I have done a bunch of measurements of stock, phase plugged and EnABLed FE167E's. There is way too much into to post here, so I am giving this link to a web page:
http://www.geocities.com/rbrines1/Pages/Articles/EnABL/EnABL.html
Have fun!
Bob
http://www.geocities.com/rbrines1/Pages/Articles/EnABL/EnABL.html
Have fun!
Bob
An externally hosted image should be here but it was not working when we last tested it.
I'm missing the point of why the smoothing of 1/3 octave is necessary and how the omitted data is 'trash'. Ha maybe I'm just sick of all the super-smoothie plots you get from Fostex and everyone else. For once, I'd like to see the real deal. But you're taking the initiative and doing the work so I applaud that. Would measuring them outside offer a cleaner plot compared to in the garage?
I just can't see how the driver could be within 5db 1K to 20K.
I just can't see how the driver could be within 5db 1K to 20K.
Inclined Plane,
I think Bob's 5 dB criterion is within plus or minus 5 dB so it is a 10 dB window.
My initial view of these plots convince me that Enabling doesn't significantly improve the performance unless your criteria is to have lower output (sensitivity) level and more rolloff once you go off axis. Phase plugging shows imrovements in certain frequency ranges but overall it is not a clear winner. No doubt that a listener will hear sound differences between these modifications.
Bottom line for me is that a 'standard' unit is flatter and better off axis vs. either modification. Save your money and time and if you want a better driver purchase a better one.
Jim
I think Bob's 5 dB criterion is within plus or minus 5 dB so it is a 10 dB window.
My initial view of these plots convince me that Enabling doesn't significantly improve the performance unless your criteria is to have lower output (sensitivity) level and more rolloff once you go off axis. Phase plugging shows imrovements in certain frequency ranges but overall it is not a clear winner. No doubt that a listener will hear sound differences between these modifications.
Bottom line for me is that a 'standard' unit is flatter and better off axis vs. either modification. Save your money and time and if you want a better driver purchase a better one.
Jim
It looks like the FR differences are from the phase plug installation... from the difference curves, both the phase plugged drivers track fairly tightly. Certainly within sample-to-sample variation.
Jim, that is only a valid statement if a FR measure as Bob made actually tells you everything about a driver's performance. Certainly that is a very short-sighted view... i view this as akin to saying you know everything about the ocean by looking only at the surface.
It is very important to know what the driver is doing 10, 20, 30, 40 dB down from this signal in its precense. I've yet to see any measures that tell us anything about this.
dave
Jim, that is only a valid statement if a FR measure as Bob made actually tells you everything about a driver's performance. Certainly that is a very short-sighted view... i view this as akin to saying you know everything about the ocean by looking only at the surface.
It is very important to know what the driver is doing 10, 20, 30, 40 dB down from this signal in its precense. I've yet to see any measures that tell us anything about this.
dave
Well. frequency response curves are a nice way to imagine what a speaker might sound like. Unfortunately, they do not provide an indication of information content. As an example, a ribbon tweeter will have much the same frequency response as other tweeters, but in reality, the information content is off the scale in comparison to any other form of driver. Even comparisons of off axis frequency response do not provide any better clue as to the information content that you, as a listener, can make, sense of.
A very famous designer of some very famous speakers (think Bozak) once commented that no driver he had every encountered allowed coherent information content below -40db down from the reference signal. Everything below that floor was unintelligible. This is where EnABL does it's immediately apparent work. After you get over hearing sound that is clearly audible and comprehensible as far down in dynamic range as your electronic equipment can provide, the next thing that becomes apparent is that all of the rest of the information you are hearing is complete. Piano notes are exactly like they are in real time, same strike, same carry through and the same decay as you find in reality. Transients have internal tones, decay structure and sound like the original event, rather than a whack or a smack without any further descriptors available.
EnABL does this without really affecting the frequency response of the driver, unless we use it to subdue those resonance nodes that are so prominent in CSD plots of full range drivers.
Looking at published data is not in any way going to inform you of what EnABL provides, period. You will have to experience it for yourself before you have any idea of what the rest of us are talking about.
You might look here to get a clearer presentation of what I am providing here, which I am fairly sure will raise your hackles due to it's absolutist tone.
http://www.diyaudio.com/forums/showthread.php?postid=1883833#post1883833
Bud
A very famous designer of some very famous speakers (think Bozak) once commented that no driver he had every encountered allowed coherent information content below -40db down from the reference signal. Everything below that floor was unintelligible. This is where EnABL does it's immediately apparent work. After you get over hearing sound that is clearly audible and comprehensible as far down in dynamic range as your electronic equipment can provide, the next thing that becomes apparent is that all of the rest of the information you are hearing is complete. Piano notes are exactly like they are in real time, same strike, same carry through and the same decay as you find in reality. Transients have internal tones, decay structure and sound like the original event, rather than a whack or a smack without any further descriptors available.
