What are good measurements? Please!!!

Joffieb · Dec 12, 2024

I have been using REW for years now and read hundreds of posts but the one thing that I have struggled to find answers to is, "What is a good measurement?"

Floyd E Toole says, (rightly or wrongly) speakers that measure well, sound good in blind tests, BUT what does "measure well" mean?

For the SPL I may have seen that the whole measurement must be within 10dB, or no peak or null should be more than 5 dB from the middle.

So please someone answer:

1. What is a good SPL reading? (What smoothing is okay to apply when looking? to me 1/1 gives a very but readable view)
2. What is a good distortion graph? (at what frequencies)
3. What is a good spectrogram?
4. What is a good IR or Filtered IR reading?
5. Any other widows that indicate a good set up/room?
6. What is Rt60 and what's a good one to aim for?
7. How can I get my SPL to show +/- 10 dB range with 0 in the middle?

Thanks!!!

FargateOne · Dec 12, 2024

Unfortunatly, it is not that easy to answer your questions because you must to read a lot and learn a lot. Acoustics and sound reproduction is not our day to day job. Did you read all the references given in post 1 of this thread:

Some guides to REW and acoustic measurement

MiniDSP UMIK-1 setup article MiniDSP loudspeaker measurement article using REW and UMIK-1 REW, UMIK-1 and HDMI on Windows AustinJerry's guide to setting up REW with USB microphones and HDMI How to set up your mac for REW Automotive Sound with miniDSP Products and REW Realtraps Room Measuring...

www.avnirvana.com

And I mean : ALL
After that you must to be sure to really understand what we read (that's a challaenge for me at least)
After more than 10 years in the rabbit hole, I would have like to get answer to your questions; my quest would have been more simple. But it is a wish.

Joffieb said:
1. What is a good SPL reading? (What smoothing is okay to apply when looking? to me 1/1 gives a very but readable view)

1/1 gives hope but it is a lure. Var smoothing or psychoacoustic smoothing is better. If you are not afraid of bad looking graph but more truthfull ones, try 1/24 smoothing.

Joffieb said:
4. What is a good IR or Filtered IR reading?

It depends of what you are searching for.If it is timing between speakers = all first peak on top of each other, for instance.

Joffieb said:
Floyd E Toole says, (rightly or wrongly) speakers that measure well, sound good in blind tests, BUT what does "measure well" mean?

I don't mean to be rude but read his book. Well written, easy to read and soooo helpfull. You will have your answer in it.

Joffieb said:
6. What is Rt60 and what's a good one to aim for?

Begin with HELP in REW it is a concentrated of university grade formation in all those matters.

Joffieb said:
7. How can I get my SPL to show +/- 10 dB range with 0 in the middle?

Don't understand . Read AustinJerry guide (see above) and you will have what you need to read your SPL. Target 0 IMHO isn't necessary. I measure at 75dB and I am ok .

dan77 · Dec 12, 2024

I struggle with this as well.
I also have been fooling around with REW for about 12 years.

i really lean towards the waterfall graph lately. Its an SPL graph that has a third axis - time
Very helpful to see decay issues. So you can see where to locate speakers, but at the same time (pun intended) see how treatments factor in.

I had never seen anyone post a ‘good’ graph versus a ‘bad’ graph (in the same post). So I never felt I really understood what I was chasing.

Then, a while back I found these images from the REW help PDF: https://www.roomeqwizard.com/REWhelp.pdf#page180

So anyway, the PDF above has other examples, but for the waterfall graph here is what an absolute perfect waterfall chart would look like. Getting a speaker to chart like this is impossible, but it is without a doubt, your goal:

This is a waterfall chart that shows exceptionally poor characteristics.
Peak to peak delta is around 17 db. The ringing (delay/echo) is well in excess of 300 ms around 28 hz and 65 hz.
There does seem to be some good news above 90 hz as the chart ‘smooths’ out and no decay longer than 240 ms.
Meaning it does start to look a bit like the ‘perfect’ chart above.

So those charts are from the REW help PDF
They give you two extremes

For the heck of it, here is my actual HT. You can see that I have some serious decay issues below 40 hz, but I have my bass turned up too much for the average person. I just like it that way and it does sound really good (at least to me)
From 6 hz and up the decibal difference is rarely above 10 db. I have worked off and on for years to get all the speakers to play well with each other.

dan77 · Dec 12, 2024

Here is the link to the REW help PDF: https://www.roomeqwizard.com/REWhelp.pdf#page180

dan77 · Dec 12, 2024

That same guide shows what a ‘perfect’ spectrogram graph looks like.
Again, not directly obtainable in real life, but it is without a doubt, your goal and it should be what you are chasing:

The PDF also shows a ‘what not to do’ spectrogram graph as well
Note the decay (ringing/echo) times around 25 hz are almost one full second! Not good!
Also around 58 hz are exceptionally long decay times
Anything over 200 ms should be looked at and anything over 300 ms should be addressed
Below 40 hz the rules change. We try to keep it under 400 ms.

