ZDL

[el`Ol]

This is where I see most driver manufacturers are treating incorrectly by using damping material on diaphragms. What happens is that the soft diaphragms, although well damped, also absorbs some energy transferred from the VC, thus much low level detail is lost in the process. This is one reason why well damped soft material diaphragmed drivers sound quite "dry" when in a room that is also quite dead. A more adequate approach is to use stiff diaphragm, but use the surround to damp it out properly. The skill is how to handle this properly. Ted Jordan's drivers are good examples of utilizing cone curvature profile to do this.

Could the reason be not damping per se, but non-linear damping behaviour, i.e. level-dependent damping? I have no idea whether that's really the case, but usually nothing in polymers is linear. Interesting in this context are the P2C cones, which measure very much like polypropylene, but don't sound boring or muffled at all. This was my impression when I listened to Blumenhofer loudspeakers and Vance Dickason also describes a very open sound.
http://www.p2ccone.de/pdf/test_Vance_Dickason.pdf
On the German Physiks website they argue that using the surround for damping the cone is quite impossible. Where we are talking about German Physiks: Do you think the horror-prices are justified (from the production point of view)?

 

[john k...]

Hi John, probably if you show a driver operating from 1KHz up, it would be more convincing.
My understanding is that energy stored mechanically cannot be solved by anything electrically feeding it. As for UE, I have not done any software debugging efforts, and am not able to do so as I have done on flight safety related software, so I don't think we have enough evidence what exactly the issue is. I will be increasing my system memory to see if the problem goes away, but to ask everyone to use a specific hardware is not reasonable. If every media playing software did that, this market would not expand as it has. That said, It is well known that I am a fan of SoundEasy mainly because it does things in a way I have found very usefull for speaker development. I tend to look at UE as a more end user oriented software, the concept is very good, and I will support it as long as it also satisfies my needs.

There are a couple of issues. First, stored energy and or linear distortion arises from two factors; amplitude distortion and phase distortion. You can stick a mic out and measure the acoustic output of a system and at that point you have the amplitude and phase response. If the system is minimum phase then they are one and the same and minimum phase amplitude equalization will correct it. If the system is not MP, then amplitude eq is required to correct amplitude errors and phase correction required to correct the phase aspects. It matters not how things at the observation point get to the state they are in, all that matters is that this is a form of linear distortion and linear corrections fix it. The caveat is that sound radiation is 3-dimensional so the fix may not work at all positions. This is why there is an increasing interest in controlled or constant directivity systems. If the system is constant directivity over some given listening window then the corrections will apply over that window as well.

Re the UE, I run it perfectly on my XP machine which is a 2.8gHz P4 with 512 meg RAM. It uses about 30% CPU. When I use it with the medial player running, either through analog out/in or with VAC 3 the total load is about 40%. In both cases I'm using less that 50% of the available memory. I agree that it should not be necessary to lock the UE to specific hardware but if it is required to assign outputs to specific channels of the sound card and the sound card drivers don't allow that it isn't a UE problem. Obviously this is something we need to continue to look at. I really don't know why you are having so many problems. It does not jive with Bohdan's or my experience or some of the others who have tested the UE.

Also, remember that most media software is not trying to assign output channels specifically. Your typical 7.1 sound card simply decodes the 7.1 signal and sendes it on the path elected by the device manufacture when he writes the drivers. Multi-channel sound cards for mixing and recording are also just that as well. I think one place to start with your sound card is to look at what can be accomplished within the Mixer/router that came with the sound device.

ANyway, let's not hijac this thread. Enough about the UE, ok?

 

[Joachim Gerhard]

At Essex we used a phase linear version of a forth order Linkwitz crossover. We found that higher order crossovers, although easily implemented in the digital domain, gave interference off axis. The new Grimm speaker uses a linear phase version of a forth order Linkwitz too out of the same reasons. The work at Essex preceded that by more then 15 years though.
How did it sound ? Unspectacular and natural.

 

[Joachim Gerhard]

Here is the original paper about digital equalization of loudspeakers at Essex University.
http://www.essex.ac.uk/csee/research/audio_lab/malcolmspubdocs/J19 Efficient filter design.pdf

 

[soongsc]

Could the reason be not damping per se, but non-linear damping behaviour, i.e. level-dependent damping? I have no idea whether that's really the case, but usually nothing in polymers is linear. Interesting in this context are the P2C cones, which measure very much like polypropylene, but don't sound boring or muffled at all. This was my impression when I listened to Blumenhofer loudspeakers and Vance Dickason also describes a very open sound.
http://www.p2ccone.de/pdf/test_Vance_Dickason.pdf
On the German Physiks website they argue that using the surround for damping the cone is quite impossible. Where we are talking about German Physiks: Do you think the horror-prices are justified (from the production point of view)?