EnABL does this without really affecting the frequency response of the driver, unless we use it to subdue those resonance nodes that are so prominent in CSD plots of full range drivers.
Looking at published data is not in any way going to inform you of what EnABL provides, period. You will have to experience it for yourself before you have any idea of what the rest of us are talking about.
You might look here to get a clearer presentation of what I am providing here, which I am fairly sure will raise your hackles due to it's absolutist tone.
http://www.diyaudio.com/forums/showthread.php?postid=1883833#post1883833
Bud
Very interesting set of plots Bob -many thanks.
One general observation WRT the FR plots is that I wouldn't draw too many conclusions about EnABL per se from these. I gather the 167eN does not just have the EnABL pattern applied to a plugged stock driver, but several other mods too, so it's really a comparison in the frequency domain between stock, stage 1 & stage 2 general mods.
On that basis, both the plugged and fully modded units appear on-axis to be somewhat smoother in detail, although the general trend of the stock unit may end up slightly flatter. I'd expect both to sound smoother & better controlled than the stock driver, purely on this basis -can't speak for the rest of course.
One general observation WRT the FR plots is that I wouldn't draw too many conclusions about EnABL per se from these. I gather the 167eN does not just have the EnABL pattern applied to a plugged stock driver, but several other mods too, so it's really a comparison in the frequency domain between stock, stage 1 & stage 2 general mods.
On that basis, both the plugged and fully modded units appear on-axis to be somewhat smoother in detail, although the general trend of the stock unit may end up slightly flatter. I'd expect both to sound smoother & better controlled than the stock driver, purely on this basis -can't speak for the rest of course.
InclinedPlane said:
Indeed you are missing the point.
If I didn't smooth these spaghetti charts, they would be unreadable.
Do you really want to see all of the cabinet defraction effects and the reflections within my garage?
We are looking at the DIFFERENCES between three different cone treatments. The actual performance of the stock driver is irrelevant.
The measurements were done with a 2k FFT. That in itself smooths the data, and I didn't feel the need to take the time to increase the data points.
Finally, you brain is going to smooth what you hear to something like 1/3 octave (actual number debatable). Unless narrow spikes are very large, you won't here them.
But since you ask,
An externally hosted image should be here but it was not working when we last tested it.
Bob
Right, in general terms, the unsmoothed data isn't especially practical to work with. If you want really raw, see attached Clio measurement.
This was of a 167 MLTL I did a few years back. 1m, on axis, room heavily damped for the measurements. Driver had a phase plug; otherwise stock.
This was of a 167 MLTL I did a few years back. 1m, on axis, room heavily damped for the measurements. Driver had a phase plug; otherwise stock.
Attachments
I'm going to let this thread ride a day or two before I post MY impressions of this study. I expected Bud and Dave to be somewhat defensive, and they have a point -- almost.
How many times have you heard "You can't tell how a speaker sounds from a frequency plot"? For the untrained, this is true. But, this is not my first dance. Almost all of the voicing aspects of speakers appear in the time and/or frequency plots. Micro-detail and "speed" -- whatever that is -- do not. When I here something in a speaker, I what to see what is producing that particular effect. When I first started installing phase plugs, I heard a smoothing of the sound. When I looked at the plots, I found that the phase plugs put a hole in the FR where sibilance occurs. In effect, the phase plugs changes an artifact in the human voice.
Those who know what they are looking at will understand why EnABLing is perceived as an improvement in the sound of a speaker, and in my experience, there is a 50/50 split between "smooth" and detail. BTW, my impressions of the SOUND of EnABL vs stock FE167E is http://www.diyaudio.com/forums/showthread.php?s=&threadid=147611HERE .
Bob
How many times have you heard "You can't tell how a speaker sounds from a frequency plot"? For the untrained, this is true. But, this is not my first dance. Almost all of the voicing aspects of speakers appear in the time and/or frequency plots. Micro-detail and "speed" -- whatever that is -- do not. When I here something in a speaker, I what to see what is producing that particular effect. When I first started installing phase plugs, I heard a smoothing of the sound. When I looked at the plots, I found that the phase plugs put a hole in the FR where sibilance occurs. In effect, the phase plugs changes an artifact in the human voice.
Those who know what they are looking at will understand why EnABLing is perceived as an improvement in the sound of a speaker, and in my experience, there is a 50/50 split between "smooth" and detail. BTW, my impressions of the SOUND of EnABL vs stock FE167E is http://www.diyaudio.com/forums/showthread.php?s=&threadid=147611HERE .
