PDA

View Full Version : Do the CBM-170s "lack top-end extension?"



Eddie
05-13-2005, 10:50 PM
In this thread: http://forum.ecoustics.com/bbs/messages/1/136535.html

One poster posted this link:

http://www.soundstagemagazine.com/measurements/ascend%5Fcbm170/

and interpreted the numbers like this:

"Edster, if you look at the frequency response graph, you'll notice the -3db points are approx. 80hz in the low end and 10Khz on the top end. The bass response drops very quickly below 100hz, which may be a good thing, since 99.9% of all homes on the planet are not able to do really low bass anyway. What is somewhat bothersome is the lack of top end extension. Although not much music is above 10khz there are harmonics that reach those ranges frequently. The midrange looks very smooth. There is some noise at approx 50hz, probably from the port that seems to increase as spl increases. Just posted it for your benefit. Nice to see your speakers performance, no?"

Agree/disagree? Comments?

I'm fairly clueless around this much techie stuff so would like to hear your thoughts on this...

davef
05-13-2005, 11:41 PM
In this thread: http://forum.ecoustics.com/bbs/messages/1/136535.html

One poster posted this link:

http://www.soundstagemagazine.com/measurements/ascend%5Fcbm170/

and interpreted the numbers like this:

"Edster, if you look at the frequency response graph, you'll notice the -3db points are approx. 80hz in the low end and 10Khz on the top end. The bass response drops very quickly below 100hz, which may be a good thing, since 99.9% of all homes on the planet are not able to do really low bass anyway. What is somewhat bothersome is the lack of top end extension. Although not much music is above 10khz there are harmonics that reach those ranges frequently. The midrange looks very smooth. There is some noise at approx 50hz, probably from the port that seems to increase as spl increases. Just posted it for your benefit. Nice to see your speakers performance, no?"


Hi Eddie,

Just read that thread on ecoustics. I am really not too sure what that person is talking about.. First of all, the on-axis frequency response graph clearly indicates dead FLAT to 20kHz.. The graph he decided to base his comments on is the 30 degrees off-axis response, at which point the CBM-170 is indeed ONLY about -3dB at 20Khz. That is quite remarkable actually :)

Regarding bass response... this measurement was taken in a true anechoic chamber. Imagine listening to a speaker in a wide open field. -3dB in that particular anechoic chamber is very respectable.

Either this person made an honest mistake by posting comments regarding the 30-degree off-axis response, or he really has no experience at all evaluating loudspeaker measurements.. or he is just trying to criticize a speaker he knows nothing about....

The CBM-170 is a remarkable +/- 1.5 dB from 90Hz to 20kHz, as proven by the most respected loudspeaker measurement authority on the planet, Canada's NRC.... For some unknown reason to me, the person who posted that graph on ecoustics left off the text, which clearly says:


Top curve: on-axis response
Middle curve: 15 degrees off-axis response
Bottom curve: 30 degrees off-axis response

This person was evaluating the bottom curve....

Hope this helps...

shane55
05-13-2005, 11:56 PM
Well... here's how I interpret these graphs (and I'm no expert).

On axis, these are virtually flat from (I'll call it) 80hz to about 20k.
It's 'off axis' where things fall apart a bit... not in the low end, but in the highs. To 30 degrees they still remain pretty true, but after that (45 and up), there is quite a drop.

Now, one thing for certain... these on, or near axis are amazing speakers. They are dead-neutral and sound great! Off axis, all speakers are going to sound worse, and I mean ALL speakers.

Check out the others on this link:
http://www.soundstageav.com/speakermeasurements.html

You'll understand what I'm talking about. Each speaker represented falls apart at greater angles off-axis. It's the nature of the beast.

That said... remember. These graphs represent numbers, which represent how the speaker 'performs'. It does not truly represent how it 'sounds'.

Cheers.

shane

Eddie
05-14-2005, 10:22 PM
Dave,

Thanks for clearing that up. Yeah, I thought there was something out of wack there, esp. since I've never been anything but very happy with how my Ascends deliver both mids AND highs.

Eddie
05-14-2005, 10:23 PM
Shane,

> That said... remember. These graphs represent numbers, which represent how the speaker 'performs'. It does not truly represent how it 'sounds'.

This is especially true for someone as technically challenged as myself! :D

S_rangeBrew
05-16-2005, 01:42 PM
Off axis, all speakers are going to sound worse, and I mean ALL speakers.


Hence, B ose inventing Direct Reflecting speakers, which minimize this problem. I wonder what Dave F. thinks of this idea, and would he ever design it into any of his speakers? :)

bikeman
05-16-2005, 02:10 PM
Hence, B ose inventing Direct Reflecting speakers, which minimize this problem.

In theory they minimize this problem. In reality, they did anything but. Thrirty four years ago I owned 901's. I was young and naive. It's been a long road to the Ascend's and I definitely have learned something along the way. ;)

David

soundfreak38
05-18-2005, 06:49 PM
Hey, I read on some site that it is better to post a response curve using ranges of 5hz as opposed to 10 or greater. The site claims by doing this you can get a much more realistic idea how a loudspeaker performs. Is this true? That is, using ranges 10 hz and up makes the response curve to appear more flat?? Just curious if this were true.
Phil

davef
05-18-2005, 11:43 PM
Hi Soundfreak,

Thought I would answer this for you...


Hey, I read on some site that it is better to post a response curve using ranges of 5 Hz as opposed to 10 or greater. The site claims by doing this you can get a much more realistic idea how a loudspeaker performs. Is this true? That is, using ranges 10 hz and up makes the response curve to appear more flat?? Just curious if this were true.

