Actually William of Occam had to flee to France because of his "Razor." This is not "A needless plurality." But I know you meant the modern widely used mistranslation of "The simplest possible explanation is the best."
You don't see the distortions because if they exist they are below the limits of measurement. You can't assume that they don't exist as that is attempting to prove a negative result without having tested all possible conditions.
(Yeah long ago I did a paper on old Billy...)
Last edited:
Your opinion of course
. JC is claiming levels far above what you show here also the more sensitive the measurement the more subtle the unknowns become i.e. even the AP2 will fail at some point and you end up with just the same situation at 20 or 30dB lower levels.You would do well to repeat the above graph with a couple of instrument grade oscillators.
Last edited:
Your opinion of course
You would do well to repeat the above graph with a couple of instrument grade oscillators.
Scott,
My original goal was to see if lead free solder had more distortion than the 63/37 that I was using. I spent quite some time on it. Prepared samples of 200+ joints of four solder types and was able to determine that there was no significant difference. In developing my measurement instrument I did try conventional oscillators using the dbx vca as my feedback control element. I could get very nice low distortion but keeping everything working was a real pain. I could get the same results on the samples I tested but it was not as reliable.
I then tried re-soldering all the connections on an audio PC card and found that did result in significant and measurable differences. I attributed that to filling in the vias.
It was an interesting side effect that different audio cables showed different results. It was completely unexpected that reversing cables showed consistent but sometimes different results. I presented my method for making the measurements along with some brief examples of results so that anyone interested could pursue it.
I have also presented the spectrum of the transversal filter oscillators that I used. The results I got on solder and cables do differ and do not show what I would expect from oscillator issues. That is OPINION and yours may differ, that to me is not an issue worth pursuing at the moment.
However my quest was for solder information. I have that. I have also measured resistors and capacitors. I am currently working on contacts and how feedback influences distortion.
The second plot shown is roughly 60 db below the first data plot. (0db is approx .38 VRMS) My design goal for audio is 160 db S/N. So audio interconnects properly used will meet that design goal. I can pick resistors that meet the goal. I can pretty much avoid capacitors, so they are not as important. In my OPINION that is audio engineering, not science research. Confirming the results and determining the mechanism is to me worthy of a PHD, anyone who would like to pursue it is free to do so.
I have not examined J.C.s setup or determined his actual operating or measured levels. He is looking at THD I look at IM. He is entitled to his OPINION and to do things his way for his work.
BUT THERE AIN'T NO EVIDENCE OF STINK'N MICRODIODES! (Just had to mention that as it keeps resurfacing.)
Last edited:
Right on that, who ever said there were? I'm of the same mind on the dead zones, ghosts.
Hi Sy,
Ahhh, this is interesting. So if I add more noise to my Phonostage it will increase resolution?
Patently NOT. Please stop making categorical statements that lack qualification and are patently false.
Adding dither adds noise. Full Stop.
It degrades Signal/Noise Ratio and results in a reduction of ENOB (effective number of bits), so clearly it CANNOT improve resolution by itself any more than adding noise to an analogue system will improve the "resolution" (we used call this the SNR, FWIW, in the olden analogue days).
Now if use dither and massive averaging we MAY produce the apparency of an increase resolution that is in fact a measurement artefact. FOr each individual sample accuracy is reduced.
Now some theorise that the human hearing includes such averaging mechanisms. I do know some well known Mastering engineers that used to be proponents of using dither when reducing the bitrate, who have reversed their position and not state that simple truncation is preferable, which would argue against such. I have no opinion for this matter, there is insufficient evidence to postulate such seriously. No credible large scale listening tests have been performed.
So I think stating that dither improves resolution, even in a purely subjective sense (we have already dealt with the objective side) is not something I would expect.
It is complete and utter bunkum!
Ciao T
Ahhh, this is interesting. So if I add more noise to my Phonostage it will increase resolution?
