Help with the basics of Frequency Response Testing

[Xmall]

Hello Audio science forums,
This is my first post - I've recently started dabbling in frequency response measurement and I was hoping for some pointers or maybe someone to point me to Frequency response measurement for dummies

Long story short, I recently repaired a Sony SS-NX1 speaker (I assume its fairly low end, but I'm upgrading from a Klipsch Quintet) and I was interested to have a look at the frequency response between the original and repaired speaker - and between my old speakers and new.

Picture of my room and recording setup (currently at the listening position - but I've also recorded either speaker directly) here:

I should note I'm very much a newbie - have a fragmented understanding of a few areas like frequency response, crossover types, room modes etc - but I might be making some very basic and fundamental errors.

My current recording setup is:

Room: Living room - for listening only. Large room, lots of furniture - very irregular space.
AMP: Denon AVR-1713 (Testing both with the room EQ (Audessey) and without)
Device: Android phone
Recording App: AudioTool with calibration files loaded - using the 'Peak Response' function to record a log frequecny sweep.
Mic: Dayton Audio IMM-6C - Calibrated mic: https://www.daytonaudio.com/product/1974/imm-6c-idevice-usb-c-calibrated-microphone
PC Software: Room EQ Wizard (REW).
Sweep generator: Online sweep generator, set to sine wave generator, logarithmic scale: https://onlinetonegenerator.com/frequency-sweep-generator.html

I've been exporting just the Hz and Db from Audiotools into REW (though i also seem to have 'voltage' measurements from the microphone - not sure if these are important).


Getting some results - a couple of samples from the listening position below (Psychoacoustic smoothing applied):

Blue trace is the Sony Speakers in Stereo, no room Equalisation.
Purple my little Klipsch Quintet in Stereo, no room Equalisation.

One thing I am quite confused about - is the significant decay in volume I'm setting in the upper frequencies when looking at the REW plot. When i look at AudioTools - it appears to indicate a much flatter response (At least upto 4K):

(This should match the blue curve above)

I thought maybe it was a scale issue - but looking the humps at 500hz and 4k REW seems to be presenting a significant reduction compared to Audiotools.


So some key questions for me are:

1) With this kind of setup, can I expect any kind of accuracy (I'm only really after sufficient for comparison)?
2) Is the 'voltage' measurement in the AudioTools file of any use in REW?
3) Can anyone comment on why REW seems to be showing a different response to Audio tools?
4) Any other tips would be very much welcome.

Thanks in advance
LJ

 

[kemmler3D]

Hello Audio science forums,
This is my first post - I've recently started dabbling in frequency response measurement and I was hoping for some pointers or maybe someone to point me to Frequency response measurement for dummies

Long story short, I recently repaired a Sony SS-NX1 speaker (I assume its fairly low end, but I'm upgrading from a Klipsch Quintet) and I was interested to have a look at the frequency response between the original and repaired speaker - and between my old speakers and new.

Picture of my room and recording setup (currently at the listening position - but I've also recorded either speaker directly) here:

View attachment 385048

I should note I'm very much a newbie - have a fragmented understanding of a few areas like frequency response, crossover types, room modes etc - but I might be making some very basic and fundamental errors.

My current recording setup is:

Room: Living room - for listening only. Large room, lots of furniture - very irregular space.
AMP: Denon AVR-1713 (Testing both with the room EQ (Audessey) and without)
Device: Android phone
Recording App: AudioTool with calibration files loaded - using the 'Peak Response' function to record a log frequecny sweep.
Mic: Dayton Audio IMM-6C - Calibrated mic: https://www.daytonaudio.com/product/1974/imm-6c-idevice-usb-c-calibrated-microphone
PC Software: Room EQ Wizard (REW).
Sweep generator: Online sweep generator, set to sine wave generator, logarithmic scale: https://onlinetonegenerator.com/frequency-sweep-generator.html

I've been exporting just the Hz and Db from Audiotools into REW (though i also seem to have 'voltage' measurements from the microphone - not sure if these are important).


