Low Frequency Measurement using Test CD and SPL Meter

[aashish351]

Hi,

I came across this link and decided to do what it said.

RealTraps - Test Tone CD

What I did exactly is:
1. Burned the test frequency tracks ranging from 10 Hz to 300 Hz on a CD.
2. Calibrated the SPL reading to 75 dB using the Pink Noise track in the CD
3. Played the CD in the system.
4. Used my Android phone app Sound Meter to carry out dB measurements as suggested.
5. Plotted in Excell using x-axis as log scale and y-axis as linear with 75db as reference value 0.

The resultant graph is attached below:

By aashish351 at 2012-08-06

I only went till 180 Hz although the CD is till 300 Hz. Will do it again also and will also measure the speaker from 1 m distance and plot graphs.

Now, can someone familiar with room response measurement or anybody with knowledge in this area make an assessment of what we see here and what I did is correct or incorrect.

 

[sidvee]

aashish,
The plot is showing that the low frequency response of your room begins to decline at 100hz with severe bass suckouts at 160hz, 100 hz, 85hz and 80hz and -6db point at approx. 60hz below whcih the bass rolls off significantly. It indicates that there is not much bass punch in your room and neither is the bass response smooth.
Not sure whether what you measured is correct or not, because one has to know the accuracy of the sound meter used, and not too sure how accurate the app is. Better to use a sound level meter like the radio shack model (though these are not too accurate at low frequencies either, but corrected values are available).
Cheers,
Sid

 

[aashish351]

Thanks sid...

I too feel that there is a margin for error in this measurement because subjective listening does not suggest very low bass in my room. But it is not sufficient either, I have felt many times that the bass in my room is not smooth. Sometimes I feel the vibration and sometimes just a faint note. so the measurements are generally in the correct direction but not very accurate as you pointed out. The link mentioned that for low frequencies, spl meter inaccuracies do not matter much, hence i tried the mobile phone app. I will try and get my hands on a good microphone and soundcard from a friend in the recording business. do you think that will work? I don't want to invest in a sound meter. my purpose is to correct my room and get done with it.

Also, do you think measuring the speakers from nearfield will give a better understanding of this graph also. Dali Ikon 6 are not very well known for their low bass. They are supposed to be midbass champs.

 

[theredcommando]

Have a look at -

Equal-loudness contour - Wikipedia, the free encyclopedia
and also,
A-weighting - Wikipedia, the free encyclopedia

You will need to consider this before you make any conclusions(apart from speaker response, mic accuracy).
The freq vs db graphs you usually see are A weighted.

 

[anm]

how reliable is the measurement with the android phone? When the output from speakers is 100db, does the phone measure 100, or 80? How can we be sure? Sorry for the noob question.
Also, the frequency response of the microphone may not be linear, so say it measures 100db for 100db, but may measure 80db when the actual volume is 90db?

 

[Thad E Ginathom]

I'd love to calibrate the software on my phone against an SPL sometime. Until then, the closest I can get is this trick: on the diy-hearing-test page of audiocheck.net. They think we can get within 10bD

The other standard warning for using phones as SPL meters is that the mics are made for human-voice frequency ranges.

 

[aashish351]

Thad: thanks for the link, but it is not working

10 dB is a huge variation is it not? I mean a -6dB or +6dB in a particular bass frequency can have a great affect on the listening experience.

 

[Thad E Ginathom]

The link works for me; if not for you, go to www . audiocheck . net and then Audiogram Hearing Test under "what's new."

Warning: do not go anywhere near that site unless you have an hour to spare :lol:

(nb, the site needs flash enabled to actually play the test sounds and tones when you click the buttons, otherwise they may load in an external media player. Took me a couple of visits to work this one out.)

 

[sidvee]

. I will try and get my hands on a good microphone and soundcard from a friend in the recording business. do you think that will work? I don't want to invest in a sound meter. my purpose is to correct my room and get done with it.

Also, do you think measuring the speakers from nearfield will give a better understanding of this graph also. Dali Ikon 6 are not very well known for their low bass. They are supposed to be midbass champs.

Yes, if you can get a properly calibrated microphone and accessories it should give a accurate representation of your room curve. Also nearfield measurement of speakers, IMO is useless, unless you are a speaker designer, because what matters most is the acoustics at the listening position that takes into account all the surrounding room surfaces - unless of-course you are sitting at 4 to 5 feet away from the speakers - which would be nearfield listening.
Finally correction in your case may prove challenging, (hopefully the measurements are erroneous), because bass peaks can be dealt with by listening position changes, bass traps, etc., but how to retrieve information that has been lost in bass suckouts? Anyways before you get to this step, you have to make sure your measurement system is accurate.
Cheers,
Sid

 

[Hari Iyer]

Measurements like these are always at the near field @ 1w/1m preferably in an anechoic room. If you do not have an anechoic room then the SPL meter should be very close to the speaker cone to prevent room reflections. These are all small signal measurements with pink or white noise (even a computer based signal generator will do). You have plenty of analyzers available as a free-ware to measure the speakers response. But the trick is to use the right sound card and the calibration mic (with the calibration files) to measure these signals. The mic can be a major culprit here in the measrurements. The measurement can vary a lot with the position of the mic changing and hence the mic should be very close to the speaker for minimizing losses and reflections.

 

[aashish351]

The link works for me; if not for you, go to www . audiocheck . net and then Audiogram Hearing Test under "what's new."