Patently NOT. Please stop making categorical statements that lack qualification and are patently false.
Adding dither adds noise. Full Stop.
It degrades Signal/Noise Ratio and results in a reduction of ENOB (effective number of bits), so clearly it CANNOT improve resolution by itself any more than adding noise to an analogue system will improve the "resolution" (we used call this the SNR, FWIW, in the olden analogue days).
Now if use dither and massive averaging we MAY produce the apparency of an increase resolution that is in fact a measurement artefact. FOr each individual sample accuracy is reduced.
Now some theorise that the human hearing includes such averaging mechanisms. I do know some well known Mastering engineers that used to be proponents of using dither when reducing the bitrate, who have reversed their position and not state that simple truncation is preferable, which would argue against such. I have no opinion for this matter, there is insufficient evidence to postulate such seriously. No credible large scale listening tests have been performed.
So I think stating that dither improves resolution, even in a purely subjective sense (we have already dealt with the objective side) is not something I would expect.
It is complete and utter bunkum!
Ciao T
Hi,
Most of this requires no more that EE101 or even the electrics in high school physics to understand.
But because it is not explicitly taught in most EE courses it seems it is by far too frequently ignored.
I am just wading through a week worth of dither in this thread (which I can assure everyone has neither improved resolution, nor increased the amount of usable information in the thread and certainly not the SNR) after soaking up see air, sunshine and way too Margarita's and Mojito's and what I am most shocked by is just how many really basic EE knowledge is shown to be absent.
Several decades I worked in industrial process control electronics, where the "stake" if something was wrong was 140 tons of liquid steel (which in the worst screw-up would destroy thoroughly the machinery and the lesser ones just meant the loss of a LOT of money).
Things like the errors introduced from metal/metal contacts and all sorts of noise sources where understood, accounted for and hence caused no issues (unless some ejit in maintenance ignored the instructions on how to make contacts etc. and hence nothing worked as expected - that ejit usually never got the the chance to do it again either).
Meanwhile in audio the same issues are simply attempted to be "talked away" by some while others, who try to draw attention to such issues are laughed at and all this is called "bunkum" while obvious bunkum is talked up to the level of factual science...
If I was not such a hardened cynic I would probably despair, but as it is, while I delight in the occasional flashed of sense and attempts to general betterment, I am never surprised by the opposite, so it's basically just back to "business as usual".
Well, the sun and sea was nice while it lasted... Maybe I should give up and go back to the seaside and just be known as "Son Of the Beach"...
Ciao T
Most of this requires no more that EE101 or even the electrics in high school physics to understand.
But because it is not explicitly taught in most EE courses it seems it is by far too frequently ignored.
I am just wading through a week worth of dither in this thread (which I can assure everyone has neither improved resolution, nor increased the amount of usable information in the thread and certainly not the SNR) after soaking up see air, sunshine and way too Margarita's and Mojito's and what I am most shocked by is just how many really basic EE knowledge is shown to be absent.
Several decades I worked in industrial process control electronics, where the "stake" if something was wrong was 140 tons of liquid steel (which in the worst screw-up would destroy thoroughly the machinery and the lesser ones just meant the loss of a LOT of money).
Things like the errors introduced from metal/metal contacts and all sorts of noise sources where understood, accounted for and hence caused no issues (unless some ejit in maintenance ignored the instructions on how to make contacts etc. and hence nothing worked as expected - that ejit usually never got the the chance to do it again either).
Meanwhile in audio the same issues are simply attempted to be "talked away" by some while others, who try to draw attention to such issues are laughed at and all this is called "bunkum" while obvious bunkum is talked up to the level of factual science...
If I was not such a hardened cynic I would probably despair, but as it is, while I delight in the occasional flashed of sense and attempts to general betterment, I am never surprised by the opposite, so it's basically just back to "business as usual".
Well, the sun and sea was nice while it lasted... Maybe I should give up and go back to the seaside and just be known as "Son Of the Beach"...