Getting some results - a couple of samples from the listening position below (Psychoacoustic smoothing applied):

Blue trace is the Sony Speakers in Stereo, no room Equalisation.
Purple my little Klipsch Quintet in Stereo, no room Equalisation.

View attachment 385058

One thing I am quite confused about - is the significant decay in volume I'm setting in the upper frequencies when looking at the REW plot. When i look at AudioTools - it appears to indicate a much flatter response (At least upto 4K):

(This should match the blue curve above)

View attachment 385052

I thought maybe it was a scale issue - but looking the humps at 500hz and 4k REW seems to be presenting a significant reduction compared to Audiotools.


So some key questions for me are:

1) With this kind of setup, can I expect any kind of accuracy (I'm only really after sufficient for comparison)?
2) Is the 'voltage' measurement in the AudioTools file of any use in REW?
3) Can anyone comment on why REW seems to be showing a different response to Audio tools?
4) Any other tips would be very much welcome.

Thanks in advance
LJ

Hey, welcome to ASR, you are in the right place for this topic.

1) Kind of, although you don't need to spend a lot more to get a mic that is considered better and can be used directly with REW (MiniDSP UMIK-1)

2) Not sure, probably redundant?

3) There's a chance that the correct calibration file isn't being used at some point in the process here. It may be that Audio Tools is not saving the data with the calibration applied, so it would show up differently in REW.

4) There are tutorial threads on this stuff here, I recommend starting here: https://www.audiosciencereview.com/...om-measurement-tutorial-for-dummies-part-1.4/

5) If you want to get really serious about measuring speakers per se (not the room) you can check this out: https://www.audiosciencereview.com/...ents-spinoramas-with-rew-and-vituixcad.21860/

 

[Xmall]

Thanks Kemmler3D,
That is helpful. I'll read up on the tutes. I couldn't see a method to apply a calibration in REW - so i wrote a little excel tool as a passthrough.

Looks like you were on the money - much better response now and appears to match Audio Tools (Note the graph below is from new measurements, doesn't quite match the test completed above:

Re. direct use of REW - would the general preference be to use REW for all measurements rather than something like Audiotools? I got the USB-C version of the Dayton IMM-6, and it looks like REW recognises it and can use it.

Maybe I'll do a trial run between the two, see what i get

 

[dualazmak]

Let me share just one suggestion based on my long-year experiences on room air sound measurements.

Have you ever tested/tried "cumulative recorded white-noise averaging method" for Fq-SPL measurements instead of REW's short sine-sweep method?

At least in my case, I have been mainly doing "cumulative recorded white-noise averaging method" as summarized here #404 on my project thread.

Spoiler: Pros of Cumulative Recorded White-Noise Averaging Method for Fq Spectrum

I wrote in my post #44 on my project thread;
Throughout these careful measurements, I confirmed and validated that primitive "cumulative (recorded) white noise averaging method" is really powerful and reliable in terms of;
1. the method is universally applicable in the stages of digital out of crossover software (EKIO), DAC's analog out, amplifier SP out, and of course in the actual room SP sound,
2. the method is accurate, sensitive and reproducible, having little or no statistical fluctuation, because of the FFT averaging analysis on the "accumulated rich data" of the recorded sound,
3. the recorded "white noise tracks" can be re-analyzed any way, anytime, afterwards,
4. flexible mix-paste (sound mixing) can be done to virtually simulate any combination of the channels, especially in amplifiers' SP out signals before going into SP drivers,
5. if needed, the environmental "continuous room background noise" can be reduced/removed by the Adobe Audition's "noise capture - noise reduction" function,
6. if needed, suitable gain/level adjustment can be applied for "level matched comparison" of Fq response shapes between the different series of the recorded data,
7. flexible and suitable FFT size (as smoothing intensity) can be selected depending on the frequency zone of interest.

I once asked to Dr. Floyd Toole regarding the pros and/or cons of "cumulative recorded white-noise averaging method" against ordinary(?) sine-sweep method, and Dr. Toole kindly replied by writing here "If properly done both swept tone and noise analysis should give identical answers. It is a choice. The principal difference is in the heating of the drivers in sustained tests at high sound levels - power compression. Low frequencies require longer averaging times."

