REW Graphic analysis

[Tony Teng]

Hi John:
I use REW generator 5K,10,15,20K HZ and play with my Macbook. meanwhile I record with RTA and the plot are as attached. The peak at individual frequency have similar dB( see attached). Then I copy each frequency wav files to my pocket SSD and play them by Cambridge 851N network player + burmester ampifier+monitor audio gold 300 speakers. The RTA plots show great different. It may be the reason why my REW file always roll-off the high frequency section?

 

[jtalden]

Are you sure the speaker switches are set to "mono-pole"?

Per the speaker manual (excerpt attached) the front tweeter is disabled when set to "di-pole". The crossover to the tweeter is at 2.3 kHz so your graph is consistent with the tweeter not working.

 

[ddude003]

Are you sure the speaker switches are set to "mono-pole"?

Per the speaker manual (excerpt attached) the front tweeter is disabled when set to "di-pole". The crossover to the tweeter is at 2.3 kHz so your graph is consistent with the tweeter not working.

View attachment 9581

Good catch jtalden... I was wondering about single wiring per page 9...

 

[Tony Teng]

I am using 2 channel system, not 5.1 or 7.1
So, monopole or dipole may not be the issue.
regarding single wiring, I try three difference wiring( as attached pdf). Still make no different.

 

[ddude003]

I am using 2 channel system, not 5.1 or 7.1
So, monopole or dipole may not be the issue.
regarding single wiring, I try three difference wiring( as attached pdf). Still make no different.

I am sorry to have led to any confusion... To be more clear...

As jtalden suggests... Please ensure the speakers switches are set to "mono-pole"... If the 300 has this option, this is the correct position for a 2 channel system.

Please ensure the two "terminal links" (the short wires connecting the HF and LF terminals) are in place per page 9...

 

[Tony Teng]

The engineer of monitor audio dealer came today and find both tweeters of speakers was out of order. They are confuse why both broken at the same time. They will send new tweeters from UK to replace it.

 

[ddude003]

The engineer of monitor audio dealer came today and find both tweeters of speakers was out of order. They are confuse why both broken at the same time. They will send new tweeters from UK to replace it.

I am not saying that this is your problem and may be something to consider... A classic symptom of an underpowered system employing a solid state amplifier is fried tweeters... Clipping and Distortion can play havoc with speakers... The amp appears to be spec'ed at 240 watts @ 4 ohms... The speakers @ 8 ohms... If the standard math applies to this amp, and it may not depending on design, @ 120 watts for 8 ohms is near the bottom of what your speakers require... If Clipping were to occur your tweeters could suffer...

 

[Eric SVL]

I am not saying that this is your problem and may be something to consider... A classic symptom of an underpowered system employing a solid state amplifier is fried tweeters... Clipping and Distortion can play havoc with speakers... The amp appears to be spec'ed at 240 watts @ 4 ohms... The speakers @ 8 ohms... If the standard math applies to this amp, and it may not depending on design, @ 120 watts for 8 ohms is near the bottom of what your speakers require... If Clipping were to occur your tweeters could suffer...

It only takes 57 watts to drive those speakers to 95dB at 3 meters...I do not believe they are under-powered.

 

[AJ Soundfield]

The engineer of monitor audio dealer came today and find both tweeters of speakers was out of order. They are confuse why both broken at the same time. They will send new tweeters from UK to replace it.

That looks like a true ribbon (won't mention manufacturer), so I'm a bit surprised by the lowish 2.3KHz specified crossover employed. I assume when you bought the speakers, they sounded fine.
Looks like somewhere they were blown during sweeps, which you should be careful with volume initially. Sweeps contain a lot of energy that can stress tweeters at high volume. Your amp is plenty powerful. True ribbons are susceptible to failure at even modest power levels, so use care.
Music will be less stressful and less likely to damage.
Since you should be primarily concerned with cutting bass peaks via EQ and not touching the frequencies > 500hz or so, next time, you can limit your sweeps in REW (on the "Measure" screen) to around 800Hz. You will get your info and protect your speakers.

btw, your speakers are monopoles, there is no dipole switch, etc., different surround models.
Oh and beautiful setup ;-)

cheers,

AJ

 

[Tony Teng]

The speaker was installed 30th April 2018. The high frequency was not crystal and the bass was not powerful from the first beginning. I thought that may because of new speaker without fully run-in. But after 1 month, the sound is still wrong. So, I told the dealer came to with Omnimac and test result is as attached. I, at least, don't know what the plot means at that time. The dealer said no problem and told me to adjust speakers position to get better sound. Later I knew REW + MiniDSP may do the same job like OmniMac and I start learning how to analysis my room and gears with you guys' great help. The first plot I did is as attached (20180807 REW).The measurement is not 100% correct. And with your help, I finally did it right at 20180814. I really appreciated what all of you had taught me. I am not sure that didI burn the tweeters during testing?(I did it at relative low volume less than 95dB). When the dealer finally found the tweeters are broken on 30th august, we can not believe both of them was gone at the same time. It is almost impossible but it just happened. Before the new tweeters arrived, I hope to hear some more opinions from all of you to avoid another disaster.