Bob
Here is a link to the only CSD plot that I think shows EnABL activities on a driver. It is a meticulously done comparison by John K, not a supporter of all of our hand waving claims. You must wait a moment for it to load and then roll your pointer over the graph and then back off of it.
http://planet10-hifi.com/johnK-test/
In the most general sense, there is no difference. This echoes my findings from tests run with my Liberty Audio test set up, with dedicated computer, calibrated mic and echo DSP sound card with the needed modifications for low frequency stability.
EnABL on it's own, had not shown me a measurement that would be considered any greater in difference than those normal to a run of drivers. Until John K's blink comparison.
I would be pleased to have any or all of you describe what differences you see and what sonic differences you would expect. All comments will be equally welcome and I will add mine after some period of time.
Bud
http://planet10-hifi.com/johnK-test/
In the most general sense, there is no difference. This echoes my findings from tests run with my Liberty Audio test set up, with dedicated computer, calibrated mic and echo DSP sound card with the needed modifications for low frequency stability.
EnABL on it's own, had not shown me a measurement that would be considered any greater in difference than those normal to a run of drivers. Until John K's blink comparison.
I would be pleased to have any or all of you describe what differences you see and what sonic differences you would expect. All comments will be equally welcome and I will add mine after some period of time.
Bud
Bob Brines said:
Yes, let's allow tempers to settle a bit before digging too deep, but if I may opine pedantically on the semantics - "almost"? - one either has a point or not - its veracity or your particular agreement with it is another matter.
Where in the recording/reproduction chain do you think this specific "artifact in the human voice" arises? If such a problem can be mitigated for some listeners, either mechanically/acoustic or electronically (EQ, digital room correction, etc.), whether objectively verifiable by current measurements (or our understanding of captured data), how "wrong" can that be?
This is a phenomena that is particularly troublesome for me with female vocals. I happen to have had the occasion to hear a female trio (Wailin Jennies) live in a small local venue ( Alix Golden Hall, a converted church) a couple of times. Their performance includes much completely acoustic and several acappella numbers, with none of the percussive vocal sibilance I think you're referring to. When they step up to the mikes, it's a different story - is it just poor P/A equipment or sound engineering - who knows for sure? So far as I'm aware there have been no live recordings of these particular events, but the studio recordings can display some of the sibilance issues you describe on some, but not all systems. In fact, these tracks are included on my system test playlist, and can reveal aspects of imaging - stability, spatial separation, etc., often buried in more complex, conventionally over-produced multi-tracked recordings.
At the end of the day, I think we're in general agreement that the EnABL process (and in the case of the FE167, phase plugs as well) can elicit an audible "improvement" in the musical listening experience. When that can be quantified and reliably packaged or manipulated, then these arguments will become moot.
Some folks hear more than others beyond simple smoothing of FR or relaxation of presentation, and on a wider range of drivers and systems than have been discussed in this thread.
oops, I guess I didn't take my own advice
I have no problem with this statement, at all. A generic EnABL only treatment just does not alter the frequency plots to a degree I find indicates an "improvement". Frequency plots tell all there is to know about balance, "voicing", what to expect for false emphasis and how reliable a reproduction you can expect of the original event, with respect to ...frequency response.
A CSD FFT processing of the frequency response provides even more information about signal decay, but all of it provides only first arrival, largest amplitude characteristics. This is not a negative, it is good information and crucial to the design of any driver and the use of those drivers.
EnABL changes none of those characteristics to an appreciable degree, unless we have used it to treat a specific resonance node. If just the generic EnABL pattern, used on all drivers as a base line treatment, is applied, what you find in John K's blink comparison is what you will always get.
Scott brings up a good point when he says that an EnABL only treated driver needs to be in those plots you have. Without any question, the alterations in frequency response, other than removal, and I do mean complete dispersal, of a resonance node are due to Dave's processes.
I do not mean to pick an argument here. There is no basis for an argument. There is also nothing to defend, EnABL only does what it does and it will do the same amount and kind of thing it does, regardless of how you alter the frequency response with physical alterations to the driver.
Bud
Last edited:
Ty, Scott and Bob. It's amazing how you can have such drastic differences in output with frequencies a dozen (or less!) hertz apart. How lucky for us that, as you say, narrow spikes aren't especially audible. I wonder what the unsmoothed plot would look like with Enabl AND cone coating treatments.
Bob
Well done Bob,
that shows a lot of wish ful thinking,
and remembers me Wiki:
http://de.wikipedia.org/wiki/Phase-Plug
the critics.
Well done Bob,
that shows a lot of wish ful thinking,
and remembers me Wiki:
http://de.wikipedia.org/wiki/Phase-Plug
the critics.
G'day dave,
Do you think Mark might be prevailed upon to do some stereo recordings - one for each driver?
Then we can compare and interpret using ears rather than eyes...
Cheers,
Alex
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.