Furthermore, we, myself and @Keith_W, recently discussed on this point here and thereafter with reference to the invaluable comments given by Dr. Floyd Toole.

In any way, if you have not yet tested "cumulative recorded white-noise averaging method", I would like to recommend you testing it, and compare the results with ordinary(?) sine sweep method, especially in low Fq zone (below 100 Hz) as well as in high Fq zone, say above 8 kHz.

 

[kemmler3D]

I got the USB-C version of the Dayton IMM-6, and it looks like REW recognises it and can use it.

It might be easier to do it that way. REW is a really good tool for room / speaker measurement, although the interface can be a little opaque sometimes. You load the calibration file in the preferences. You will need to do that for sure, as it looks like the mic (pre-calibration) is more than a few dB soft in the treble. https://www.roomeqwizard.com/help/help_en-GB/html/calfiles.html

@dualazmak brings up a good point. White noise measurements can be good in part because they can also be used for spatially averaged measurements and you also avoid issues with stuff in your room resonating.

 

[dualazmak]

Thank you @kemmler3D for your quick follow-up.

And, yes, I agree the importance of knowing/understanding the Fq response of our measurement microphone.

Very fortunately, at least in my case, my specially selected (in 2008) rather affordable BEHRINGER ECM8000 microphone was found to have still fairly flat response throughout 10 Hz to 20 22 kHz (plus 0.7 dB, minus 0.5 dB, total 1.2 dB range). Last year in October, I could cross-calibrate it with recently calibrated (expensive!) EARTHWORK M50 (ref. here).

 

[Keith_W]

That room of yours is gorgeous, that view is to die for!!

Before you take any measurement, you need to ask yourself what you are measuring, and what is the intention of your measurement. If your intention is to correct your speaker with DSP, it is absolutely vital to get a meaningful measurement otherwise you will be performing wonky corrections. The photo shows a tripod perched on a sofa, with a coffee table in front of the measurement position. This means that a lot of early reflections will arrive at your microphone and contaminate the measurement.

You also need to think of low and high frequencies differently. Because of the long wavelength of low frequencies (a 20Hz wavelength is 17.15m, or 56 feet) it is impossible to take an in-room measurement free of reflections. So the bass is corrected together with the room, with the understanding that the correction is only valid for the measurement position. With high frequencies (a 20kHz wavelength is 17.15mm, or 0.7 inches) we aim to correct the loudspeaker alone, free of room influences. If you have a measurement setup where reflections arrive early, it will become impossible to window out the upper frequencies (windowing = rejecting sound that arrives later than a certain time threshold). This means that the lower frequency limit will become much higher.

I will illustrate with this example. Suppose you are in a perfect anechoic chamber where there are no reflections. You will be able to take a meaningful measurement from 20Hz - 20kHz. Now we place a sofa in the anechoic chamber, and your mic is perched on the sofa, 50cm from the closest reflecting surface. This means that measurement of any frequency with a wavelength longer than 50cm (686Hz) will not indicate what your speaker is doing, but what the speaker + sofa is doing.

In your case, you own a Dayton IMM-6. This microphone, calibrated or not, WILL suffer from reflections from the USB device you are plugging it into (phone, tablet?) not to mention the tripod itself. I am guessing a distance of 10cm at most, meaning that you will be unable to window out anything less than 3430Hz. To calculate the limit yourself, get out a tape measure and plug the distance into this calculator.

Sad to say, your microphone is only suitable for getting a rough idea what is going on. If you want to coax better accuracy with your setup, use a USB cable and mount your microphone on a boom. Move furniture out of the way. This will lower your limit where you can take meaningful measurements. Or better still, upgrade to a 48V Phantom Power mic (like a Behringer ECM8000 as the absolute minimum) + microphone interface + proper microphone tripod. Of course, whether you need all this accuracy depends on what your intentions are.

 

[dualazmak]

Yes, as @Keith_W kindly pointed, not only Fq response spectrum, but also time-domain tuning, i.e. time alignment between all the SP drivers at your listening position, would be critically important for optimization of your objective and subjective audio listening tuning and enjoyments.

