Been sleeping, Kind as you said there are other dynamic un linear and noise contributing elements in a speakerdriver , spider, surround and glue joints also contributes to the sound of a driver, as you clear the big stones in your path those things demand your focus as well. Cone noise is the top, then VC material, then VC inductance, then the other factors, often you end in choices off what to accept and what not to accept, having built basically the same driver for more than 10 years, I know it's not about revolutions, but gradual focused evolutions. I am very privileged as my play with things have no limit, and where we are now it different from where we are a few years from now.
The speaker drivers constitutes the heart of the speaker, and regardless how good boxes or crossovers you make you will never raise performance above the drivers potential. Yet the box is a great determining factor when it comes to interfacing to the amplifier.
Traditional optimization thinking around boxes has hand mainly one objective, how to extract as much lower bandwidth as possible. For this you use spring loads and the resonances that goes with that. This is in my book a no way path- just exporting the dynamic issues to the amplifier.
The speaker drivers constitutes the heart of the speaker, and regardless how good boxes or crossovers you make you will never raise performance above the drivers potential. Yet the box is a great determining factor when it comes to interfacing to the amplifier.
Traditional optimization thinking around boxes has hand mainly one objective, how to extract as much lower bandwidth as possible. For this you use spring loads and the resonances that goes with that. This is in my book a no way path- just exporting the dynamic issues to the amplifier.
Last edited:
MIIB,
I do find working on speakers so interesting in all the small details and both on the motor side and the mechanical side. The electronic side has been so well analyzed and understood for so long by those in the know that I leave that to people like John Curl and others to work on what little is left to solve. What we can do on the mechanical side still has many areas to explore such as your and my development of alternative to paper cones. I still have somewhere an old pair of aluminum Bozak drivers, I would say this was a historical link to what you are doing today. It seems that some things get rediscovered that are lost to history from many corners. Those also used an anodized aluminum cone and that was so long ago. I agree that to many think all that is important is a fancy box stuffed with commodity drivers and a fancy crossover network. You can never get better than the driver is capable, you can't do with electronics what is needed in the driver, if you don't get the driver right you can never get where we all seem to want to go, as close to live as we can get.
One of the biggest hits to the audio industry in my eyes has been the misconception of the mini speaker and subwoofer concept as Marketed by the B*se company. The marketing dollars spent convincing the public that you could get the same sound as what we are attempting to do out of a 3" cone and a sub was bought hook line and sinker as they say. This was the biggest hit to our industry that has happened by any one company. It is a shame how the industry was hurt by this campaign.
I do find working on speakers so interesting in all the small details and both on the motor side and the mechanical side. The electronic side has been so well analyzed and understood for so long by those in the know that I leave that to people like John Curl and others to work on what little is left to solve. What we can do on the mechanical side still has many areas to explore such as your and my development of alternative to paper cones. I still have somewhere an old pair of aluminum Bozak drivers, I would say this was a historical link to what you are doing today. It seems that some things get rediscovered that are lost to history from many corners. Those also used an anodized aluminum cone and that was so long ago. I agree that to many think all that is important is a fancy box stuffed with commodity drivers and a fancy crossover network. You can never get better than the driver is capable, you can't do with electronics what is needed in the driver, if you don't get the driver right you can never get where we all seem to want to go, as close to live as we can get.
One of the biggest hits to the audio industry in my eyes has been the misconception of the mini speaker and subwoofer concept as Marketed by the B*se company. The marketing dollars spent convincing the public that you could get the same sound as what we are attempting to do out of a 3" cone and a sub was bought hook line and sinker as they say. This was the biggest hit to our industry that has happened by any one company. It is a shame how the industry was hurt by this campaign.
32 bits suggested for audio DSP -
Very nice ! In the mid-upper-mid where our hearing is most sensitive you are at about 20dB spl. At least an 80-100dB dynamic listening range or more. That is also the noise level of good studios.
It has been well reported that 16 bits didnt sound good enough. While back in the late 1990's Floyd Toole with usual thorough listening tests, said:
View attachment FToole.pdf
Meanwhile, a 2015 paper: Here is a DSP App engineer (John Tomarakos 2015) who recommends 32 bit for audio processing. he says low cost, higher performance DSP are now appearing on the market to satisfy high dynamic range requirements.