The concept sounds very interesting, seems like another composite type structure. Since the properies of such material is unknown, I really cannot make any judgement.
Whether surround can damp a cone or not depends on both cone and surround design since the two are made of different material. This is where Jordan drivers do better then most metal drivers. I have listened to German Physics speakers briefly, but cannot remember which model it was. The impression was, they certainly disappeared in the room they were played in, the sound was realistic but the background details seemed lacking. Everyone has a different idea what is worth what, it's like my wife never fancied LV purses, etc. It's not that she can't afford it either.

 

[Joachim Gerhard]

Here is another paper about digital equalisation and a passage about reduction of energy storage.
http://www.essex.ac.uk/csee/research/audio_lab/malcolmspubdocs/C72 Active loudspeakers.pdf

 

[Joachim Gerhard]

I many cone speakers there certainly is a problem when the cone meets the surround.
Usually you see this as a ragged response in the 1kHz area. This can be solved with edge coating or using a highly damped surround. Both measures have audible consequences
though. The Scan Speak wideband i will use does not suffer from that problem. The cone is a open weave and to make it air tight it is coated with a propriatary material that also gives the cone it´s acoustical properties. What i also like is the very low moving mass.

 

[soongsc]

There are a couple of issues. First, stored energy and or linear distortion arises from two factors; amplitude distortion and phase distortion. You can stick a mic out and measure the acoustic output of a system and at that point you have the amplitude and phase response. If the system is minimum phase then they are one and the same and minimum phase amplitude equalization will correct it. If the system is not MP, then amplitude eq is required to correct amplitude errors and phase correction required to correct the phase aspects. It matters not how things at the observation point get to the state they are in, all that matters is that this is a form of linear distortion and linear corrections fix it. The caveat is that sound radiation is 3-dimensional so the fix may not work at all positions. This is why there is an increasing interest in controlled or constant directivity systems. If the system is constant directivity over some given listening window then the corrections will apply over that window as well.

...

Never wanted a UE discussion here, I certainly intend to measure when I get it running. If it shows what you say, I will admit I have learned something new. If improvement in stored energy results, the same trend will be noticeable regardless whether measured on axis or not. From an AES paper I have read, it does seem that certain directivity pattern supports certain listening window as well as how the speakers are orientated.

 

[Joachim Gerhard]

The off axis response is as important as the on axis response and in the old times it was measured on a turntable. It usually was an expensive B&K system and the operators knew what they did. When MLSSA came at the end of the 80th that has changed and the focus was on CSD waterfalls and acoustic phase measurements for a long time. Some time ago the focus changed again to off axis performance. Isn´t it weird how the way we measure influences our opinions on what sounds good ?

 

[soongsc]

I many cone speakers there certainly is a problem when the cone meets the surround.
Usually you see this as a ragged response in the 1kHz area. This can be solved with edge coating or using a highly damped surround. Both measures have audible consequences
though. The Scan Speak wideband i will use does not suffer from that problem. The cone is a open weave and to make it air tight it is coated with a propriatary material that also gives the cone it´s acoustical properties. What i also like is the very low moving mass.

The effective moving mass will change with frequency as the diaphragm starts flexing. This is why wide band drivers work the way they are. I would be interested in your final listening impressions.

 

[Joachim Gerhard]

Yes, i think that is the case in the Scan. In a certain way it is a bending wave driver with a high coincident frequency. I will try to avoid that region though. That is one reason why i will cross to a tweeter. I have already build a dipole speaker with 3 Scan Speak widebands where they work without tweeter in the front. It was presented on the ETF Festival in France this year. I have a "helper" tweeter in the back to arive at my "G-Pole" radiation pattern though. Part of that work you can see on my MPL thread. The Scan without tweeter is one of the best widebands i heard so far but there is a resolution boundary very high up in the treble. You get used to that rather quick though. Nobody during the ETF Festival complained about the treble. Sometimes a little blending in the treble helps to tame agressive material but in the ZDL i do not want to hide information.

 

[soongsc]

Thank you for the papers Joachim, it would be very interesting to see what I will hear and measure. I think different types of measurements have different strengths and weaknesses. It all depends on how we use them.

 

[Joachim Gerhard]

Yes sure, but we should listen too. For me unamplified life music in a good acoustic is still the reference. See a picture from the back of my G-Pole at ETF. The helper tweeter is not connected yet. At lower frequencies all drivers radiate and from 300 Hz up the outer drivers are filtered out so that the middle driver works only at higher frequencies.

 

[ThorstenL]

Hi,

But have you actually listend to music through that kind of setup? Can you describe the audible difference?

I did a job a long time ago at a friends studio (mid towards late 90's, when DSP based crossovers etc. first became viable).

We first activated his PMC Monitors using the PMC recommended Bryston Amp's and their active crossover (on approval). This was a step back in several areas from the Krell KAS Amp's he had used on the passive system, but some areas improved.

So the Bryston stuff was duly send back and I went to work.

I modified a very expensive BSS digital crossover with just transformers after the DAC Chips and did a lot of work on the powersupplies. After testing ton's of Amp's for mid and treble he heard my "David Jerico" system driven by my special DC coupled 300B Amp and he decided he wanted the Focal Tweeters in his speakers and he wanted to try some SE Amp's for Mid and Treble.