As @Keith_W also pointed, objective and subjective measurements and understandings on reflective wave(s), standing waves, and so on, i.e. so called "room ambience" or "room reverberation" (which cannot be completely eliminated since our home setup can never be anechoic) would be also important.

I objectively and/or semi-objectively measured and tuned also these features using (again!) my rather primitive but reliable/reproducible methods/procedures. If you would be interested, just for your initial reference, please visit these posts under the spoiler cover on my project thread.

Spoiler

- Precision measurement and adjustment of time alignment for speaker (SP) units: Part-1_ Precision pulse wave matching method: #493
- Precision measurement and adjustment of time alignment for speaker (SP) units: Part-2_ Energy peak matching method: #494
- Precision measurement and adjustment of time alignment for speaker (SP) units: Part-3_ Precision single sine wave matching method in 0.1 msec accuracy: #504, #507

- Measurement of transient characteristics of Yamaha 30 cm woofer JA-3058 in sealed cabinet and Yamaha active sub-woofer YST-SW1000: #495, #497, #503, #507
- Identification of sound reflecting plane/wall by strong excitation of SP unit and room acoustics: #498

- Short (but important) communication on my preference in room acoustics: #502

- Perfect (0.1 msec precision) time alignment of all the SP drivers greatly contributes to amazing disappearance of SPs, tightness and cleanliness of the sound, and superior 3D sound stage: #520

- Not only the precision (0.1 msec level) time alignment over all the SP drivers but also SP facing directions and sound-deadening space behind the SPs plus behind our listening position would be critically important for effective (perfect?) disappearance of speakers: #687

- Reproduction and listening/hearing/feeling sensations to 16 Hz (organ) sound with my DSP-based multichannel multi-SP-driver multi-amplifier fully active stereo audio system having big-heavy active L&R sub-woofers: #782

- A nice smooth-jazz album for bass (low Fq) and higher Fq tonality check and tuning: #910, #63(remote thread)
- A new series of audio experiments on reflective wide-3D dispersion of super-tweeter sound using random-surface hard-heavy material: Part-4_Provisional conclusion to use Case-2 reverse reflective dispersion setting in default daily music listening: #929

By the way, only if you would be interested, you can find here #931 on my project thread the details of the latest (as of June 26, 2024) setup of my PC-DSP-based multichannel multi-SP-driver multi-amplifier stereo audio system, including all the features/aspects I pointed in the above.

 

[Xmall]

Thanks all for the great info.
@kemmler3D - Thanks I managed to get my calibration files loaded into REW - I don't currently have the requisite cables so I'll continue dabbling with the phone app for now, then try again.

@dualazmak - Looks like you have been doing some amazing work! I wont profess to understand more than about 0.5% of it at this point but it gives me something to work towards. Will look into the use of the white noise, is this something that is available through REW - or does that need to be provided through an external source?

@Keith_W - thank you. The house has been a long running renovation (much to my wife's dismay) - but 6 years on and we are nearly there!

"Tripod perched on a sofa" - That pretty much sums up my audio prowess unfortunately! This is very early days for me - as you can probably tell. You raise some good points in your post.

What are my intentions:
1) Learn something about audio and audio measurement (even if the answer ends up being, "its damned hard and I'll just fall back on Audyssey )
2) Do an objective test between my old speakers (Klipsch Quintet) and my new speakers (A pair of $15 (AUD) repaired tip shop Sony SS-NX1's) and try to understand WHY I think they sound better than my old tiny Klipsch.
3) Try and do some objective tests on my Sony Speakers. One had a bad cap in the crossover, led to a blown tweeter. I replaced the tweeter, but used an 8Ohm instead of 6 Ohm - it was also louder so I added a series and parallel resister with the tweeter to bring the overall impedance of the tweeter line back to 6Ohm, and drop the volume. I think it worked out OK - but would like to see some measurements.
4) If I haven't already fried my brain, lost all my hair - or ended up in a mental ward trying to figure out how all of this works, see If I can tune the system a bit further and improve the sound . I'm quite pleased with it as is - so this would really just be a bonus.

