Understanding the State of the Art of Digital Room Correction

[dasdoing]

And, what smoothing is "correct" in order to create the "right" spectral envelope?

Denis uses ERB,

but imo no smoothing should be used at all. and FDW doesn't make sense either with this concept, unless you hear an echo

 

[markus]

Looks like "state of the art" means different things to different people

 

[dasdoing]

but imo no smoothing should be used at all

this obviously results in a correction valid for a few cms only though.
but you can flaten the envelope for multiple positions by overlaying their messurements (overlaying ≠ averaging)

Looks like "state of the art" means different things to different people

we only have theories afaics.
one thing that will never escape my mind is that the correct theory has to produce a flat graph for neutral. if it is not flat and sounds neutral-ish...the method of looking at it is wrong. JJ's theory was one way to target flat (flat direct), the flat envelope is another. all the others are just trail and error (house curves) that are aproximated to a neutral response in a given room which is not realy transfareble to other rooms

 

[dasdoing]

Measuring the (quasi) anechoic speaker response: This is virtually impossible even when using windowing as room boundaries, furniture (seat back!) and objects are too close to the microphone. The magnitude response is skewed and the resolution is coarse.

also FDW is not intended to produce (quasi) anechoic speaker response. it intends to produce the direct sound response. this is not the same, as there are (very close) reflections in the direct sound. it makes no sense at all to EQ the (quasi) anechoic speaker response, since we can't hear those very early reflections seperated from it.
At least this is how I understand JJ's concept

 

[markus]

also FDW is not intended to produce (quasi) anechoic speaker response. it intends to produce the direct sound response. this is not the same

Well, if you place objects and surfaces close to a speaker in an anechoic chamber it is the same (although FDW also reduces frequency resolution). But that's just a technicality. The real question is still "what constitutes direct sound" and how does it relate to perceptual qualities like (phantom) source location, timbre, spaciousness, etc. pp. A lot has been written but here we are still discussing these basic things...

 

[dasdoing]

what constitutes direct sound

I have been questioning this. And I don't know how Bruggeman arrived at 15 cycles as a default.
I am not intrested in this anymore since I adopted and am experimenting with spectral envelope,
but I imagine that you can somehow test it with 2 speakers, with one functioning as a aritifical reflection. than you send frequency filtered impulses through them with one of the speakers having a delay. if you percieve them as one sound the phantom center will be somewhere towards the earlier speaker. than you make the delay bigger. once you are at the boundery you will hear the speakers speratly; the phantom center colapses

 

[mitchco]

First of all thank you for taking the time to make this video and explaining your process with Acourate and other software which you – as I understand it – offer as a commercially available service. No excuse for people anymore to not know about the fundamentals.

Cheers mate!

As much as I tend to agree with JJ's "laws" there has been no formal verification, i.e. scientific studies. Such a study would need to include many different configurations, small rooms, large roooms, different reverberation times, different ratios of direct and reflected sound, reflections angles, timing and spectrum, stereo, multichannel, etc.pp.

I have worked on over 125 rooms that are indeed of all various sizes, shapes, reverb times, different loudspeakers, some with subs, some with multiple subs, some MCH, the whole gamut. I have system descriptions, rooms ratios, reverberation times, pics of the rooms, etc. While not a "formal" verification, given the amount of independent data and the results, I would say at this stage @j_j was right on the money. I hope to anonymize and share this data.

Latency with ultra long FIR filters: The signal has to pass the whole filter first before sound comes out at the end. This prohibits such filters for any application that requires (near) realtime processing like gaming or really any video streaming. One could build a video buffer to sync audio and video but at this point such a solution does NOT exist in consumer AV space (and the gaming problem remains).

I think this is about the 6th time that I have seen you make this statement One can simply turn the linear phase filter into a minimum phase filter and voila! That's what I do when I stream Netflix for example. While you lose out on the excess phase correction, you still benefit from a high resolution frequency correction.

Measuring the (quasi) anechoic speaker response: This is virtually impossible even when using windowing as room boundaries, furniture (seat back!) and objects are too close to the microphone. The magnitude response is skewed and the resolution is coarse.

You may want to revisit FDW math to show that it is not only possible, but is actually so. And as I have mentioned so many times in so many threads, no obstructions between mic and speakers. Move the chair or couch temporarily out of the way for measurements and place back once finished. We don't want to "bake in" any comb filtering into the correction filter.

Psychoacoustic filtering: This is something Uli Brüggemann introduced without providing any information how he arrived at it. It seems more like an educated guess that this is closer to what we hear but it is certainly not backed up by any scientific study (I know of).
"We don't hear dips (as much)": While I agree they greatly contribute to perceived overall timbre. Simply ignoring them (by visually filling them in) isn't probably helpful in that regard.