In the final accounting the speakers ended up with the BSS Digital XO (direct digital input from his DAW - this was before the return of the analog summing mixer) with the BSS employed to time-align the whole shooting match and to equalise it flat. Krell KAS something on the bass (dual Volt 15" Transmission line loaded), Cary 845 Amp on the (Vifa?) 3" Dome midrange and Cary 300B on the Focal tweeter.

Crossovers incidentally where whatever this digital XO had as 4th order LR, at 700Hz and 4KHz about IIRC, speakers where carefully soffit-mounted of course, so they actually where genuinely Zero Diffraction. Before EQ'ing the speakers we carefully matched the mid and treble drivers off-axis response (felt wedges). The STudio of course was quite substantially treated for a low and even RT60 and so on. Hence not much what was done thre can be translated back into a domestic context. It is also interesting that using MSL based, gated measurements for EQ-ing the system lead to a rather un-natural sound, in the end we used pink noise 12th octave smothed with a slow average to give us a FR that measured like it sounded...

This system now sounded nothing like the original or the Bryston active version. Walking from the control room over into the recording areas was spooky, as what was audible was essentially the character of the microphones and the positioning.

In terms of Studio systems that I have heard over time it came closest to an "open window" in terms of far field systems. The SET Amp's in the middle and High Channels maybe gave a little extra "bloom" and "space" but allowed very loud monitoring levels with little fatigue. Some carefully done on-location recordings made with the classic Decca Tree showed amazing ability to re-create the acoustic space on this system.

Given the low cost of six or eight channel Pro-Sound external sound cards and the ability of modern PC's to handle all the processing and relatively cheap decent multichannel amplification (Chipamp's) I am surprised we do not see more such systems in DIY.

Ciao T

 

[Joachim Gerhard]

After all this discussion about dispersion and diffraction here is what Peter Walker had to say. He calls the radiation pattern of a loudspeaker "directivity". It is also a parameter defined for speakers that are used in German Radio Stations or for TV broadcast. I think there is also a "Directivity Index". How that index should be is an ages old problem. Peter Walker talked about that somewhere else but i can not find the original material at the moment. As far as i know he optimised the directivity of his speakers by listening tests, based on measurements of cause and came to the conclusion that it had to be optimsed by ear.

 

[Joachim Gerhard]

I found that the German word for "Directivity Index" is "Bündelungsgrad".
Here is the Wikipedia excerpt :
Bündelungsgrad ? Wikipedia
Yes, Thorsten, to be in the studio and have the life music on hand is a very good situation.
I heard from old Siemens engineers that they did the same. Walking between life music and the control room and adjusting the system until it sounded convincig.

 

[ThorstenL]

Hi Joachim,

After all this discussion about dispersion and diffraction here is what Peter Walker had to say. He calls the radiation pattern of a loudspeaker "directivity". It is also a parameter defined for speakers that are used in German Radio Stations or for TV broadcast. I think there is also a "Directivity Index". How that index should be is an ages old problem. Peter Walker talked about that somewhere else but i can not find the original material at the moment. As far as i know he optimised the directivity of his speakers by listening tests, based on measurements of cause and came to the conclusion that it had to be optimsed by ear.

I have not really kept up to date on this topic with the IRT publications, but fairly recent (last decade) research from that corner that came my way made a few critical points about DI (short for Directivity index, something anyone in Pro-Sound tends to be intimately familiar with and which seems an enigma in HiFi circles!?).

If I remember this right, crucial points are:

1) DI should change smoothly, if at all, sharp discontinuities in DI cause audible colourations.

2) DI should be reasonably constant with frequency, falling DI beloiw 500Hz and rising above 5KHz are considered undesirable but unavoidable problems, rather than desirable charateristics.

3) DI for two channel and for Multichannel application needs normally greater DI for multichannel.

4) Practical DI of 6dB in the 500Hz-5KHz or wider range is suggested, cardiodid pattern (which is why you see more and more MEG "K" Series monitors in German Studios).

My experiences so far make me mostly agree with the IRT on these issues, for farfield systems anyway. Nearfild systems are a whole other kettle of fish.

Speakers that sound fine far-filed, sound almost uniformly awful nearfield and near boundaries and visa versa.

Ciao T

 

[Joachim Gerhard]

Good information Thorsten. Thanks, and i hope you have a good xmass.

 

[ThorstenL]

Hi Joachim,

Seems my nearly 1 year old Daughter enjoys X-Mass a lot (especially her new toys and all the attention from everyone of the new people coming over). Even for a hardened X-mess cynic like me it was nice to have a few days with family and friends.

I hope you also had/have a good time, and "guten rutsch und hals und beinbruch!" for the new year.

Ciao T

 

[Joachim Gerhard]

Hi Torsten !
Wait until she is 4. That was the best time i had with my child and now he is 17 !
Time runs fast and i have work to do.