So, I would expect it very possible and do-able especially when pro gear starts applying 32b DSP, for lowly consumer gear to soon follow. And, for mastering and used in HiRes downloads. That way we will get the most important low-middle listening level range to sound a lot better.
BTW - John Tomarakos works for Analog Devices. See 'The Relationship of Data Word Size to Dynamic range and Signal Quality in Digital Audio Processing Applications.'
Guess I wasnt so far off the mark, after all. Its coming.
THx-RNMarsh
Very nice ! In the mid-upper-mid where our hearing is most sensitive you are at about 20dB spl. At least an 80-100dB dynamic listening range or more. That is also the noise level of good studios.
It has been well reported that 16 bits didnt sound good enough. While back in the late 1990's Floyd Toole with usual thorough listening tests, said:
View attachment FToole.pdf
Meanwhile, a 2015 paper: Here is a DSP App engineer (John Tomarakos 2015) who recommends 32 bit for audio processing. he says low cost, higher performance DSP are now appearing on the market to satisfy high dynamic range requirements.
So, I would expect it very possible and do-able especially when pro gear starts applying 32b DSP, for lowly consumer gear to soon follow. And, for mastering and used in HiRes downloads. That way we will get the most important low-middle listening level range to sound a lot better.
BTW - John Tomarakos works for Analog Devices. See 'The Relationship of Data Word Size to Dynamic range and Signal Quality in Digital Audio Processing Applications.'
Guess I wasnt so far off the mark, after all. Its coming.
THx-RNMarsh
Last edited:
I couldn't agree more.
But by the same token, I don't think that same market would want what the high end industry has to sell, i.e. a collection of wood and metal boxes connected with a tangle of wires.
I really think, and have for the past 7 years or so, that it's time for a revival of the old console stereo concept.
Something more the size of a credenza. Something designed like furniture and look like it actually belongs in a living room and not a laboratory. Something that would actually be APPEALING to women. Something that would just use a single socket on a wall outlet and not necessarily need any other wire because it would have wireless capability.
It would basically be the console stereo of the past, but without any source components. Just analog, digital and wireless inputs, amplification and loudspeakers. And since it wouldn't have any source components, it wouldn't require any hinged top to access them and be perfect for putting a large screen TV on top of it.
Someday...
se
Nothing less than 32-bit DSP has been used for years. This is a blind track of reasoning, again and again. You may have quite "perfect" digital recording. It is a consumer market demand that destroys it by dynamic compression etc., to make it playable from pocket digital in underground train.
Go for good classical music, forget pop and consumer music.
If a good digital recording sounds poor, start finding problems in analog path. Tape and vinyl are preferred only for the reason that they mask problems by their inherent low dynamic range.
That someday came and went just a few years ago. I suggested to your friend Noel Lee he start just such a line of audio furniture... to be sold in furniture stores (and they were)... until the economy went over the cliff. Came in a variety of colors and sizes. Speakers of your choice would go at each end (the grill is hinged to open). Or outside speakers and the subs inside (which were also offered). I got a free sample with built-in subs to hold my gear. But there is no significant market for them now.
THx-RNMarsh
Last edited:
That's not quite my idea of a console stereo. It sounds like it was little more than a glorified equipment rack. I mean, if you have to go looking for speakers to use with it, it's not really what I have in mind.
I've done a number of searches and haven't been able to find any such animal. Do you have any photographs or other evidence of its existence? I've seen other "console" attempts by others, but none, save for one has been quite what I have in mind.
The one exception was a demo by Klipsch at the (I think) 2013 CES. It was virtually what I had in mind. No source components, built in amplifier and loudspeakers with analog, digital and wireless inputs. Their projected retail price was even exactly the same as a figure I had in mind ($6,000).
But like virtually all of high end audio, they were marketing it strictly toward the testosterone crowd, i.e. men.
First, it was all black.
Second, the high point of their demo was putting a glass of water on top of it and cranking Metallica's Enter Sandman at ear-splitting levels to show that the glass of water didn't vibrate, like that scene form Jurrassic Park when the T-Rex is approaching.
They said they planned to release it first quarter of 2014.
I haven't seen or heard of it since.
And I wouldn't try selling through furniture stores. Ideally, I'd like to sell through Costco.
