Klipsch RP-160m vs Phil BMR.

[speakerman39]

In addition to the Omni mic thing that Dennis mentioned, I think to get a better picture we really need to see 1/24 or even 1/48 smoothing.

Peng, would you mind to take a few minutes and explain the differences between 1/24 and/or 1/48 smoothing? Which is better and more accurate? Still a bit confused as to what the differences are, if any, between various scales of smoothing. It would be most appreciated. Thanks!


Cheers,

Phil

 

[yepimonfire]

Did not realize the omni mic had that issue. Glad you brought that to my attention.

Sent from my LM-X210(G) using Tapatalk

 

[PENG]

Peng, would you mind to take a few minutes and explain the differences between 1/24 and/or 1/48 smoothing? Which is better and more accurate? Still a bit confused as to what the differences are, if any, between various scales of smoothing. It would be most appreciated. Thanks!


Cheers,

Phil

According to the definition per Wiki: "In music, an octave (Latin: octavus: eighth) or perfect octave is the interval between one musical pitch and another with half or double its frequency."

So 1/48 octave smoothing is more accurate than 1/24 octave smoothing because it will be the moving average of twice as many data as it will be for 1/24 octave. If you want the FR plots to look smooth, you would choose 1/12 or even 1/6 smoothing but resolution will suffer accordingly.

AH has an article for every question an audioholic may ask, so here's one for your question:

https://www.audioholics.com/loudspeaker-design/audio-measurements

 

[speakerman39]

According to the definition per Wiki: "In music, an octave (Latin: octavus: eighth) or perfect octave is the interval between one musical pitch and another with half or double its frequency."

So 1/48 octave smoothing is more accurate than 1/24 octave smoothing because it will be the moving average of twice as many data as it will be for 1/24 octave. If you want the FR plots to look smooth, you would choose 1/12 or even 1/6 smoothing but resolution will suffer accordingly.

AH has an article for every question an audioholic may ask, so here's one for your question:

https://www.audioholics.com/loudspeaker-design/audio-measurements

Thanks for the info Peng. I am a bit more clear now. Also, thanks for the link. Will be sure to read it. It is most appreciated.


Cheers,

Phil

 

[TheWarrior]

According to the definition per Wiki: "In music, an octave (Latin: octavus: eighth) or perfect octave is the interval between one musical pitch and another with half or double its frequency."

So 1/48 octave smoothing is more accurate than 1/24 octave smoothing because it will be the moving average of twice as many data as it will be for 1/24 octave. If you want the FR plots to look smooth, you would choose 1/12 or even 1/6 smoothing but resolution will suffer accordingly.

AH has an article for every question an audioholic may ask, so here's one for your question:

https://www.audioholics.com/loudspeaker-design/audio-measurements

However it makes it difficult to identify the center frequency of a bass resonance. But of course that applies to sounds below transition, above, 1/12 is sufficient.

Short of the most analytical usage in an anechoic chamber, 1/48 is simply too much information.

 

[PENG]

However it makes it difficult to identify the center frequency of a bass resonance. But of course that applies to sounds below transition, above, 1/12 is sufficient.

Short of the most analytical usage in an anechoic chamber, 1/48 is simply too much information.

Didn't Floyd Toole recommended 1/20 octave though, according to the AH article I linked?

 

[PENG]

Thanks for the info Peng. I am a bit more clear now. Also, thanks for the link. Will be sure to read it. It is most appreciated.


Cheers,

Phil

You are welcome, Gene has done such a good job acquiring a wealth of AV knowledge on this site so we should all take advantage of it.

 

[speakerman39]

You are welcome, Gene has done such a good job acquiring a wealth of AV knowledge on this site so we should all take advantage of it.

I agree 100% Peng. Gene and Hugo have both worked very hard on educating forum members/visitors. We are very lucky to be quite candid here. Thanks once again!


Cheers,

Phil

 

[TheWarrior]

Didn't Floyd Toole recommended 1/20 octave though, according to the AH article I linked?

Yes, and he elaborates in his book (which I really think you should get) that when it comes to dealing with resonances below transition, anything below 1/20 (1/24 is REW) can mask the size of the resonance and given insufficient data.

Above transition, EQ is to be used sparingly, which is why he goes in to great detail about what is, and how to define a 'good' loudspeaker that should need little/no EQ to sound 'good' in a given room. When you're looking at extremely high resolution measurements above a few hundred hertz, there is little sense to be made of the measurement, because you are measuring room reflections.