Warning: do not go anywhere near that site unless you have an hour to spare :lol:

(nb, the site needs flash enabled to actually play the test sounds and tones when you click the buttons, otherwise they may load in an external media player. Took me a couple of visits to work this one out.)

Thanks... the link worked later.

Sent from my Nexus S using Tapatalk 2

 

[aashish351]

Yes, if you can get a properly calibrated microphone and accessories it should give a accurate representation of your room curve. Also nearfield measurement of speakers, IMO is useless, unless you are a speaker designer, because what matters most is the acoustics at the listening position that takes into account all the surrounding room surfaces - unless of-course you are sitting at 4 to 5 feet away from the speakers - which would be nearfield listening.
Finally correction in your case may prove challenging, (hopefully the measurements are erroneous), because bass peaks can be dealt with by listening position changes, bass traps, etc., but how to retrieve information that has been lost in bass suckouts? Anyways before you get to this step, you have to make sure your measurement system is accurate.
Cheers,
Sid

My idea for nearfield measurement was to compare speaker response curve to room response curve in order to assess whether the low bass is because of room or the setup is not producing it at all. Is that correct thinking?

Sent from my Nexus S using Tapatalk 2

 

[Hari Iyer]

After you have measured the room response of the speakers in your room, you will require an Equalizer for EQ the response of the speakers to your room. Let me know once you have done this as i have a very simple excel based solution for EQ from your 20 channel foobar player if you intend to use your computer, foobar player and DAC for your music.

 

[aashish351]

that's great hari... i am using a usb dac. please send me the excel file and i will do the EQ and measure the response again and put up the graph.

 

[sidvee]

My idea for nearfield measurement was to compare speaker response curve to room response curve in order to assess whether the low bass is because of room or the setup is not producing it at all. Is that correct thinking?

Sent from my Nexus S using Tapatalk 2

Actually I would skip this step directly to room response and figure out how to smoothen/correct the room response. Regardless if you want to compare the 2 response curves then it is fine, but in your case it is almost certain that the room is causing the anomalies.
Cheers,
Sid

 

[sidvee]

DALI Ikon 6 loudspeaker Measurements | Stereophile.com
Here are the measurements of the dali ikon6 speaker in an anechoic setting. fig. 4 is the measurement that corresponds to your plot. As you can see in the measurement, the speaker's bass drops off steeply from 60 hz and is almost -5db at 50 hz. Sort of consistent with your findings, but beyond that peaks and valleys of your graph are significantly different from the smooth response shown in the measurement graph, and that is where your room/listening position/speaker placement etc. is effecting the curve.
Cheers,
Sid

 

[aashish351]

Thanks sidvee... that was great help. Now the most difficult task is to figure out how improve the bass drop between 60 and 100 and flatten the curve as much as possible.

On a different note, subjective listening does not suggest such a big drop between 60 and 100. So it could be an effect of phone's Mic not absorbing the frequencies below 100 properly.

However, I do feel on many occasions that the bass has peaks and dips in my room because some bass notes sound louder and some weaker specially when a cello piece is being played in a jazz song.


Sent from my Nexus S using Tapatalk 2

 

[sidvee]

Now the most difficult task is to figure out how improve the bass drop between 60 and 100 and flatten the curve as much as possible.

On a different note, subjective listening does not suggest such a big drop between 60 and 100. So it could be an effect of phone's Mic not absorbing the frequencies below 100 properly.

Responding to your second point first, IMO you have to make sure that what is measured is correct before you consider step 1. If subjective listening is not correlating to the measurements, then that is a red flag right there.
Room acoustic correction is really a combination of subjective listening and objective measurement, and a blend of both will be required to arrive at your optimal preference, which might vary based on personal perferences. I have been in rooms where the curve at 30-40hz has had 10db peaks! and the listener preferred it that way (the word headbanger comes to my mind:lol.
Once you are reasonably sure about your measurements, consider moving the listening position in small increments back and forth and re-measure. Though this may not get rid of all suckouts it may help the worst offenders which are nulls in your listening position.
Cheers,
Sid

 

[aashish351]

Actually I am not so unhappy with the bass response being lower by say -3db or -6db. Dali's have a very fast, tight & lean bass and I usually get just that, except when only bass notes are playing and two subsequent notes at different frequencies / octaves have variation in dB level. I am more interested in smoothening that curve - removing peaks and suck-outs rather than increasing the overall bass response by 6 or 10 db!! Have read at Realtraps and other places that Bass Traps not only help peaks reduction but also improve something called a "bass null".

However, one thing is absolutely clear that first doubt on measurement has to be removed.

 

[sidvee]

I am more interested in smoothening that curve - removing peaks and suck-outs rather than increasing the overall bass response by 6 or 10 db!! Have read at Realtraps and other places that Bass Traps not only help peaks reduction but also improve something called a "bass null".

Yes bass traps in room corners will help greatly in smoothening out the bass response curve, but they cost money (you will need minimum 2 at the corners behind the speakers and 4 in all corners for the best effect). Meanwhile you can also try different listening positions and also speaker positions if you can, since these are just free tweaks (there are various recommendations on the net). In fact for optimum bass reponse, the listening position in the room, IMO is the most important variable as is the positioning of the speakers. Anyways it is commendable that you are taking pains for objective measurements rather than rely on subjective listening alone.
Cheers,
Sid