Anyway, I'd like to see what this Monster console was. Sounds like it was more of a Man Toy. What, you're going to send women out to go shopping for loudspeakers to use with it?
se
People here speak of high or higher efficiency speakers. Exactly what would you consider to be a higher efficiency speakers? How many dB SPL at 1 m using 2.83V?
If we look back, we had speakers declared as 103 dB/2.83V/1m by Altec, and we had speakers like Dahlquist declared at 82 dB/2.83V/1m, one hell of a range.
That said, I must admit I prefer more efficient to less efficient speakers, in my view more efficient starts at say 92 dB/2.83V/1m and goes up, but that's just me.
If we look back, we had speakers declared as 103 dB/2.83V/1m by Altec, and we had speakers like Dahlquist declared at 82 dB/2.83V/1m, one hell of a range.
That said, I must admit I prefer more efficient to less efficient speakers, in my view more efficient starts at say 92 dB/2.83V/1m and goes up, but that's just me.
You're going to have to show me because I've been to IKEA's and Spectral's websites and don't see anything like I have in mind.
se
I know, I guess Dick is applying selective neglect to prove some point. Almost 20yr. Today's benchmarks, biquad filter 4.4nS per sample, fir 1.1ns per point. Also I don't see a player needing any DSP if you are simply recovering data and sending it to a DAC. miniDSP has a floating point SHARK based FIR room corrector that is pretty cheap ~10,000 points on all channels IIRC.
Last edited:
Err, I don't think that's what Pavel's saying.
For audio quality, go for Classical (and I'd add jazz)
I assure you, Pavel rocks like the rest of us!
And here (see attachment) you will say that it is ugly but it could be worse!
Upper trace, same apparatus, same cal file, 12 Feb 2015 18:48 same room, with very low outside noise (it was snowing, no car traffic outside) but an air-conditioning unit was in operation on another neighbouring space, 6m away from the mic.
That must be the theoretically achievable.
From the few times I have casually monitored the min and max SPL while listening to music, I don’t think that the dynamic range was more than 60-70dB.
You have intrigued my curiosity for doing a more cautious observation on what is the average dynamic range in my living room with the music I listen to (classical and opera, the later only when I am alone at home
) https://web.archive.org/web/20130213113846/http://www.analog.com/en/content/relationship_data_word_size_dynamic_range/fca.html
George
Attachments
Last edited:
Here is introduction to DSP with convolution:
http://www.diyaudio.com/forums/mult...sum-monopole-dipole-speakers.html#post4227362
I would guess you HAVE experienced power compression but just like instantaneous voltage clipping, it is not audible as a distinct flaw until it becomes large enough.
The thing that has changed since the drivers of days gone by, isn’t the conductor but rather the insulation and glue that holds the wire to the VC former. A “Voice of the Theater” never exhibited power compression because the wire would come off the VC former at about 160 degrees C.
Modern high power Voice Coils’s can go up into the 300’s and that is degrees C not F.
Since the resistance is still tied to temperature, what one finds is modern high power drivers HAVE much more power compression at their rated power than drivers from the say 70’s.
The “power compression” shows up at least two ways, one the SPL will be less than the sensitivity and power math would suggest and this can be 6 or 8 or more dB reduction vs ideal. Also though, for a woofer, the doubling or more of the Rdc causes the T&S parameters to change and so the frequency response shape also changes.
For a state of the art Pro-sound woofer may well measure using a Klipple machine to have 6 to 9 dB of “power compression” and this is evident in less than 60 seconds, fully established shortly after, going from cold to full power.
The funny thing is, if one can keep the wire glued on, one eventually reaches a temperature where increasing the input power further actually reduces the sound level making it truly bullet proof.
Beyond the effect of a changing Rdc, there are other things which impact the dynamic behavior of a loudspeaker. While the air compliance follows the gas law, mechanical springs (like the suspension and spider) usually have level dependant properties as well and hysteresis gives things a dependence on the recent history of operation as well.
The only way insure dynamics is to be using a loudspeaker and power amp FAR FAR below it’s maximum capacity. The only way to insure you don’t have Voltage clipping on peaks somewhere is to look at the amp output with an oscilloscope.
Best,
Tom
Here they used two 16's to get 24. But it was DC, and they used a 10 digit hp meter to create the lookup table.
The app was essentially DC, it was back in '93.
It certainly was NOT 32 bit monotonic, and probably not 24 bit mono either.
jn
- Status
- Not open for further replies.