 

[PENG]

Yes, and he elaborates in his book (which I really think you should get) that when it comes to dealing with resonances below transition, anything below 1/20 (1/24 is REW) can mask the size of the resonance and given insufficient data.

Above transition, EQ is to be used sparingly, which is why he goes in to great detail about what is, and how to define a 'good' loudspeaker that should need little/no EQ to sound 'good' in a given room. When you're looking at extremely high resolution measurements above a few hundred hertz, there is little sense to be made of the measurement, because you are measuring room reflections.

Haha, I don't need to own that book to know he said those things but I will eventual buy the book, when I know I will have time to read. I think he was just making the point that it is futile to try and EQ the higher frequencies. That does not mean one should never attempt. I, and probably @Pogre too, appear to belong to the group that would want to EQ whatever the software/hardware can do well.

The fact is, and being fact I am sure Dr. Toole would agree, that if one sits in one spot and do not move his head when listening to music, those bouncy waves will hit him/her quite consistently if not totally consistently. That means in theory, if you can EQ things to the flattest possible FR you can hear the effect, until you move your head then all bets are off. Of course I am making a point that has no practical meaning, making it simply for argument sake. To be more practical, those smart PhDs (Chris K for sure, had said it, sort of) all talked about trying to collect a lot of data and create a small bubble, that within it one can hear the "best" possible FR. Not knowing how the software and algorithm function it is not possible for us to know if such were just talks, or at least true to a point. In theory though, it is certainly possible.

Now back to manual EQ that's what I am doing with the help of REW, I am definitely sticking to EQ'ing up to about 800 Hz. I pick that frequency based on my crawling (speakers, subwoofer, furniture, mic, and me) experience and findings. After spending hours, including my one hour that turned into 2 hours, I was able to gain about 2 to 5 dB in several dips on the right channel while less successful on the left channel. I cannot do better than that because of there is a practical/physical limit for how much I can move them.

The phases at the dips happens to be roughly the sames between the left and right, so I am only to EQ them separately and see/hear what happens. After I applied the EQ (in JRiver) by just looking at the FR, without running the EQ program, I already heard significant improvements.

 

[Pogre]

Haha, I don't need to own that book to know he said those things but I will eventual buy the book, when I know I will have time to read. I think he was just making the point that it is futile to try and EQ the higher frequencies. That does not mean one should never attempt. I, and probably @Pogre too, appear to belong to the group that would want to EQ whatever the software/hardware can do well.

The fact is, and being fact I am sure Dr. Toole would agree, that if one sits in one spot and do not move his head when listening to music, those bouncy waves will hit him/her quite consistently if not totally consistently. That means in theory, if you can EQ things to the flattest possible FR you can hear the effect, until you move your head then all bets are off. Of course I am making a point that has no practical meaning, making it simply for argument sake. To be more practical, those smart PhDs (Chris K for sure, had said it, sort of) all talked about trying to collect a lot of data and create a small bubble, that within it one can hear the "best" possible FR. Not knowing how the software and algorithm function it is not possible for us to know if such were just talks, or at least true to a point. In theory though, it is certainly possible.

Now back to manual EQ that's what I am doing with the help of REW, I am definitely sticking to EQ'ing up to about 800 Hz. I pick that frequency based on my crawling (speakers, subwoofer, furniture, mic, and me) experience and findings. After spending hours, including my one hour that turned into 2 hours, I was able to gain about 2 to 5 dB in several dips on the right channel while less successful on the left channel. I cannot do better than that because of there is a practical/physical limit for how much I can move them.

The phases at the dips happens to be roughly the sames between the left and right, so I am only to EQ them separately and see/hear what happens. After I applied the EQ (in JRiver) by just looking at the FR, without running the EQ program, I already heard significant improvements.

All I know is, every time I play with REW and my mini I learn more. I'm getting better at dialing in a flat response with fewer and lighter adjustments. If you took away my DSP I know it wouldn't sound as good as it does now. Before DSP there was a lot of bass left on the table that I was just not hearing before.

I'm still gonna try to isolate the sub sections of my towers and see what I can make happen. Re routing my main speakers (full range) through the DSP turned up some interesting and not bad results, but I lost a little in the lower frequencies. I like to play with it because it gives me something to do, has absolutely improved my FR and sq in a positive way and I learn more every time.

 

[PENG]

All I know is, every time I play with REW and my mini I learn more. I'm getting better at dialing in a flat response with fewer and lighter adjustments. If you took away my DSP I know it wouldn't sound as good as it does now. Before DSP there was a lot of bass left on the table that I was just not hearing before.