Reflections:
You also make some interesting points re. reflections. When trying to tune speakers - should we be:
A) Tune them individually minimising reflections - and that will lead to better overall sound; or
B) Tune them in the listening position with the stuff which will actually be in the room?
I'm curious, if I were to remove furniture - open doors (we have a big set of double sliders about 4m behind the sofa) and reduce reflections - would I then be 'randomising' the sound once everything went back to its usual spot?

Gear:
I do have a boom mike, audio interface and an AKG perception 220 condenser mic (back when I was recording some guitar work) - but I don't think that one was calibrated so assuming I cant use it (and have no way to calibrate). I'll probably start with a long USB C cable, have a good play with this Dayton mic then if I'm having fun - invest in some better kit (its currently hard to justify, when I'm running a 10YO amp and $15 repaired speakers).

At this stage I've used Audyssey which came with my Denon amp to tune the setup, but I found that hit and miss (Kept on way over boosting the high frequencies) - got it to what I consider an improvement over Audyssey off - but would like to dabble and see if I can improve.


Thanks again for the detailed responses from everyone - Very much appreciated

 

[Verig]

The UMM version of the Dayton is better suited for the job but it costs the same as Umik so not sure why one would chooce that (other than availability). But yes, for now a long usb-c cable and a boom instead of a camera tripod do nicely.

 

[dualazmak]

@dualazmak - Looks like you have been doing some amazing work! I wont profess to understand more than about 0.5% of it at this point but it gives me something to work towards. Will look into the use of the white noise, is this something that is available through REW - or does that need to be provided through an external source?

In order to eliminate any "black box type" (to me) analysis/internal-processing, I dared not to use REW or any similar tuning software tools; I just played very carefully prepared/QC-ed "white noise" and recorded the room air sound by using measurement microphone (cross-calibrated BEHRINGER ECM8000) connected to an audio interface (TASCAM US1xHR) for digital recording with the second independent PC.

The recorded track(s) were then analyzed with PC in my office upstairs for gain-time spectrum, gain-Fq-time 3D color spectrum, and FFT averaging analysis for Fq-SPL spectrum, using ADOBE Audition 3.0.1. You may use free Audacity (I have/use ver.3.0.2) for this kind of analysis as replacement for Adobe Audition 3.0.1, even though I highly recommend you using Adobe Audition for this kind and other audio analysis.

Of course, the "precision" of the white noise is critical in this case. I always use "track-11: White Noise" of "SONY Super Audio Check CD" which has ultra-high precision of eleven nine (99.999999999) or higher (ref. here #651). Attached please find the PDF booklet of "SONY Super Audio Check CD" (English translation by myself.)

In case if you would be seriously interested in using any of the intact tracks of "SONY Super Audio Check CD", and all the tone-burst test signals I prepared and used in the posts under the spoiler cover of my above post #8, you would please simply let me know through PM (person-to-person) communication system of this ASR Forum writing your wish.

BTW, suitable and consistent your own "Audio Sampler/Reference Music Playlist of Excellent Recording Quality" should be also indispensable for your coming/future audio and room tuning exploration. If you would be again seriously interested in using intact track(s) of my "Audio Sampler/Reference Music Playlist" (ref. my thread here), please simply PM me writing your wish.

 

[pma]

Long story short, I recently repaired a Sony SS-NX1 speaker (I assume its fairly low end, but I'm upgrading from a Klipsch Quintet) and I was interested to have a look at the frequency response between the original and repaired speaker - and between my old speakers and new.

Picture of my room and recording setup (currently at the listening position - but I've also recorded either speaker directly) here:

Would you try to move the speakers a bit forward, to the front edge of the shelf they are standing on? As of now, there will be a dip due to subtraction of reflections from the woofer and tweeter from the shelf surface.

 

[Keith_W]

Ah, so you are in Australia! Pretty sunny where you are, I am guessing you are not in Victoria because it is really miserable down here!