 

[Tony Teng]

By the way. I borrowed a pair of Monitor Audio Gold 50 bookshelf speakers(with same tweeter as gold 300) and plot as attached.

 

[Tony Teng]

Today I try to reposition my gold 300 speakers. At the first 15 time measurements, 30 Hz is still a mode. But later,something strange happen. The low frequency below 50Hz were gone! ( comparison as attached) Did I burn my bass?

 

[Tony Teng]

I made a comparison test of Gold 50 today. The plot is similar to 20180831(gold 50). I really worry about my speakers were damaged during testing.

 

[John Mulcahy]

What SPL are you doing your tests at Tony? 75 dB SPL is plenty for room measurements.

 

[Tony Teng]

Hi John:
I don't know where to read the SPL during my test. Attached is the parameters I set for the measurement (most of them are default). I always ckeck level before measurement.

 

[John Mulcahy]

I believe you have a UMIK-1, so if you select the UMIK as the Input Device and Microphone as the input on REW's soundcard preferences the REW measurements will show actual SPL. When measuring the sweep should be at a level you could comfortably listen to.

 

[John Mulcahy]

As an aside, that Omnimic measurement shows the same treble roll-off seen with your REW measurements.

 

[Tony Teng]

Regarding the tweeters problem, I start to reposition my speakers and analysis by REW.
While moving forward & backward, the measurement is as plot 1( 110CM,130CM,150CM,170Cm & 190CM mean distance from speaker face to the rear wall)
While moving left & right, the measurement is as plot 2( 55cm,60cm & 65cm mean distance from speaker face to the side wall)
My question is Which position have better performance?

 

[Tony Teng]

The mode at 30Hz exist no matter what position I move. My plan is to make a helmholtz resonator to absorb the mode. But, how about the null at 40~50 Hz ?

 

[John Mulcahy]

That 30 Hz is probably the first length mode. If you measure half way down the room it may disappear, but that isn't likely to be a practical place either to sit or to locate your speakers, so absorption on the front and/or rear walls (tuned or broadband) is your main option, assuming you don't wish to apply some EQ. There looks to be a small dip at 40-50 Hz rather than a null (hard to say for sure without seeing an unsmoothed plot, nulls are very sharp and very deep). It is likely mid way between the first and second length modes and so reflects the lack of modal boost rather than a dip as such.

 

[Tony Teng]

Finally I finished 2 helmholtz resonators and put them in my audio room. I try several different location to find the best result. And I rear wall seems the best place so far.
I make a comparison with and without HR as plot. the peak at 30HZ slightly go down.
I will try to put some sponge inside the HR to see any different. Do you think the HR work at my room?

 

[John Mulcahy]

Looks to be making a difference. Note that even very small amounts of polyester wadding can broaden the tuning of the resonators.

As an aside, there appears to be a board underneath the microphone in your picture. Surfaces near the microphone generate reflections, affecting the results especially at high frequencies.

 

[Matthew J Poes]

John is right on, I would ditch that wood panel.

The Helmoltz resonators are quite a bit more effective once you add some damping to the inside chambers. You might even play with different types. Mineral wool, poly batting, and recycled cotton are all good options. Fiberglass works but is not great for air quality. Damping the box sides with bitumen panels or CLD can also sometimes be good. These things are fairly sensitive to the rigidity and damping of the enclosure.

I would consider building more of these and maybe consider making them physically larger. Also modify the tuning slightly so they are about 3-5hz higher and loswer by slightly changing the length of the ports. You want as much poet surface area as possible for a given tuning to get he most effect.

I have to be honest, I’m very impressed with that result. You nailed the tuning frequency. I rarely see it that perfect and it’s not uncommon for the effect to be barely noticeable. Great work!

 

[Tony Teng]

I tried to cover the openings of HR with different weave fabric or different openings.
I found cover the top two opening or low two opening have significant difference as plot.
For me it looks like a trade off , while peak at 30hz was smooth, by 40-50Hz get worse.
Which curve perform better?

 

[Matthew J Poes]

I tried to cover the openings of HR with different weave fabric or different openings.
I found cover the top two opening or low two opening have significant difference as plot.
For me it looks like a trade off , while peak at 30hz was smooth, by 40-50Hz get worse.
Which curve perform better?

If you have the capacity to apply eq then the red plot is best. You can’t really eq a dip but you can eq a peak. On the other hand, peaks are more noticeable than dips and so if you are working passively and can’t eq, anything that reduces the peak is best.

I would start with some listening tests though. Those really low peaks probably are t that audible or objectionable. I would listen and see what sounds best, rather than focus too much on those plots.