I'm still gonna try to isolate the sub sections of my towers and see what I can make happen. Re routing my main speakers (full range) through the DSP turned up some interesting and not bad results, but I lost a little in the lower frequencies. I like to play with it because it gives me something to do, has absolutely improved my FR and sq in a positive way and I learn more every time.

Just a silly question, have you been EQ'ing left and right independently or you do it for both as a whole?

 

[Pogre]

Just a silly question, have you been EQ'ing left and right independently or you do it for both as a whole?

Both as a whole at the end of the day. Tho I've taken measurements individually and noticed that when I do both as a whole my FR suffers (before eq) from the interaction between both transducers? Just guessing, but I definitely have a smoother response measuring independantly, before any eq'ing.

My thought process is, I listen to both as a whole so I should eq both as a whole.

 

[PENG]

Both as a whole at the end of the day. Tho I've taken measurements individually and noticed that when I do both as a whole my FR suffers (before eq) from the interaction between both transducers? Just guessing, but I definitely have a smoother response measuring independantly, before any eq'ing.

My thought process is, I listen to both as a whole so I should eq both as a whole.

I think it depends on a few things. Even though you listen as a whole, as long as you eq each individually to the same target curve you should be fine too. I have been doing so deep thinking about this and may be my next step is to PM JM (author of REW) over at the HTS. Or TheWarrior could do me a favor and email Dr. Toole or Dr. Olive, assuming he has a direct line to them.

 

[TheWarrior]

Haha, I don't need to own that book to know he said those things but I will eventual buy the book, when I know I will have time to read. I think he was just making the point that it is futile to try and EQ the higher frequencies. That does not mean one should never attempt. I, and probably @Pogre too, appear to belong to the group that would want to EQ whatever the software/hardware can do well.

The fact is, and being fact I am sure Dr. Toole would agree, that if one sits in one spot and do not move his head when listening to music, those bouncy waves will hit him/her quite consistently if not totally consistently. That means in theory, if you can EQ things to the flattest possible FR you can hear the effect, until you move your head then all bets are off. Of course I am making a point that has no practical meaning, making it simply for argument sake. To be more practical, those smart PhDs (Chris K for sure, had said it, sort of) all talked about trying to collect a lot of data and create a small bubble, that within it one can hear the "best" possible FR. Not knowing how the software and algorithm function it is not possible for us to know if such were just talks, or at least true to a point. In theory though, it is certainly possible.

Now back to manual EQ that's what I am doing with the help of REW, I am definitely sticking to EQ'ing up to about 800 Hz. I pick that frequency based on my crawling (speakers, subwoofer, furniture, mic, and me) experience and findings. After spending hours, including my one hour that turned into 2 hours, I was able to gain about 2 to 5 dB in several dips on the right channel while less successful on the left channel. I cannot do better than that because of there is a practical/physical limit for how much I can move them.

The phases at the dips happens to be roughly the sames between the left and right, so I am only to EQ them separately and see/hear what happens. After I applied the EQ (in JRiver) by just looking at the FR, without running the EQ program, I already heard significant improvements.

The phrase Toole uses regarding EQ above transition in both the book and video is 'cheating'. If you start with good speakers, there is a lot less to deal with. Both the video and book go in to great detail on what defines a 'good' loudspeaker, and price is not the essential component.

Flat has never been the goal, smooth is. Bass rises, and as frequency increases, there is typically a downward tilt in a room curve as high frequencies attenuate through air. There is no 'target curve', every room is different. Otherwise we'd all be posting pictures of ruler flat FR measurements.

The room and all speakers function together as the play back system. Measuring speakers individually, in room, is of little meaning to the quality of that play back system. Toole discusses rooms as two different sound fields. Taking physical measurements of the room to predict room modes (to know what peaks/nulls are modes, and which ones, just - are) and finding a happy balance between positional EQ and electronic EQ could yield far better results while saving you time, in regards to sound below transition. Above, 'good' speakers and attention to detail of a room's furnishings in and around the listening area are all that should be needed, save for basic tone controls due to the variability of recordings.

If you're looking for answers, I will continue to recommend the same book to anyone, 'Sound Reproduction'. The answers are in there, you just need to be willing to find them, even if they go against your firmly held beliefs. Science works whether you believe in it or not. If Audyssey had scientific workability, you wouldn't have to keep remeasuring or applying additional automated EQ.

