Scott,
Yes, some use multiple capsules, others use dual diaphragm capsules. I'm used to the Microtech Gefell versions of the classic variable pattern Neumann Mikes.
There is a lot that can be done with with multiple capsules, but I have specific requirements. Even what I am after can still be done variable pattern...
Ciao T
Yes, some use multiple capsules, others use dual diaphragm capsules. I'm used to the Microtech Gefell versions of the classic variable pattern Neumann Mikes.
There is a lot that can be done with with multiple capsules, but I have specific requirements. Even what I am after can still be done variable pattern...
Ciao T
But I never did. I speak about A/D all the time. Without accurate A/D, we have no exact data for music replay.
Again the same question - is sigma-delta A/D able to yield accurate sample values, in correct time, for transient input signal (bandlimited <Fs/2)? If the answer is yes, I would like to see the see similar plot as I have shown for the pipeline A/D converter.
Array of 1/4" capsules could work but the combing effects of 16 of them is problematic. How about 16 A/D's and massive DSP to make them a virtual point? 😀
Last edited:
Position them in concentric circles and it just takes simple time delays to make them into a virtual hemisphere. For a large enough array, it could be done crosspoint on a rectangular grid.
edit: Or be coaxial 2 way.
Hi,
Hmm, if we have for example a 500Hz crossover and the minimal possible spacing for a 1/4" and a 1" capsule, just how much real phaseshift would we be able to get at 90 degrees vertical off axis?
Well, I'm really upset by people telling me I know nothing about digital audio and then showing they are completely ignorant of really basic electronics. Tough luck.
Ciao T
Hmm, if we have for example a 500Hz crossover and the minimal possible spacing for a 1/4" and a 1" capsule, just how much real phaseshift would we be able to get at 90 degrees vertical off axis?
Well, I'm really upset by people telling me I know nothing about digital audio and then showing they are completely ignorant of really basic electronics. Tough luck.
Ciao T
Uncorrelated noise sums to 3dB. Signal and correlated noise sum to 6dB. That's a 3dB gain (re uncorrelated noise) if the outputs of the two ADCs are summed. I don't see the issue.
edit: Same argument for "differenced" as well.
edit: Same argument for "differenced" as well.
or maybe adding 2 signals of full scale versus 2 signals of half scale ?
BTW I read the the PCM 1704 DS to indicate the 2 23 bit converters have a common ladder.
Thanks
-Antonio
BTW I read the the PCM 1704 DS to indicate the 2 23 bit converters have a common ladder.
Thanks
-Antonio
I see confusion between the issues around D/As and A/Ds. Also between correlated noise, uncorrelated noise and correlated error signals treated as noise.
As to a dual capsule microphone, the space limitations will depend on the intended working distance. A large capsule is often used for close mic'ing to reduce breathing noise. At a distance the capsule spacing becomes less critical.
Ureda did a paper on the difference between near field and far field. You are in the far field when the phase cancellation due to additional movement goes away. When inline with both elements there is no far field. Of course if the spacing is tight enough the issue could be above the frequency range of interest.
As to a dual capsule microphone, the space limitations will depend on the intended working distance. A large capsule is often used for close mic'ing to reduce breathing noise. At a distance the capsule spacing becomes less critical.
Ureda did a paper on the difference between near field and far field. You are in the far field when the phase cancellation due to additional movement goes away. When inline with both elements there is no far field. Of course if the spacing is tight enough the issue could be above the frequency range of interest.
A lot of posts in the last few days claim that the sigma delta DACs output "fuzzy distortion" while the ladder DACs don't, and therefore the ladder DACs are better (at least that's condensed version).
The term "fuzzy distortion" is too vague, too imprecise, too "fuzzy" 😀.
Let's try to clarify what "fuzzy distortion" means.
The main spec parameters of a DAC are ENOB, sample frequency, etc..
Let's consider two DACs, a ladder one and a sigma delta one with the same sample frequency, ENOB, linearity, etc..
If we feed the two DACs with the same stream of numbers they should output the same voltage up to a maximum error of 1 ENOB LSB.
If this is the case, then we shouldn't care whether one is a ladder DAC and one is sigma delta or whatever architecture.
Is anyone here claiming that the sigma delta DACs don't satisfy this basic specification ?
In more detail, that the voltage error at their output is larger than 1 ENOB LSB is specific cases like transients, etc. ?
If that's the case, then what kind of measurements exposes it ?
The term "fuzzy distortion" is too vague, too imprecise, too "fuzzy" 😀.
Let's try to clarify what "fuzzy distortion" means.
The main spec parameters of a DAC are ENOB, sample frequency, etc..
Let's consider two DACs, a ladder one and a sigma delta one with the same sample frequency, ENOB, linearity, etc..
If we feed the two DACs with the same stream of numbers they should output the same voltage up to a maximum error of 1 ENOB LSB.
If this is the case, then we shouldn't care whether one is a ladder DAC and one is sigma delta or whatever architecture.
