something for you big speaker haters: http://www.fostex.co.jp/int/pages/products/spunits/unitpdf/fw800n.pdf
I don't know what Dan mounted the mass with, but I think mounting it to the dust cap was a good idea in that it would dampen the cone resonances the least. Were you implying that the dust cap is not part of the mass?Kuei Yang Wang said:
The added mass curve exactly lays on the free air curve on the rising edge. It differs on the falling edge due to the mass. That's the physics of it. Look at a tweeter vs any woofer impulse response. Did I miss your point entirely?
I think you'll have to be the one to run the field strength test. I don't have one and Dan showed us his side of the argument. ;^)
"Practically" does not belong with "absolute and infinite", would you agree? As engineers would say, "close enough". You would be hard pressed to find a difference that was meaningful.Kuei Yang Wang said:
I don't know what other methods you are refering to but it is of the utmost importance to correcty match the level of the test signals. One example would be the difference between a MLS measurement and a swept or even stepped sine measurement. Windowing and the type of window and more importantly how they are applied also make a big difference in the accuracy.Kuei Yang Wang said:
I work for the Test Systems Design Center at a Defence contractor and I can tell you the pen and paper analyzer would not be my choice.
If you heard a difference with 2 systems that were EQ'd to the same response curve and same LF extension, I have no explanation. Assuming the HF was equally matched of course, and you were able to switch back and forth relatively quickly.Kuei Yang Wang said:
Russ
The answer sits in the waterfall diagram.
Kind of similar to ringing in digital filter response, worser even, pre- and post ringing have influences imho: a note can sound different on its attack depending on the previous one.
But I bet KYW knew that!
A music player doesn't play individual notes, waiting for one to die out before playing the next.
Kind of similar to ringing in digital filter response, worser even, pre- and post ringing have influences imho: a note can sound different on its attack depending on the previous one.
But I bet KYW knew that!
A music player doesn't play individual notes, waiting for one to die out before playing the next.
Konnichiwa,
Not quiet. But it looks to me as if the mass was not rigidly coupled to the Cone/Voicecoil assembly, or as I wrote "decoupled". The measured result certainly fit's the hyphotesis.
Yup.
Well, I have to say that as long as measurements show little to no congruence to how something sounds I must suggest that measurements tell you rather exactly and detailed something not worth knowing.
Nor would it be my first choice in this day and age.
However, the point remains that swept sinewaves and impulses analysed via FFT tend to give widely differing results and noisloading another set of differing results. I tend to go with noiseloading, as the test signal most closely approaches music in it's nature. The Pen/Moving Paper Strip analyser systems are still widely in use for response testing in production envoironments (my DL-103 cartridge came with a response "proof" generated thusly) and give the right answer to the right question...
All the above requirements where fulfilled. And yes, there where audible differences and NOT ONLY related to room modes (these where equalised as well anyway).
Sayonara
russbryant said:
Not quiet. But it looks to me as if the mass was not rigidly coupled to the Cone/Voicecoil assembly, or as I wrote "decoupled". The measured result certainly fit's the hyphotesis.
russbryant said:
Yup.
russbryant said:
Well, I have to say that as long as measurements show little to no congruence to how something sounds I must suggest that measurements tell you rather exactly and detailed something not worth knowing.
russbryant said:
Nor would it be my first choice in this day and age.
However, the point remains that swept sinewaves and impulses analysed via FFT tend to give widely differing results and noisloading another set of differing results. I tend to go with noiseloading, as the test signal most closely approaches music in it's nature. The Pen/Moving Paper Strip analyser systems are still widely in use for response testing in production envoironments (my DL-103 cartridge came with a response "proof" generated thusly) and give the right answer to the right question...
russbryant said:
All the above requirements where fulfilled. And yes, there where audible differences and NOT ONLY related to room modes (these where equalised as well anyway).
Sayonara
I've not looked at the impulse data with regards to mass so I'm not sure of how the data doesn't match. Can you elaborate? Never too late to learn you know!Kuei Yang Wang said:
Haven't heard of noiseloading unless it goes by another name? I normally use MLS data for the upper frequencies and splice in stepped sine measurements for the lower frequencies then add close mic'd data to see the lowend rolloff. MLS data by itself does not contain enough valid data in the lower frequencies by the time you gate out the reflections. As you know it takes great care to get reliable below 200 hz or so. Few take the time to do it and fewer still take the time to do it right.Kuei Yang Wang said:
Russ
Konnichiwa,
In my view we the the sharp initial spike from radiating surfaces that are "decoupled" from the added mass, hence no material difference is observable. The usual approach of using blue tack or the like to "add mass" will give such results. To be realistic with these measurements (and measure what you wish to measure) you would have to build two versions of the same driver, one with a drastically heavier cone, all else being equal or you need to use a Fieldcoil driver whcih allows the manipulation of the "B" element of the BL factor directly.
Sorry, slip of tounge/key. Noiseloading is more applicable to Amplifiers.It uses actually comb filtered white noise. I meant to use pink noise.
Sayonara
russbryant said:
In my view we the the sharp initial spike from radiating surfaces that are "decoupled" from the added mass, hence no material difference is observable. The usual approach of using blue tack or the like to "add mass" will give such results. To be realistic with these measurements (and measure what you wish to measure) you would have to build two versions of the same driver, one with a drastically heavier cone, all else being equal or you need to use a Fieldcoil driver whcih allows the manipulation of the "B" element of the BL factor directly.
russbryant said:
Sorry, slip of tounge/key. Noiseloading is more applicable to Amplifiers.It uses actually comb filtered white noise. I meant to use pink noise.
