John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
1audio said:


Sometimes you make do with what you have-

Start out with two oscillators, a very stable low distortion low frequency oscillator and a modulatable high frequency oscillator.

Run the lf oscillator a low level through the chain into a sensitive phase meter. Mix into the input the HF oscillator pushing the system to close to its power limits. Cycle the HF oscillator at various rates consistent with the thermal time constants in question. Pass the output through the LPF and see the effect on the phase meter comparing the input phase to the output phase.

I have read but can't cite references that the relative phase of a composite signal are not audible, however if the phase relationship is changed the change is audible. I don't know if there is any data on this.

It would work;

but what I had in mind, to make 2 very clean very stable HF oscillators with LF difference between their frequencies, run them both equally mixed, filter out them on output and see how big is their product of inter-modulation, and how big are phase differences when level of input signals changed slower than a thermal time constant.

One more option is (if an intermodulation is too low) to rectify an output precisely to get it's elvelope then filter, and all the same.

The third option is, to use the 2'nd one, but with a synchronous demodulator. Phase shifts would be easy detectable in such case.
 
I have read but can't cite references that the relative phase of a composite signal are not audible, however if the phase relationship is changed the change is audible. I don't know if there is any data on this.

This might be of interest:

An introduction to the psychology of hearing
By Brian C. J. Moore

http://books.google.com/books?id=tk...0JD0CA&sa=X&oi=book_result&ct=result&resnum=1


I'm loathe to quote very much of this, because it's copyright and I'm not sure this is a legal download in the US or even possible to download there. Its certainly not possible to copy it.

Moore is writing about the psychoacoustics of loudspeakers and says experiments indicate we can discriminate between transient sounds differing only in the relative phases of the components - durations as small as 2 - 3 ms.

Also, (I'm typing from my copy):

Changes in the relative phases of components are much more noticeable when these components are close together in frequency than when they are widely separated (see Chapt 3, Sec 4D), so phase changes which occur abruptly as a function of frequency have a larger subjective effect than phase changes which occur gradually over a wide frequency range. Unfortunately, the former is the common situation..

And:

Blauert and Laws (1978) investigated the audibility of phase distortion by taking a variety of complex sounds, including speech, music, noise and brief impulses, and delaying a group of frquency components with respect to the remainder. They found that the minimum detectable delay was about 400 microseconds when the frequencies delayed were in the range of 1 - 4 kHz.

PP 345 -346

I can't recommend this book too much. Wonderful lots of references.

See also:

http://www.jstor.org/pss/3066509
 
Frank;

subconscious mind can distinguish 10 microseconds, when conscious mind 10 miliseconds only.
How to prove?
2 snaps near the same ear VS 2 snaps near different ears.

That's why I absolutely don't care of conscious judgments of experts, but pay very close attention to emotions caused by sounds in persons who are far from music. Why? Because musicians and others who often deal with music have well trained musical imagination so they easily delete distortions, add missing information, and hear in imagination another instruments than are reproduced.
 
scott wurcer said:


John, I'm really sorry you have been having problems. I have several relatives in the same situation and my father went years with these issues. We always think things will go on and on, no worries.

My father in law wears black glasses after each shot. It is pain to see him...

Get well John, we need you! We need you exactly who you are: rude, nice, don't matter. However, it is better to see you smiling...
 
brianco said:
Wavebourn,

I like what you say! It is much closer to reality than are isolated lab produced experiments. They too often become ends in themselves rather than a means to an end....there is a point where instinct and intuition aka as an act of faith must play a part.

Heh-heh...


:cool:

By the way, what I said that info about 10 mili- and micro- seconds was proved in the lab by (AFIAIR) the same Blauert that Frank mentioned. ;)

...and 30 years after people continue arguing about phase modulations and listening tests. :D

You are absolutely right: results of experiments die in files. :angel:
 
AX tech editor
Joined 2002
Paid Member
1audio said:


THD and IM stuff is interesting but my interest was time based modulations. They may show as sidebands but may be more audible as time shifts. Much harder to measure.


Again, jitter comes to mind. When we assume that, say, 100pS jitter is audible, how would that translate to time/phase modulation in the analog domain?

jd
 
AX tech editor
Joined 2002
Paid Member
PMA said:
I often think about a kind of analogy, because poor analog amplifier sounds similar to poor digital source. IMO, we are seeking for non-harmonic phenomena.


Yes, if you think about dynamic phase modulation, I can't help 'seeing' it as jitter - the zero-crossings moving to and fro, as it were.
If we can make that connection, that would help. We could find an audibility threshold for jitter and from that deduce how much phase modulation would have the same effect in terms of jitter. I think.

jd
 
janneman said:



Yes, if you think about dynamic phase modulation, I can't help 'seeing' it as jitter - the zero-crossings moving to and fro, as it were.
If we can make that connection, that would help. We could find an audibility threshold for jitter and from that deduce how much phase modulation would have the same effect in terms of jitter. I think.


Jan,

You are not alone in that thought OK, you have to read between the lines of the marketing puff, but DPA c. 1992 were probably looking at the same thing sideways.
 
scott wurcer said:
As a practical matter the slope at some of the zero crossings will be so low that the time resolution will be very poor.

Sure. That's why I prefer not to create such problems, hence not to use symmetrical complementary topologies. It is much easier than to create them then hunt for resulting errors that are hard to measure, but easily audible on subconscious level. What they kill, they kill reverberation that is vital for nice music listening. However, when the record was made using multiple microphones in anechoic chamber, with hundreds of opamps, then artificial reverberation applied, then fine details lost in digitizing, what hiend are we talking about at all?
 
Status
Not open for further replies.