Reflections:
You also make some interesting points re. reflections. When trying to tune speakers - should we be:
A) Tune them individually minimising reflections - and that will lead to better overall sound; or
B) Tune them in the listening position with the stuff which will actually be in the room?
I'm curious, if I were to remove furniture - open doors (we have a big set of double sliders about 4m behind the sofa) and reduce reflections - would I then be 'randomising' the sound once everything went back to its usual spot?

How you tune your speakers depends on your philosophy of tuning. Some people tune full range, some people leave the upper frequencies alone and only EQ below the Schroder frequency. Remember that I said it is fundamental to understand that upper and lower frequencies need to be thought of differently and a different solution applied for each? This is a very important point! Below the Schroder frequency, what you hear is the room + the speaker. Above the Schroder frequency, the speaker's performance predominates.

Now you need to be very clear about what you are doing. If you were a speaker designer or a reviewer, you need to know what your speakers are doing from 20Hz to 20kHz without any room influence. But you are a hobbyist, so all you need to know is what your room + speaker system is doing below Schroder, and what your speaker alone is doing above Schroder. This greatly simplifies your approach and makes measurement much easier. But OTOH if you are only doing this for fun and learning, and do not plan to use the information to manipulate the sound ... then you don't need to take as much care and consideration when you are taking measurements. You can still get useful information from your mic, you just need to know what the limitations are.

A good approach which most of us in ASR use (and what Dr. Floyd Toole recommends) is to EQ the lower frequencies to provide the smoothest possible bass response, and either leave the upper frequencies untouched, or with very gentle "broad and low Q" tone control type EQ. If this was the approach you decide on, then I would leave the sofa and coffee table in place because the bass will be corrected as part of the room, and completely ignore the rubbish results that you would see in the upper frequencies. It doesn't matter because you won't be correcting them anyway.

Opening the doors behind the sofa will not prevent reflections from contaminating the measurement. Sound reflects from the closest reflecting surface - this will be the floor or the ceiling (assuming you have cleared everything else out of the way). Big doors which open to the outside can be thought of as the ultimate room treatment, because sound goes out the door and never comes back (assuming you don't have anything that reflects sound in your backyard). So you might choose to have two sets of EQ - one for doors open, another for doors closed.

Gear:
I do have a boom mike, audio interface and an AKG perception 220 condenser mic (back when I was recording some guitar work) - but I don't think that one was calibrated so assuming I cant use it (and have no way to calibrate). I'll probably start with a long USB C cable, have a good play with this Dayton mic then if I'm having fun - invest in some better kit (its currently hard to justify, when I'm running a 10YO amp and $15 repaired speakers).

At this stage I've used Audyssey which came with my Denon amp to tune the setup, but I found that hit and miss (Kept on way over boosting the high frequencies) - got it to what I consider an improvement over Audyssey off - but would like to dabble and see if I can improve.

Your AKG 220 microphone is unsuitable for measurement because it is not omnidirectional, it has a cardioid pattern.

I don't use Audyssey or Dirac or any of those "one button DSP" type solutions because I don't want software making decisions for me. I have little faith in those mass market DSP solutions. Just look at MiniDSP - if you order a mic from them, it comes with a mini tripod. If you attempt to take a measurement with a mini tripod, the result will be meaningless. You then feed this meaningless measurement into automated software and it spits out a meaningless result.

BTW in my previous post I forgot to say something about the psychoacoustic smoothing you are using. In short: use 1/12 or 1/24 smoothing instead. Psychoacoustic smoothing tries to mimic what your ears are hearing, and it will smooth over many flaws in your frequency response.

 

[Verig]

I don't use Audyssey or Dirac or any of those "one button DSP" type solutions because I don't want software making decisions for me. I have little faith in those mass market DSP solutions. Just look at MiniDSP - if you order a mic from them, it comes with a mini tripod. If you attempt to take a measurement with a mini tripod, the result will be meaningless. You then feed this meaningless measurement into automated software and it spits out a meaningless result.

