Thank you! I thought it was an analog one...i don't understand this language unfortunately
At least PMA can hear the different of cymbal recording at 44.1 and 96 and you don't. You will call PMA an alien?
Maybe PMA just a normal person, and you are....
THx-RNMarsh
Suggest this work be turned into a paper for publication (AES? IEEE? EDN?)
Ditto JN's study.
Finally, getting some info that could explain hearing differences besides just looking at thd/IM and FR etc.
Last edited:
Of course, I still interesting.
PMA proved that he can hear the different.
It proved technically that envelope is different.
You can ignore who can not hear the different and concentrate in technique/science.
For noise voltage 10Hz-100kHz :
National, TI and NJR 7805 datasheet quoted 40uV
LT1693 LDO datasheet quoted 45uV (the one Markw4 compared with on the AK4499 eval board)
What do I miss?
I didn't even bother...some people hear strange noise just by looking on their analyzer screen.
Actually it's 45uv for the NJR part, but PSRR add up too.Just use bigger capacitors in and out...I've seen 10uv noise for 10v output /10w/10 days up to 1mhz smps power sources, but they weren't for sale...
Last edited:
I'm having trouble imagining what the envelope of a chaotic waveform means from an audible perspective. Calculating the envelope is a (highly) non-linear operation that introduces new frequencies. Differencing the envelopes in this case produces a spectrum from DC to Nyquist. What is the significance of that?
You lose nothing by ignoring his witch hunts.
It is enough that someone he hates has said something for him to take the opposite side. A unique party, of course.
Here, he strangely puts on the boots of a behavior usually attributed to audiophiles: the more expensive the better.
We now still have to analyze the influence of the feedback loops from regulators on the way the sound of the circuits they supply is affected.
Scott,
Interesting image.
It is obvious that the envelope changes when reducing the frequency content.
It is also to be expected that Gibbs ringing will be added to the filtered signal.
Could you additionaly make a plot of the signal’s magnitude, its HP filtered version from 0 to 220500 and their subtracted difference ?
Hans
Please, consider 3 things:
1- What you qualify of "chaotic" is called transients.
2- Our hearing system has adapted, during a long evolution, to offer us the maximum chance of survival: It is perhaps more sensitive to these transients than what you imagine: the sinusoidal signals that we use to determining our limits of perception of high frequencies do not exists in the nature.
3- Between the electric signal and our ears is an electro acoustic transformer called a "Speaker". How the slew-rate of a transient signal affect its responses ?
The problem with your question is included in the word "imagine". You just have to make experiments, listening by yourself, to figure out if it makes, or not, a difference for you.
To answer point 3, I suggest a fairly simple test: Compare the acoustic differences (with a microphone rising fairly high in frequency) between a sinusoidal signal and a square signal at 12kHz, applied to your speakers on an oscilloscope.
Scott,
Just one more question.
I heard you mentioning the use of Hilbert afair.
Did you really use this ?
After filtering the complex Fourier transform of the time signal, when going back to the time domain, all that has to be done is to calculate the inverse Fourier without Hilbert.
Or did you use the magnitude of the Fourier transform in combination with Hilbert.
In that case you will create a slightly different time domain signal IMO.
Hans
Just one more question.
I heard you mentioning the use of Hilbert afair.
Did you really use this ?
After filtering the complex Fourier transform of the time signal, when going back to the time domain, all that has to be done is to calculate the inverse Fourier without Hilbert.
Or did you use the magnitude of the Fourier transform in combination with Hilbert.
In that case you will create a slightly different time domain signal IMO.
Hans
THx-RNMarsh
Suggest this work be turned into a paper for publication (AES? IEEE? EDN?)
Ditto JN's study.
Finally, getting some info that could explain hearing differences besides just looking at thd/IM and FR etc.
Richard, this is observable with an incredibly long FFT. Will this happen with ordinary D-A’s?
Just posing the question.
Back to 78xx style regulators. I actually quite like the 7815 - it’s about 1/3 the noise of an equivalent 317 reg circuit and great for general purpose work. I would not of course use it to power an MC head amp. I always use split secondaries and rectify, smooth, regulate and then combine to form a split +- supply so I therefore only use one type of regulator IC.
Opamp PSRR is so good now that 100uV of PSU noise ends up buried way, way below the noise floor of any line stage or MM EQ amp. Decent decoupling and layout all have to be beyond reproach of course.
Opamp PSRR is so good now that 100uV of PSU noise ends up buried way, way below the noise floor of any line stage or MM EQ amp. Decent decoupling and layout all have to be beyond reproach of course.
Last edited:
Gibbs phenomena
Gibbs phenomenon - Wikipedia
Am I right in saying this phenomena is only observable in discontinuous functions - ie like a square wave? It is not an issue in continuous functions where the Fourier series converges more quickly?
Gibbs phenomenon - Wikipedia
Am I right in saying this phenomena is only observable in discontinuous functions - ie like a square wave? It is not an issue in continuous functions where the Fourier series converges more quickly?
Last edited:
THx-RNMarsh
Suggest this work be turned into a paper for publication (AES? IEEE? EDN?)
Ditto JN's study.
Finally, getting some info that could explain hearing differences besides just looking at thd/IM and FR etc.
Without knowing what you see and why, you already tend to jump to conclusions.
Hans
- Status
- Not open for further replies.