I hope this helps

JStewart · Dec 12, 2024

Joffieb said:
Floyd E Toole says, (rightly or wrongly) speakers that measure well, sound good in blind tests, BUT what does "measure well" mean?

My interpretation of Dr. Toole’s work is different. My interpretation is that speakers that measure well have a statistical preference in blind tests. In context of the OP question I think the distinction becomes important.

Here is a quote from his book Sound Reproduction The Acoustics and Psychoacoustics of Loudspeakers and Rooms Third Edition Chapter 5.9, my emphasis added, that I think sums up the value of spinorama measurements nicely…

“From the simple analog ink lines on graph paper of the pre-digital era, we have moved on to computer-controlled collection of unlimited amounts of data, elaborate digital processing of that data, useful predictions of steady-state room curves, and ultimately to the prediction of listener subjective ratings of sound quality. We may never eliminate the need for subjective evaluations, but it is greatly reassuring to know that there are measurements that add guidance to the evaluation of a product. In fact, one may venture a challenge that an examination of the right set of anechoic measurements may well be more reliable as an indicator of sound quality than a “take it home and listen to it” subjective evaluation, and far more trustworthy than an uncontrolled listening session at an audio show or store. Now, with brick and mortar audio stores difficult to find, and Internet sales becoming more popular, it is highly useful if data like spinoramas, or a reasonable facsimile, are available to shoppers.”

So to the question of what constitutes a good measurement characteristics, I think the answer is they don’t exist. I‘ve not come accross standards defined and proofed by scientific rigor. Not even for frequency response, let alone the other characteristics mentioned in the OP. I recall reading somewhere, and I hope this is correct, that with respect to the measurements (referring to CEA 2034 aka Spinorama), Dr. Toole stating only that smoother is better. So there you have it. The measurements are comparative tools and do not define what is good or even minimally acceptable.

This is not to say the various measurement characteristics have no value. They can each highlight aspects which can sometimes be changed, for example, an impulse response can show and perhaps aid in locating, unwanted reflections. An SPL response can show sharp resonances or nulls.

A move in the right direction of establishing minimum standards for and defining objective degrees of “goodness” is CEDIA/CTA-RP22. Copy available here:

RP22 - Immersive Audio Design Excellence | CEDIA

Redefine how you quantify system performance - Visit CEDIA to learn more and download our RP22 Document: Immersive Audio Design Recommended Practice.

cedia.org

Another document attempting to establish standards and which may provide some guidance for questions in the OP is from Nyal Mellor, Acoustic Frontiers LLC & Jeff Hedback, HdAcoustics. 2011 found here:

https://cdn.shopify.com/s/files/1/1639/4021/files/acoustic_measurement_standards_4e7bbd86-32f8-45a8-9a3e-5467aecb2073.pdf?v=1682245032

I think @FargateOne has it right. Read and study everything one can and understand it probably won’t be enough. I’ll add to that, experiment oneself making changes and looking at resultant measurement changes. Post measurement data and ask questions. If we’re lucky someone will have insightful advice.

ddude003 · Dec 12, 2024

I would add that even good measuring and great sounding speakers may not measure or sound so great in any given room...
Physical acoustic absorbers and diffusers along with Digital Room Correction is the way to go... Mitch Barnett has some great resources at https://accuratesound.ca/resources-audio-digital-signal-processing/

Joffieb · Dec 13, 2024

FargateOne said:
Unfortunatly, it is not that easy to answer your questions because you must to read a lot and learn a lot. Acoustics and sound reproduction is not our day to day job. Did you read all the references given in post 1 of this thread:

Some guides to REW and acoustic measurement

MiniDSP UMIK-1 setup article MiniDSP loudspeaker measurement article using REW and UMIK-1 REW, UMIK-1 and HDMI on Windows AustinJerry's guide to setting up REW with USB microphones and HDMI How to set up your mac for REW Automotive Sound with miniDSP Products and REW Realtraps Room Measuring...

www.avnirvana.com

And I mean : ALL
After that you must to be sure to really understand what we read (that's a challaenge for me at least)
After more than 10 years in the rabbit hole, I would have like to get answer to your questions; my quest would have been more simple. But it is a wish.

1/1 gives hope but it is a lure. Var smoothing or psychoacoustic smoothing is better. If you are not afraid of bad looking graph but more truthfull ones, try 1/24 smoothing.

It depends of what you are searching for.If it is timing between speakers = all first peak on top of each other, for instance.

I don't mean to be rude but read his book. Well written, easy to read and soooo helpfull. You will have your answer in it.