Dirac works great. You don't have to leave it full auto, there's plenty to adjust with more advanced taste. Correcting a set with two subs is a breeze. For a newcomer with REW and a mic... well, there's a learning curve for sure. Depending on how far OP wants to go I'd say a little bit of cheating is not always bad.

I agree about wrong kinds of mic supports and placement though. Why go all the trouble and have the mic basically in your butt.

 

[DWPress]

Another measurement method to try is the MMM approach which is summed up nicely in this post. You will not get any of the time domain information but it will give you a more accurate frequency response than a single sweep in a static position. Use your mic boom and move slowly.

 

[Xmall]

This is great, thanks all.

@Verig you ask why I ended up with the IMM not the UMM. Basically I found a youtube clip that talked through measurements with a smart phone and i thought it looked fun (

if interested). REW looked 'complicated, scary and like something only professionals would use' so I didn't originally intend to use that. Now that I've got some guidance though I'll give it a try.

In order to eliminate any "black box type" (to me) analysis/internal-processing, I dared not to use REW or any similar tuning software tools;

That sounds like a dedicated approach I will probably continue with measurement and REW for now - but I'm intending to isolate and test variables in sequence - to try and form some kind of understanding of how each impacts the sound.

Would you try to move the speakers a bit forward, to the front edge of the shelf they are standing on? As of now, there will be a dip due to subtraction of reflections from the woofer and tweeter from the shelf surface.

Thanks PMA, I will do that - and I might test some isolating foam as well (have some stiff foam off-cuts). Curious to see how much impact I see.

How you tune your speakers depends on your philosophy of tuning. Some people tune full range, some people leave the upper frequencies alone and only EQ below the Schroder frequency. Remember that I said it is fundamental to understand that upper and lower frequencies need to be thought of differently and a different solution applied for each? This is a very important point! Below the Schroder frequency, what you hear is the room + the speaker. Above the Schroder frequency, the speaker's performance predominates.

Thanks Keith.

One thing I noticed going from my Klipsch Quintet, to the Sony SS-NX1 - was much better clarity in instruments. I.e. with the Klipsch (tiny satellite) it felt difficult to follow a tune from a specific instrument - where with the Sony it felt like you could hear each and every one clearly and track them.

Photo of the two below (They are paired with a sub)

From what I read, I thought that 'clarity' came from the mids so I was going to try and tune these to see if I can improve further - but I might simply be getting mixed up with the obviously more prominent bass response.

On Schroder - how much does the '250Hz' rule of thumb vary by room size?

I've probably got a large room, and acoustically may be complicated than many. Couple of key points:

• Main room is 6.3x7.7m ~ 49 m2 - but also part of the kitchen and entry area;
• Extended area of the larger space is about 100m2;
• I have three 'cathedral' (for lack of a better term) skylights in this area - which will probably focus reflections and make my life acoustically complicated.

Couple of screenshots of the overarching design (though refer to the photo in the first post for actual furniture - no floor boards, different couch etc).

General space:

Skylight Configuration

I think at this stage, I'll actually start playing with some variables people are pointing out - to see if I can measure the impact on sound.

Based on feedback so far, what I want to run:
1) Speakers in current position - vs speakers moved forward to edge of the TV cabinet and sitting on foam;
2) Current tripod and mic setup vs boom stand and mic setup;
3) Couch and coffee table in place vs removed;
4) Rear doors open vs rear doors closed;
5) Directly under one of the skylights - vs directly between 1 of the skylights. (maybe 4 positions, so i can get a feel for what is skylight impact vs general room varaition)
6) On-axes measurements vs well off-axis measurements
7) Cumulative white noise approach recommended by dualazmak vs sine wave sweep (if I can figure it out!)

Another measurement method to try is the MMM approach which is summed up nicely in this post.

I will give this a shot as well, and see how it differs. How long a sweep should I be running (30 seconds, 2 minutes etc)? I imagine as I sweep, if I am hitting nodes/antinodes It could give a slightly different result each time?

I should note, the tests above are not because I question the combined wisdom of the forum - I just find there is nothing quite like testing each variable yourself to understand the impact

I need to actually do some work today (work for myself from home) - but I'll try and find a time slot over the next few days where I don't have three kids running rampant and start sharing results.