Is anyone here claiming that the sigma delta DACs don't satisfy this basic specification ?
In more detail, that the voltage error at their output is larger than 1 ENOB LSB is specific cases like transients, etc. ?
If that's the case, then what kind of measurements exposes it ?
Last edited:
you have to specify that this is with analog anti-alias in front of the ADC, anti-image/reconstruction/noise rejecting filter after the DAC - and then compensate for phase shifts
some here want to claim the shaped noise of Delta-Sigma is an inherent flaw in its reproduction of audio frequency signals despite the requirement that it be filtered out
properly band limited reproduction is all we can ask of ADC/DAC
some here want to claim the shaped noise of Delta-Sigma is an inherent flaw in its reproduction of audio frequency signals despite the requirement that it be filtered out
properly band limited reproduction is all we can ask of ADC/DAC
I assume you are making a product with sufficient margin that this will be worked out in the initial calibration.
Well at the recent NY AES there were more than 25 microphone manufacturers showing product. If some of them sell 100 units a year I would be surprised. So there is a market for high priced low production products... maybe.
Well, some people will pay 'any price' for the best sounding stuff. That I rely on. We, 'normals' will never have enough extra money and hearing perception in the same individual to do the same. Some of us have 'good' ears, but limited resources. Others of us have 'tin' ears and lots of resources. Only a few have both 'good' ears and lots of resources, and these are our best customers. They do exist, and keep me designing new products at the very highest level.
Hi,
One may say this is a summary. It is however inaccurate.
I have OBSERVED (there is no claim involved) that DS DAC's operate on a principle that relies on adding very large amounts of a synthetic noise, that unavoidably involve patterns and that as a result the accuracy of a given individual sample is severely compromised.
As result instead of precise value we get a wide error band around the original value which analogue lowpass filtering cannot remove and which makes the output distorted in a way that is best described as Fuzzy, in the was as when using fuzzy logic.
Equally, unless heavy averaging is applied the "low noise" of such a DS converter does not exist. Analogue 'scopes simply display this error band as a massive widening of the signal lines with static signals, digital 'scopes give interesting results and single shots show clearly the expected fuzziness clearly.
And no, the noise common in analogue systems and the dither/fuzzy distortion are not the same or comparable any more than a milkshake made real strawberries and one made with artificial strawberry aroma taste even alike, never mind identical.
And how are these specifications derived? In ways possibly, that average out the "error band"? So we measure one thing but what matters is another.
Yup.
IF the two DAC's behave identically, this will be the case. If one is low bit DS and the other multibit, it will not be the case, UNLESS massive averaging is employed and with respect, the human ear does not average sounds over several seconds, unlike an AP2.
So, it takes no PHD in digital converter design to know what the result will be.
If the output was the same, yes.
My 'scope claims nothing, but it shows distinct differences.
So, at least in the context of SOME measurements, which do not use averaging, yes, there are observable differences that suggest that indeed DS DAC's do NOT behave like you set out.
Any that look at essentially single samples, not several seconds worth of samples of a static signal.
Ciao T
One may say this is a summary. It is however inaccurate.
I have OBSERVED (there is no claim involved) that DS DAC's operate on a principle that relies on adding very large amounts of a synthetic noise, that unavoidably involve patterns and that as a result the accuracy of a given individual sample is severely compromised.
As result instead of precise value we get a wide error band around the original value which analogue lowpass filtering cannot remove and which makes the output distorted in a way that is best described as Fuzzy, in the was as when using fuzzy logic.
Equally, unless heavy averaging is applied the "low noise" of such a DS converter does not exist. Analogue 'scopes simply display this error band as a massive widening of the signal lines with static signals, digital 'scopes give interesting results and single shots show clearly the expected fuzziness clearly.
And no, the noise common in analogue systems and the dither/fuzzy distortion are not the same or comparable any more than a milkshake made real strawberries and one made with artificial strawberry aroma taste even alike, never mind identical.
And how are these specifications derived? In ways possibly, that average out the "error band"? So we measure one thing but what matters is another.
Yup.
IF the two DAC's behave identically, this will be the case. If one is low bit DS and the other multibit, it will not be the case, UNLESS massive averaging is employed and with respect, the human ear does not average sounds over several seconds, unlike an AP2.
So, it takes no PHD in digital converter design to know what the result will be.
If the output was the same, yes.
My 'scope claims nothing, but it shows distinct differences.
So, at least in the context of SOME measurements, which do not use averaging, yes, there are observable differences that suggest that indeed DS DAC's do NOT behave like you set out.
Any that look at essentially single samples, not several seconds worth of samples of a static signal.
Ciao T
Hi,
Yes, I had the chance to listen to and play with it a few times.
I do not think I should comment on it's sound etc., as I am involved with making what many see a competing product.
Ciao T
Yes, I had the chance to listen to and play with it a few times.
I do not think I should comment on it's sound etc., as I am involved with making what many see a competing product.
Ciao T
- Status
- Not open for further replies.