Begin with HELP in REW it is a concentrated of university grade formation in all those matters.

Don't understand . Read AustinJerry guide (see above) and you will have what you need to read your SPL. Target 0 IMHO isn't necessary. I measure at 75dB and I am ok .
View attachment 76533

Thanks for taking the time and replying so thoroughly.

Joffieb · Dec 13, 2024

JStewart said:
My interpretation of Dr. Toole’s work is different. My interpretation is that speakers that measure well have a statistical preference in blind tests. In context of the OP question I think the distinction becomes important.

Here is a quote from his book Sound Reproduction The Acoustics and Psychoacoustics of Loudspeakers and Rooms Third Edition Chapter 5.9, my emphasis added, that I think sums up the value of spinorama measurements nicely…

“From the simple analog ink lines on graph paper of the pre-digital era, we have moved on to computer-controlled collection of unlimited amounts of data, elaborate digital processing of that data, useful predictions of steady-state room curves, and ultimately to the prediction of listener subjective ratings of sound quality. We may never eliminate the need for subjective evaluations, but it is greatly reassuring to know that there are measurements that add guidance to the evaluation of a product. In fact, one may venture a challenge that an examination of the right set of anechoic measurements may well be more reliable as an indicator of sound quality than a “take it home and listen to it” subjective evaluation, and far more trustworthy than an uncontrolled listening session at an audio show or store. Now, with brick and mortar audio stores difficult to find, and Internet sales becoming more popular, it is highly useful if data like spinoramas, or a reasonable facsimile, are available to shoppers.”

So to the question of what constitutes a good measurement characteristics, I think the answer is they don’t exist. I‘ve not come accross standards defined and proofed by scientific rigor. Not even for frequency response, let alone the other characteristics mentioned in the OP. I recall reading somewhere, and I hope this is correct, that with respect to the measurements (referring to CEA 2034 aka Spinorama), Dr. Toole stating only that smoother is better. So there you have it. The measurements are comparative tools and do not define what is good or even minimally acceptable.

This is not to say the various measurement characteristics have no value. They can each highlight aspects which can sometimes be changed, for example, an impulse response can show and perhaps aid in locating, unwanted reflections. An SPL response can show sharp resonances or nulls.

A move in the right direction of establishing minimum standards for and defining objective degrees of “goodness” is CEDIA/CTA-RP22. Copy available here:

RP22 - Immersive Audio Design Excellence | CEDIA

Redefine how you quantify system performance - Visit CEDIA to learn more and download our RP22 Document: Immersive Audio Design Recommended Practice.

cedia.org

Another document attempting to establish standards and which may provide some guidance for questions in the OP is from Nyal Mellor, Acoustic Frontiers LLC & Jeff Hedback, HdAcoustics. 2011 found here:

https://cdn.shopify.com/s/files/1/1639/4021/files/acoustic_measurement_standards_4e7bbd86-32f8-45a8-9a3e-5467aecb2073.pdf?v=1682245032

I think @FargateOne has it right. Read and study everything one can and understand it probably won’t be enough. I’ll add to that, experiment oneself making changes and looking at resultant measurement changes. Post measurement data and ask questions. If we’re lucky someone will have insightful advice.

Thanks for taking the time to reply.

Joffieb · Dec 13, 2024

dan77 said:
That same guide shows what a ‘perfect’ spectrogram graph looks like.
Again, not directly obtainable in real life, but it is without a doubt, your goal and it should be what you are chasing:
View attachment 76538

The PDF also shows a ‘what not to do’ spectrogram graph as well
Note the decay (ringing/echo) times around 25 hz are almost one full second! Not good!
Also around 58 hz are exceptionally long decay times
Anything over 200 ms should be looked at and anything over 300 ms should be addressed
Below 40 hz the rules change. We try to keep it under 400 ms.
View attachment 76539

I hope this helps

Thanks so much for the reply and effort.

skid00 · Dec 13, 2024

I have accumulated several mdats of various speakers over the years. Here is data from my Carver Amazing Platinum Mark IV, and a Harmon F328 (which is the best I've seen).

My Carver is heavily EQ'd, in a non-treated great room 18' x 36' with cathedral ceiling, and one wall mostly of glass. Note that it is a dipole radiator (open front and back), with a 5' vertical ribbon, and 4 12' woofers open to the room on both sides. It's hard to get the kind of bass you get from boxes that pressurize the room, so this 'excessive' bass boost is what sounds appropriate to my ears. I measured 48" from the ribbon.

I don't know the environment of the F328 measurments. A pair of these is something like $16,000 dollars.....