I believe what you are referring to is called resolution. For example, measure the loudspeaker at 20Hz, 30Hz, and 40Hz would be 10Hz resolution. Measuring the loudspeaker at 5Hz intervals 20Hz, 25Hz, 30Hz... Would be 5Hz resolution. The better the resolution, the more revealing the response measurement would be.

The measurements we use (and post), as well as the measurements taken at the NRC use an FFT size of no less than 16384. That is 16,384 measurement points taken between 20Hz and 20kHz. This translates to a resolution of 1.2Hz, or one measurement taken every 1.2Hz. Approximately 4 times greater resolution than a 5Hz interval. Even more interesting, there is occasion when I will use an FFT size of 32768, about .5 Hz resolution! This is about as revealing as it gets...


One additional comment I would like to make regarding the off-axis response of loudspeakers.. Some of you mentioned that it is a problem or a flaw with how the higher frequencies roll-off the further off-axis you are.. I can assure you, this is a GOOD thing and most desirable... Remember, the measurements you are evaluating and basing your thoughts on are anechoic measurements... free of reflections.. A speaker that it is truly non-directional in an anechoic chamber will have a sound power response that is essentially flat = All frequencies that reach your ear have equal energy.

This is undesirable and not how we perceive sound. Within the bandwidth of human hearing, we perceive high frequencies as having more weight (more energy) then lower frequencies. This is because each octave has twice as many frequencies as the octave preceding it. This is why white noise sounds like high frequency "hiss" even though the level of each frequency from 20Hz to 20kHz is the exact same. Pink noise, on the other hand, is weighted so that each octave is reduced from the preceding octave (each octave has the same weight, equal energy etc.) and we perceive this to sound more like "full-range", even though on a graph, pink noise would look like a downward angled line starting from 20Hz and ending at 20kHz, while white noise would be "flat as a board"....

A speaker that delivered truly flat on-axis and off-axis response (regardless of listening angle) would sound very, very bright (completely undesirable) in a typical room. In designing loudspeakers (at least quality loudspeakers) off-axis dispersion is carefully controlled and optimized to deliver an overall flat response at the listening position. You must remember that when listening in a typical room, we hear a combination of the on-axis and off-axis response, regardless if we are sitting on-axis or off-axis.

Here is a more useful graph than an on-axis response. This is the "listening window" response of the CBM-170 taken at the NRC.. on- and off-axis measurements are averaged together, which is more representative of what we "hear" while listening to the loudspeaker.


http://www.soundstagemagazine.com/measurements/ascend%5Fcbm170/frequency_listeningwindow.gif

Response curve is an average of five measurements:
on-axis, 15 degrees left and right off-axis,
15 degrees up and down off-axis

Notice how this graph is even flatter than the on-axis and off-axis response of the CBM-170??? This graph is an amazing +/- 1dB from about 85Hz to 18kHz... This is the flattest listening window response of any loudspeaker I have ever had the pleasure to measure... and let me say I have been measuring loudspeakers for a long, long time ;)

Don't want to talk about this too much as I am getting dangerously close to revealing some of my personal and well guarded design techniques :eek:

Sorry for my long ramble, just thought you would all find it interesting..

GaryB
05-18-2005, 11:54 PM
Sorry for my long ramble, just thought you would all find it interesting..What gave you that idea? :D

Great post, Dave - I now finally understand the difference between pink and white noise. And my CBMs sound pretty good, too! ;)

shane55
05-18-2005, 11:57 PM
Ditto...

Interesting is an understatement. Thanks for the great detailed explanation.:D

shane

curtis
05-19-2005, 12:19 AM
Dave...thanks for the great info!

I, for one, don't want you to compromise your design techniques.....unless I ever decide to DIY, which I highly doubt will ever happen. :)

bikeman
05-19-2005, 05:15 AM
Dave...thanks for the great info!

I, for one, don't want you to compromise your design techniques.....unless I ever decide to DIY, which I highly doubt will ever happen. :)


With enough beer, I could be talked into a DIY sub but there's no way I'm going to build speakers. Dave's job is safe. ;)

David

soundfreak38
05-19-2005, 08:45 AM
Yeah Dave thank you very much for your extremely detailed explanation. You are the man Dave!!! Just wish I knew what you forgot about loudspeaker design.......lol. Anyways, keep up the good work.
Phil

metalaaron
05-19-2005, 10:01 AM
i would submit that there are several loudspeakers that lack top end extension, and the ascends are not one of them.

there is a bit of a double edged sword to posting measurements, but the rationale dave provides towards design choices along with the data from 3rd parties makes perfect sense for ascend's approach.

jimsiff
05-19-2005, 10:56 AM
Dave,

Thanks a lot for helping us understand. It's nice to see you're as accurate and revealing as the speakers you design. :D

It would be really nice if somebody would/could collect your educational posts and put them in a sticky FAQ thread. I know this was talked about awhile ago, but I don't know if anything came of it.

curtis
05-19-2005, 11:41 AM
It would be really nice if somebody would/could collect your educational posts and put them in a sticky FAQ thread. I know this was talked about awhile ago, but I don't know if anything came of it.
I'll see if I can work on a DaveF "knowledgebase". :)

Now that I think about it, just searching on Dave's posts, I bet most of them are educational.

Lee Bailey
05-19-2005, 12:34 PM
DaveF, I've sent you a couple of PMs that I have yet to see a response on. :(

davef
05-19-2005, 05:34 PM
Hi Lee,

My apologies... I will be responding to the messages shortly. For future correspondance, please email me directly.