the width of the 600 Hz window I see you took 4 Hz as the lower frequency. The reason is not clear. Can you explain why 4 Hz was picked, and not, say 1 Hz which has the calculation advantage that log1 = 0, or alternatively 15 Hz, which would be a more realistic lower cut off frequency for a system with subs? Of course, the change in window width would not really matter, but it would be good to understand the reason behind your choice of 4Hz as the lower frequency.

2. In you list of recommended software packages you did not mention Rephase which is freeware. I’ve not used it, but looking at the feature list and reading the long thread on diyaudio it seems to tick all the boxes. Can you comment?

Great work. Keep it up.

P.S:- Is an updated edition of you book in the works?

Cheers! To answer your questions:

1. The calculation uses 1 kHz as reference frequency for octave treatment. And 3.9 Hz is 8 octaves below 1 kHz.
2. rePhase - awesome tool I have used, but is not advertised as DRC. I get that folks use it in combo with REW, but if going the DIY route, I would recommend DRC-FIR.
3. Yes, but no ETA yet.

 

[JRS]

The only con of DRC I can see of is that it can be done incorrectly or sub-optimally. There are a lot of factors to consider...

Arguments revolve mainly around what is the best way to do it e.g. single-point, mutli-point mic measurements, regular or weighted averaging or no averaging at all, smoothing & windowing techniques or none, correction only in the modal region or up to full-range (probably verging into speaker correction), and other stuff.

Honestly, you don't have to have a separate PC or use Audiolense or Acourate since your AVR probably already can do some form of DRC. But it's not going to be nearly as customizable in that regard. Less of a headache, so could be an advantage also, if you will.

My experience leads me to believe that when we correct the drivers and the room response simultaneously, it is difficult to tease out the impact of each. DEQX, and IIRC Acourate (I believe DRC as well) follow a different path. The first round of msmts takes place close to the speaker at 1 to 2 meters, obtaining the usual time windowed impulse response, being careful to minimize early reflections.

I might add that to get a good S/N ratio this may take 2 minutes or more of swept log freq sine waves. Having obtained a S/N ratio in the 45 to 60dB range, one then sets about to correct the speaker--this involves flattening the FR, time aligning the drivers and correcting phase issues insofar as that is possible. Generally that correction is good to about 300-400 Hz. At this point, the microphone is moved to the listening position and the room + speaker response are measured one channel at a time, and peaks are eliminated. At least in the case of DEQX, the recommendation (though this may have changed as I own an older model) is to leave the deep dips alone and attempt to smooth the response by eliminating peaks and shallow troughs. So single msmt. I know several users though who take multiple msmts and simply average the results, before applying correction. In either event, one can obtain some success in smoothing the bass response, though the work of Harmon/JBL makes amply clear that having more subs would be useful as otherwise there are too many degrees of freedom.

I like this approach because it clearly separates the impact of modifying the IR of the speaker, vs that of the entire system. And the speaker filters are portable. I normally live in homes that have decent listening rooms as it is a high priority. Based on doing it separately, I would say that 70% or more of the improvement is due to fixing the speaker. Obviously, the better the speaker, the smaller the impact, and this is where room correction enters the picture, and again the lousier the room, the greater the impact. Edit: Like Gnarly@, I do DIY active, multiway speakers exclusively-and so XO design is a huge piece of the puzzle.

I get that newer algorithms can measure both the speaker and the room and know which is which. But I also tend to believe that early reflections might be confused with the speaker response, and that's where it would seem a system like Dirac LIve might have the advantage as one is dealing with a multiple variable system of equations that would seem to require multiple msmts to isolate the impact of room aspects. But that's just a mathematical hunch--I have no real experience to base this on, and am open minded enough to have decided to try AL once I get situated. I'm a single guy and usually it's just me or my gal and I so the window of excellence need only cover an area of two recliners, and not a bunch of guests--in which case, I see no recourse but to expand the sweet spot at the expense of ultimate quality in the middle. I've looked at some of the available literature--and there has been extensive research--even a consensus has yet to e reached about achieving the best average response. There are a variety of approaches: minimizing the min/max errors, measuring an arithmetic average, measuring a median average, minimizing RMS error from target curve, and other "fuzzier" approaches. Iteration is also a feature in some approaches if that counts as a multiple msmt.

Bottom line: I'm not sure a clear winner has emerged and may very well be situation dependent. It would be great if you could call up the guy who comes by, uses his software to come up with a filter set, that you can load in JRMC and be done with it--at least until next move. Seems like a very useful service to me.

Anyway my 2 cents worth.

 

[Alice of Old Vincennes]

audio DSP software development

Oops

 

[LumbermanSVO]

Given that none of the evangelists care to provide standard measurements to show what it actually does, the threshold to pique a reasonable person’s interest has not been reached. So why sho

So, I have a couple FIR processors and access to lots of different speakers. What exact tests could I perform that would satisfy you? Please, describe the testing process in detail, or link to them.

 

[Digital_Thor]

Tried to read everything... but no matter what I do..... 24 hours is apparently all we get each day

If someone really dislike windows - I kinda agree. But why not just use Linux? It has been perfected for years now and is beginning to really be as easy as windows to use. And a simple search on the net, gives me quite a few Linux options for Convolvers. If they work or not... well... you might learn how much you dislike windows, when you figure it out

Camilla DSP could be an alternative to windows APO:
https://github.com/HEnquist/camilladsp

Also found a good list here:
EQ Software for Windows, Linux, macOS, iOS/iPadOS and Android.

 

[Le Concombre]

Dear cucumber, you are the guy that has been repeatedly banned on another forum for harassing me. You were banned again for the 6th time just last week, lol. Now you are here repeating the same nonsense. Goodbye @Le Concombre

Why you and C Connaker couldn't simply accept debate over exclusion ? I'm back whenever I wish on AS with whatever pseudo/email and VPN : banning is real nonsense,lol. As the owner of computer audiophile CC was able to delete all (on his website, not in my archives) the nonsense he wrote about asking you (and you complying) for a flat 20 20K flat response but he just can't let go the blow he took while, as a good christian, I'm glad I could help him. Attached are the initial eQ you did for him and later mod : his Wilson remain expensive pieces of... with serious directivity issues but at least now he is less crazy about his flat in room response nonsense and your "calibration". Now, about your "goodbye" : I’m aware that deceiving on the internet by presenting biased information in order to fulfill a business agenda is current practice and I have been banned from AS for hurting business interests with Wilson Audio and Accurate ; I might be naive once again but I believe Armin and ASR are about honesty and thus doubt you can get me banned from here too. I doubt that on ASR there are business links not clearly declared to the website users and money streams not clearly declared to the IRS (in-kind donations of very expensive stuff under the form of permanent loans). And I doubt that guys here in ASR will find good enough your dubbing of my advices as "same nonsense" for Audiophile Style banned me : do yourself a favour and try contradict me with real relevant arguments

 

[AdamG]

Thread closed to further comment.