Psychoacoustic filtering is in each of the SOTA DRC software I mention. And based on the 125 rooms I have worked on, I can attest, with data that it works very well. Again, I am intending to share that. And I think there still appears to be some misunderstanding on psychoacoustic filtering as we are not simply ignoring the dips...

Audibility of pre-ringing: This is not a well researched topic either. Threshold is probably depending on frequency, signal and specific room reverberation time (masking effects).

You are making guesses with no data to support your guesses Preringing has been a mathematically solved problem for quite some time and preringing compensation, if required, is in each of the SOTA DRC products mentioned.

Single mic position: We have two ears, so how do other points around a central position look like? You're only showing what looks like heavily smoothed measurements. Did these points improve too? Or are they worse? What about multiple seat optimizations?

As covered in gory detail in my DSP book and in the video, I took 14 measurements around a 6' x 2' grid area and the results show a smooth frequency response across the area with the timing response intact, with one analysis measurement. And no, the measurements are not heavily smoothed.

@markus free to respond, but how about bringing some measurements or data to demonstrate your counterpoints.

 

[JRS]

Most FIR filters use a centred impulse alignment where the latency due to the filter alone is half the sample time of the overall filter as shown above. Linear phase filters have symmetrical impulse responses so this makes sense. An acausal filter could have the impulse placed anywhere within the filter but that would limit the amount of time correction to the number of samples of delay so is more unusual.

Every soundcard, computer system, DSP, software etc. uses buffers to prevent dropouts, even DAC's have propagation delays through them. It is these in combination with the latency of the filters that make up the final delay from pressing go to when audio comes out.

An IIR causal filter in itself should cause no delay as the impulse peak is at the start, but if the software or associated processing has buffers this might not end up being true in practice. IIR/causal filter encoded as an FIR only needs to incur 1 sample of delay if the convolver is written to make it so but many are not.

Parallel processing of the same type should not incur any extra delay unless it causes the software to need to increase it's buffer settings to keep up.

Thank you, exactly what I wished to know.

 

[markus]

Cheers mate!


I have worked on over 125 rooms that are indeed of all various sizes, shapes, reverb times, different loudspeakers, some with subs, some with multiple subs, some MCH, the whole gamut. I have system descriptions, rooms ratios, reverberation times, pics of the rooms, etc. While not a "formal" verification, given the amount of independent data and the results, I would say at this stage @j_j was right on the money. I hope to anonymize and share this data.

The more data the better but I'm not sure if that data is helpful as it is just "more of the same" without really answering the psychoacoustic questions I've raised.

I think this is about the 6th time that I have seen you make this statement One can simply turn the linear phase filter into a minimum phase filter and voila! That's what I do when I stream Netflix for example. While you lose out on the excess phase correction, you still benefit from a high resolution frequency correction.

If you do just minimum phase filters then people can use Audyssey (XT32). Much more user friendly.
Dirac even uses short FIR filters which do address excess phase issues.

You may want to revisit FDW math to show that it is not only possible, but is actually so. And as I have mentioned so many times in so many threads, no obstructions between mic and speakers. Move the chair or couch temporarily out of the way for measurements and place back once finished. We don't want to "bake in" any comb filtering into the correction filter.

My perspective is not the lonesome 2 channel stereo guy with a single chair in the middle of his "sanctuary" but the regular enthusiast that can't move furniture easily or he has a dedicated listening space with rows of seating that also can't be moved.
Furthermore I do think that room correction can be more than speaker correction so furniture needs to part of the equation.

Psychoacoustic filtering is in each of the SOTA DRC software I mention. And based on the 125 rooms I have worked on, I can attest, with data that it works very well. Again, I am intending to share that. And I think there still appears to be some misunderstanding on psychoacoustic filtering as we are not simply ignoring the dips...

Applying FDW after psychoacoustic filtering is applied suggests the opposite? I've raised that question years ago when talking to Uli but never got an answer

You are making guesses with no data to support your guesses Preringing has been a mathematically solved problem for quite some time and preringing compensation, if required, is in each of the SOTA DRC products mentioned.

Hey, that's my line I've played around with room correction quite a bit. For what it's worth, I do own Acourate for example. Pre-ringing to me is a perceptual problem not a mathematical one.

As covered in gory detail in my DSP book and in the video, I took 14 measurements around a 6' x 2' grid area and the results show a smooth frequency response across the area with the timing response intact, with one analysis measurement. And no, the measurements are not heavily smoothed.