Here's Klippel measurements of the F328. https://www.audiosciencereview.com/forum/index.php?threads/revel-f328be-speaker-review.17443/

Joffieb · Dec 14, 2024

JStewart said:
Another document attempting to establish standards and which may provide some guidance for questions in the OP is from Nyal Mellor, Acoustic Frontiers LLC & Jeff Hedback, HdAcoustics. 2011 found here:
https://cdn.shopify.com/s/files/1/1...-32f8-45a8-9a3e-5467aecb2073.pdf?v=1682245032

YESSSS!!!!!!!!!!!

Thank you!!!!!!

ddude003 · Dec 14, 2024

From the Bruel and Kjaer application notes found at the link I posted above...

Screen Shot 2024-12-14 at 3.11.03 PM.png

basscleaner · Dec 16, 2024

Let me say some words about "good room acoustics". Nyal Mellor and Jeff Hedback's article introduces controversial provisions have to be proved. For instance, the sound quality for low frequencies must be based on, the first, an ear SPL sensibility. Well-known value for it is in the range 1.5 - 2.5 dB. It means, that SPL level greater than 3 dB for neighboring frequencies will strongly emphasize such a sound. It means, that we will have negative impression due to "peaks and dips" hearing. Proposed range for it in the article is +- 5 dB can't be accepted as control parameter at all for high quality sound. Next, "good quality sound" should not contain any phase mismatches for all speakers. It concerns all the sound range and must to be taken into consideration before any other room treatment. The third, room dimensions must be selected according to low frequencies sources, i.e. your room dimensions must be Acoustical.
Term "good measurements" means different criterium for low, mid and high frequencies range. The choice of 30 - 300 Hz for low frequencies range causes confusion for me, because it contains from 200 to 300 Hz too many very narrow peaks and dips zone, which needs to be averaged and not to be included into LF range at all.

basscleaner · Dec 17, 2024

And one more. Nyal Mellor and Jeff Hedback's advice for (2x) of 5/8" gypsum board like your room lightweight wall construction for low frequencies response causes confusion to me. The first, because it doesn't have sufficient mass (about 20-25 kg/m2 - too little mass surface. The surface mass has to be no less, than 35 kg/m2); the second, if such a construction placed with gap from masonry wall, hence there will be resonant absorption, depending on air gap thickness. There is the rule: "More mass - more bass"

JStewart · Dec 17, 2024

basscleaner said:
Nyal Mellor and Jeff Hedback's article introduces controversial provisions have to be proved.

Still, I think that even though not proved, the document seems fairly reasonable, appears to take real world into account, discusses both time and frequency domains with REW measurements explained. In short, some good information to learn from.

basscleaner said:
Proposed range for it in the article is +- 5 dB can't be accepted as control parameter at all for high quality sound.

The article states +/- 5dB, two speakers, 1/3 octave smoothing. In the lowest octaves +/- 15 or 20dB is common. Certainly not as good as it can be, but the guidance does not use EQ and this seems good without EQ. Maybe I’m wrong. The article itself says

So, get it the best one can and then employ EQ if one wants to. My take.

basscleaner said:
The choice of 30 - 300 Hz for low frequencies range causes confusion for me, because it contains from 200 to 300 Hz too many very narrow peaks and dips zone, which needs to be averaged and not to be included into LF range at all.

While they didn’t state it, this range would be in the modal range of the frequency band. So with this in mind and the proposed smoothing they are using and the corresponding +/- deviations, it seems to make sense if we’re not talking about employing EQ to get there.

JStewart · Dec 17, 2024

basscleaner said:
And one more. Nyal Mellor and Jeff Hedback's advice for (2x) of 5/8" gypsum board like your room lightweight wall construction for low frequencies response causes confusion to me. The first, because it doesn't have sufficient mass (about 20-25 kg/m2 - too little mass surface. The surface mass has to be no less, than 35 kg/m2); the second, if such a construction placed with gap from masonry wall, hence there will be resonant absorption, depending on air gap thickness. There is the rule: "More mass - more bass"

This is their comment from the article on the recommendation.

A difference of opinion or differing goals?

JStewart · Dec 17, 2024

One could, I suppose, just go with these words of wisdom from Dr. Toole

✂️ Listening to Crappy Speakers

15 seconds · Clipped by J Stewart · Original video "Chat with Dr. Floyd Toole" by Erin's Audio Corner

youtube.com

(ok, taken a bit out of context. Watch the video.)

ddude003 · Dec 17, 2024

This idea of learning about room acoustics and the measurement of room acoustics is not a trivial area to explore... This is where some information is good, more information is better and too much information is just right... :reading:

What are good measurements? Please!!!

New Member

Member

Member

Member

Member

Senior AV Addict

Senior AV Addict

New Member

New Member

New Member

Member

Attachments

New Member

Senior AV Addict

New Member

New Member

Senior AV Addict

Senior AV Addict

Senior AV Addict

Senior AV Addict

Popular tags