BTW (I think @Pogre has the 3rd edition, but for anyone the companion site is live! http://routledgetextbooks.com/textbooks/_author/toole/

Back on topic - I am looking forward to comparing the Phil 3 with the BMR next week, and might as well use the Klipsch too - pretty easy as we prefer to compare in mono anyway. @KEW

@Dennis Murphy I saw somewhere you are using the NRC for measurements, have you done this with the Phil 3? It would be interesting to compare that data with what I come up with from my outdoor CTA2034 measurements!

 

[KEW]

@Dennis Murphy I saw somewhere you are using the NRC for measurements, have you done this with the Phil 3? It would be interesting to compare that data with what I come up with from my outdoor CTA2034 measurements!

My understanding was that Dennis is generally not using NRC for measurements.
Rather, he sent one speaker to NRC for measurement so he could validate his measurement methods for consistency with the NRC.
Dennis would have to answer if it was the Phil3.

 

[MrBoat]

Strange all of this talk with EQ'ng speakers to room and this is what is new to me and where I fall through the cracks with my old habits. I have never sought to EQ capable speakers in capable rooms. I was instead, EQ'ng the source. It's still the same here, and perhaps more so now that a lot of speakers are more revealing.

Pink noise and other gauges don't give a damn about less than audiophile quality recordings. Then I may see what's in some people's CD rack and find very little of what would be considered even close to audiophile quality. So then, is all of this rounding and averaging for a few good sources?

Weirder yet is, trusting all of this testing for speakers, without having a way to test the sources. It seems there would have to be a 3rd party grading system that did this in which to validate such things beforehand to make all of this work. Otherwise, to say that a great speaker, or a sufficiently EQ'd one plays everything equally well seems completely erroneous to me.

 

[Dennis Murphy]

Yes, and he elaborates in his book (which I really think you should get) that when it comes to dealing with resonances below transition, anything below 1/20 (1/24 is REW) can mask the size of the resonance and given insufficient data.

Above transition, EQ is to be used sparingly, which is why he goes in to great detail about what is, and how to define a 'good' loudspeaker that should need little/no EQ to sound 'good' in a given room. When you're looking at extremely high resolution measurements above a few hundred hertz, there is little sense to be made of the measurement, because you are measuring room reflections.[/

My understanding was that Dennis is generally not using NRC for measurements.
Rather, he sent one speaker to NRC for measurement so he could validate his measurement methods for consistency with the NRC.
Dennis would have to answer if it was the Phil3.

Right--I just sent an early version of the BMR. it would be prohibitively expensive to ship a Phil 3 up there and back on top of the hefty fee they charge for testing, and I don't think the nice guy who runs the chamber would be thrilled when he saw the boxes. I also have no idea how the bass would measure given the issues I encountered with the BMR.

 

[TheWarrior]

Right--I just sent an early version of the BMR. it would be prohibitively expensive to ship a Phil 3 up there and back on top of the hefty fee they charge for testing, and I don't think the nice guy who runs the chamber would be thrilled when he saw the boxes. I also have no idea how the bass would measure given the issues I encountered with the BMR.

Well then I will email you my data, hopefully in the next month or two. Measurements are only the first part, the rest is spreadsheets - weeeee!

One advantage to testing outdoors is that I don't have to compensate for the standing waves of an anechoic chamber so they should be looking quite good, down low!

 

[TLS Guy]

Don't know about the Audyessey Mic, but when I measure the Klipsch using my OmniMic, I get pretty much exactly what the OP got. And my BMR measures way down in the highs with the Omni, which it isn't, unless the NRC and Sound and Vision mics are off in the same direction and to the same degree as mine. I've had discussions with at least 4 other Omni Mic owners, and they all have the same problem. Jim Salk's is so far off in the highs that he can't use it for anything.

Yes, my omni mic does the same thing. It rolls off at 15 KHz. However on he plus side it has no peaks, I'm certain.

Even very expensive mics are prone to aberrations in the top end.

Here is the famous Neumann U 87. I have the dual capsule stereo version. This is the omni response. This is one of the most widely used microphones of all time.

The Shure SM81

The discontinued SM80 omni which I have is similar.

The mics with the flattest extended responses are by Earthworks. They make measurement and studio mics.

The reference measurement microphone is the Earthworks M30, but it costs $700.00.

Omni mic is handy. If you are using a good tweeter and can rely on the manufacturers data like SEAS for instance, I can tolerate a 15 KHz roll off.