Looks like the data on the slide in question was taken with REW. Can you post that .mdat?

@markus free to respond, but how about bringing some measurements or data to demonstrate your counterpoints.

Not sure what data you'd like to see. Disappointingly I have more concerns and questions than answers.

 

[Digital_Thor]

I know it is counter intuitive, but multiple measurements can actually reduce the resolution of the correction at the listening position.

Is the whole point of multiple measurements not, that you actually smooth out the response, to go for the modal problems, which are low or minimum phase, which is easier to correct with basic EQ?
Also, I believe that most people are often very happy to simply get the bass just fairly ok - considering that the main speakers are good to begin with - smooth on/off-axis FR.

 

[markus]

Is the whole point of multiple measurements not, that you actually smooth out the response, to go for the modal problems, which are low or minimum phase, which is easier to correct with basic EQ?
Also, I believe that most people are often very happy to simply get the bass just fairly ok - considering that the main speakers are good to begin with - smooth on/off-axis FR.

Multiple measurements help in identifying the modal behavior of a room and how it can effectively be corrected using PEQ. A single point measurement isn't enough.

 

[Digital_Thor]

Multiple measurements help in identifying the modal behavior of a room and how it can effectively be corrected using PEQ. A single point measurement isn't enough.

Then we agree

 

[markus]

By the way, here's a good overview of state-of-the-art room response equalization approaches:

https://www.mdpi.com/2076-3417/8/1/16/pdf

 

[dc655321]

Multiple measurements help in identifying the modal behavior of a room and how it can effectively be corrected

And, when averaged, assist in reducing measurement noise.
Which can lead to incorrect compensation if there are contributions to the signal that should be compensated for, but whose effects have been diminished by said averaging...

 

[markus]

And, when averaged, assist in reducing measurement noise.
Which can lead to incorrect compensation if there are contributions to the signal that should be compensated for, but whose effects have been diminished by said averaging...

Who said you need to average them? Here's an example from Dirac how multiple measurements can inform an optimization algorithm what to do:

 

[mitchco]

The more data the better but I'm not sure if that data is helpful as it is just "more of the same" without really answering the psychoacoustic questions I've raised.

It is more of the same to show that it working as intended, robust and repeatable If you have psychoacoustic questions then perhaps reach out to JJ or to the authors of the DSP software. Myself and others are more than satisfied. This is your problem to work out.

My perspective is not the lonesome 2 channel stereo guy with a single chair in the middle of his "sanctuary" but the regular enthusiast that can't move furniture easily or he has a dedicated listening space with rows of seating that also can't be moved.
Furthermore I do think that room correction can be more than speaker correction so furniture needs to part of the equation.

You have completely missed the point.

Applying FDW after psychoacoustic filtering is applied suggests the opposite? I've raised that question years ago when talking to Uli but never got an answer

Since you have Acourate you are free to experiment yourself and apply either way to answer your own question.

Hey, that's my line I've played around with room correction quite a bit. For what it's worth, I do own Acourate for example. Pre-ringing to me is a perceptual problem not a mathematical one.

So present your data and findings...

Looks like the data on the slide in question was taken with REW. Can you post that .mdat?

Since you have Acourate why don't you do it yourself and again present your findings and data. So far you have presented no data or measurements of your own.

 

[dc655321]

Who said you need to average them?

Not I.

 

[mitchco]

Who said you need to average them? Here's an example from Dirac how multiple measurements can inform an optimization algorithm what to do:

View attachment 161852

Again, rather that posting marketing material, how about you post some of your own data and measurements that you have taken to support any of your claims.

 

[JRS]

Well, let me also say that the opinion I stated is also based on the pretty extensive personal experience with room correction, not only with "ordinary" rooms like the we usually have at our homes but also with professional recording studios.

What may be the subject of the discussion here is the size and shape of the "small" space of the single seat related to the distance to the speakers, type of the speakers, type of room treatment and is the listener sitting behind the mixing console or sitting in a sofa in the room - to name just some of the factors. But if we assume "ordinary" non-treated room, "ordinary" consumer speakers and app 3 meters or so listening distance, than Dirac single seat measurment area pretty much gets it right.

That gibes with the professional papers I have read wherein all sorts of schemes and weighing to get the mostest the bestest have been tried, all of which rely on multiple measurements, and at the same time, validates those who claim that Dirac single location produces marvelous results. Different horses for different courses it would seem.

 

[markus]

Again, rather that posting marketing material, how about you post some of your own data and measurements that you have taken to support any of your claims.

Again, what do you want to see? I'm not the one making claims here unless you think that stating "there's not enough scientific studies" is a claim that can be proven with measurements...