Ah, so you are in Australia! Pretty sunny where you are, I am guessing you are not in Victoria because it is really miserable down here!

Keith, curious how you guessed my general location? Was it timing of the posts? I'm well further South of you, in Sunny warm Hobart


Thanks again all for the amazing responses! This might be the most helpful forum I've ever joined!

 

[DWPress]

I will give this a shot as well, and see how it differs. How long a sweep should I be running (30 seconds, 2 minutes etc)? I imagine as I sweep, if I am hitting nodes/antinodes It could give a slightly different result each time?

The post I linked to gives all the information you need and screenshots on how to set up REW to do it. Very simple actually and will give you repeatable results.

Personally, I leave any furniture and doors in position for any measurements I take - removing them to measure potentially alters the result and if you're going to be sitting on the couch for music/movies anyway it's just part of the equation. Moving the speakers as PMA suggested will also help remove some negative early reflections. I also note that you are doing sweeps at a relatively low SPL - get a SPL meter app for your phone and do sweeps/MMM at 75-85dB to overcome the noise floor of your mic and the inherent background noise of your home and you'll get more accurate results.

 

[tmtomh]

Personally, I leave any furniture and doors in position for any measurements I take - removing them to measure potentially alters the result and if you're going to be sitting on the couch for music/movies anyway it's just part of the equation.

I strongly agree with this. As long as you're trying to measure in-room response (as opposed to the speaker itself independent of the room), then to me it makes sense for the measurement to include the same reflections that you'll be exposed to during normal listening.

 

[dualazmak]

BTW in my previous post I forgot to say something about the psychoacoustic smoothing you are using. In short: use 1/12 or 1/24 smoothing instead. Psychoacoustic smoothing tries to mimic what your ears are hearing, and it will smooth over many flaws in your frequency response.

Personally, I leave any furniture and doors in position for any measurements I take - removing them to measure potentially alters the result and if you're going to be sitting on the couch for music/movies anyway it's just part of the equation.

Yes, fully agree with you two. I too usually do not like, do not use, psychoacoustic smoothing* which "hides" almost all the "room mode" in mid to high Fq range, and hence it is not suitable for measurement and tuning of audio gears as well as your actual room mode including the furniture alignment.
*In REW user manual: "Psychoacoustic smoothing uses 1/3 octave below 100Hz, 1/6 octave above 1 kHz and varies from 1/3 octave to 1/6 octave between 100 Hz and 1 kHz. It also applies more weighting to peaks by using a cubic mean (cube root of the average of the cubed values) to produce a plot that more closely corresponds to the perceived frequency response."

Just for example, and a reference for OP @Xmall, please look at my almost best tuned latest Fq-SPL curve where I applied FFT Size of 4096 as smoothing factor on 20 Hz to 20k Hz recorded cumulative flat white noise;

Please note that the fine up-and-down structure of the room sound total Fq-response curve (yellow Fq curve = Total magnitude) is not statistical fluctuations, but they exactly represent/reflect room acoustics (reflections, resonances, mutual compensations, etc.) thanks to the "cumulative (recorded) white noise averaging" method. My listening room is not symmetrical in L-R, Up-Down, not symmetrical in furniture alignments, and of course not anechoic (please refer here).

Such fine structure of the Fq-SPL curve, i.e. fine room acoustic modes, can be clearly seen/observed as reproducible "horizontal stripes" in Adobe Audition's 3D Fq-Gain-Time color spectrum, as shown bellow; in this case FFT Size of 8192 was applied. Whole of the 55 sec track (rich sound data!) was selected and scanned (averaged) for FFT to give the yellow highly reproducible reliable FFT Fq-SPL spectrum curve.

 

[kemmler3D]

I strongly agree with this. As long as you're trying to measure in-room response (as opposed to the speaker itself independent of the room), then to me it makes sense for the measurement to include the same reflections that you'll be exposed to during normal listening.

+1 - measure how you listen, unless you plan to EQ above Schroder for some reason, in which case reflections could throw things off even worse.