Sayonara
regarding added mass on the woofer dome:
The stiffness of the dome is not absolute and certainly not in both moving directions.
Regarding low frequency measurements:
just a proposal
1) measure the reverbation after impuls exitement of the speaker, we're interested in the tail (reverb), not the initial impuls, memorise it.
2) We exite the system with 2 impulses, the first to get the room exited, the second superposed on the reverbation of the first, ie with a delay of the impulse time of the first impuls.
3) idem as 2) but with a delay of first + 2 times the second impuls.
So shifting each time the delay of the second impuls.
4) idem as 2) but delay of first impuls + 3 times the second impuls.
etc.
this way your window stays narrow, not influenced by anything else but the dying out reverbation wich is known by 1) wich is a large window measurement.
This is a pure theoretical presentation of something I never tried myself. But it may be a way to see the frequency behaviour in a reverbant room.
If you get what I expect: a sequence of frequency (waterfall)diagrams that 'waves' in time like the crowd in a footballstadion. The waves amplitude not only depending on reverbation of the room but also on the dampening ability of the speaker (waterfall diagram of speaker superposed to waterfall diagram of the room)
I don't know variable timeshifting of the measuring window can be done with programs like clio etc.
Eventually the first impuls can be replaced by noise (MLS, pink, white) to wake up all room modes, as long as you have an accurate room-waterfall diagram.
In fact, human ears tend to notice not only the sound character of the instrument but also the kind of space it is played in, independant one from the other.
Sayonara
The stiffness of the dome is not absolute and certainly not in both moving directions.
Regarding low frequency measurements:
just a proposal
1) measure the reverbation after impuls exitement of the speaker, we're interested in the tail (reverb), not the initial impuls, memorise it.
2) We exite the system with 2 impulses, the first to get the room exited, the second superposed on the reverbation of the first, ie with a delay of the impulse time of the first impuls.
3) idem as 2) but with a delay of first + 2 times the second impuls.
So shifting each time the delay of the second impuls.
4) idem as 2) but delay of first impuls + 3 times the second impuls.
etc.
this way your window stays narrow, not influenced by anything else but the dying out reverbation wich is known by 1) wich is a large window measurement.
This is a pure theoretical presentation of something I never tried myself. But it may be a way to see the frequency behaviour in a reverbant room.
If you get what I expect: a sequence of frequency (waterfall)diagrams that 'waves' in time like the crowd in a footballstadion. The waves amplitude not only depending on reverbation of the room but also on the dampening ability of the speaker (waterfall diagram of speaker superposed to waterfall diagram of the room)
I don't know variable timeshifting of the measuring window can be done with programs like clio etc.
Eventually the first impuls can be replaced by noise (MLS, pink, white) to wake up all room modes, as long as you have an accurate room-waterfall diagram.
In fact, human ears tend to notice not only the sound character of the instrument but also the kind of space it is played in, independant one from the other.
Sayonara
you know watt/puppy v?
they've the most beautiful waterfall response I ever saw (dead in no time), but in the same time are awefully, even legendary difficult to place correctly in a room .
The previous post is the explanation for this.
Since with these loudspeakers the room modes (waterfall) are very dominant in the restitution of music, compared to the speaker own waterfall diagram, it's very important to cancel room modes as much as possible in order to get the optimum holographic picture out of these speakers.
With less acurate speakers, the mixture of speakers/room waterfalls gives a more foggy picture, but with the advantage of being more forgivinly in placement in the room.
Horn speaker systems are also taking advantage of this theory because with them, direct radiated sound is more dominant over reflected sound. The penalty with these systems is that the sweet spot for listening is much smaller.
they've the most beautiful waterfall response I ever saw (dead in no time), but in the same time are awefully, even legendary difficult to place correctly in a room .
The previous post is the explanation for this.
Since with these loudspeakers the room modes (waterfall) are very dominant in the restitution of music, compared to the speaker own waterfall diagram, it's very important to cancel room modes as much as possible in order to get the optimum holographic picture out of these speakers.
With less acurate speakers, the mixture of speakers/room waterfalls gives a more foggy picture, but with the advantage of being more forgivinly in placement in the room.
Horn speaker systems are also taking advantage of this theory because with them, direct radiated sound is more dominant over reflected sound. The penalty with these systems is that the sweet spot for listening is much smaller.
Digging up the past, I know, but could you explain how you calculated 231Hz? My calculations used a rough approximation where the cone represented an antinode (point of maximum air velocity) and an opposing wall of the enclosure represented a node (fixed point). That way a quarter of a wavelength would be sufficient to cause a resonance in the box.Kuei Yang Wang said:
Since your calculation results in a higher frequency than my calculation despite a longer box, I assume you calculated the lowest box resonance as being 1/2 of a wavelength between 2 opposite walls? Since the speaker is the source of the sound in the box, therefore most of the resonances will be interdependent, rather than independent of the speaker.
Re: 31 inch subwoofer. Ok, now, can we see one with a cone break-up frequency above 70Hz ?
Kuei Yang Wang said:
Sorry but I can't agree with this. Your saying that the added mass, which more than doubles the moving mass, is so completely decoupled by the mounting method that in effect it's not even there. Like the man said "I gots to get me some of that!"
Kuei Yang Wang said:
I use pink noise for RTA at the listening position but not for driver measurement.
Russ
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.