Ciao T
Hi,
Not only here.
I remember that I when I once pointed out that the theoretical Dynamic Range derived by the milkmaid calculation of the number of bits is not accurate as it is not possible to encode such a signal in a digital system (e.g. -96dBFS peak level for 16 Bit's) dither came up again and again, however showing such a signal showed a noisefloor that was increased sufficiently to degrade the real SNR and only massive averaging could reveal the low level signal...
Yes, dither can reduce the appearance of distortion of quantisation distortion (again, this is apparency, not an actuality), but at a price.
One MAY consider in some specific conditions used dither, to overcome problems with measurability or sensor thresholds in specific applications, however to simply derive from such limited and specialised applications that dither has universal and general benefits and improves resolution everywhere is absurd...
Ciao T
Not only here.
I remember that I when I once pointed out that the theoretical Dynamic Range derived by the milkmaid calculation of the number of bits is not accurate as it is not possible to encode such a signal in a digital system (e.g. -96dBFS peak level for 16 Bit's) dither came up again and again, however showing such a signal showed a noisefloor that was increased sufficiently to degrade the real SNR and only massive averaging could reveal the low level signal...
Yes, dither can reduce the appearance of distortion of quantisation distortion (again, this is apparency, not an actuality), but at a price.
One MAY consider in some specific conditions used dither, to overcome problems with measurability or sensor thresholds in specific applications, however to simply derive from such limited and specialised applications that dither has universal and general benefits and improves resolution everywhere is absurd...
Ciao T
Hi,
There are of course many ways in which one may deal with ground/earth loops, secondary path issues and other "equipment ghosts" (which have not a small impact in all this cable differences and related controversies IMNSHO) without failing the electrical safety code, it just costs a little more and needs recognition and understanding of the issues, "talking it away" is cheaper and easier.
Ciao T
There are of course many ways in which one may deal with ground/earth loops, secondary path issues and other "equipment ghosts" (which have not a small impact in all this cable differences and related controversies IMNSHO) without failing the electrical safety code, it just costs a little more and needs recognition and understanding of the issues, "talking it away" is cheaper and easier.
Ciao T
Hi Simon,
I would suggest that a high impedance load is the opposite of what you would want to show anything like "micro diodes", or similar effects that are essentially rectification related.
One needs to construct a test that is appropriate to show (maximise the visibility?) of the effect we theorise about.
I would suggest that high source impedance and appreciable loading would more likely demonstrate problems in at least this domain.
You even hinted at the core of this, when you quipped that you never saw anyone ping a condensor mike to overcome bad contacts at the capsule... There are reasons for that.
Ciao T
I would suggest that a high impedance load is the opposite of what you would want to show anything like "micro diodes", or similar effects that are essentially rectification related.
One needs to construct a test that is appropriate to show (maximise the visibility?) of the effect we theorise about.
I would suggest that high source impedance and appreciable loading would more likely demonstrate problems in at least this domain.
You even hinted at the core of this, when you quipped that you never saw anyone ping a condensor mike to overcome bad contacts at the capsule... There are reasons for that.
Ciao T
Hi ThorstenL Re; your post #14,229.
The first line where you express doubt (or stronger sentiments) about noise increasing resolution.
I can't put my hand on it now but in what used to be "Wireless World" (can't think of current name.....Electronics World..is it?) there was a short piece 3-4 years ago that claimed precisely that. Now it wasn't at audio frequencies but the reason it caught my attention was that a number of us have found that vocals heard while driving a car often appear easier to understand than in the lounge room and I was just wondering if the extra low frequency noise in the car had something to do with the increased clarity.
Jonathan
The first line where you express doubt (or stronger sentiments) about noise increasing resolution.
I can't put my hand on it now but in what used to be "Wireless World" (can't think of current name.....Electronics World..is it?) there was a short piece 3-4 years ago that claimed precisely that. Now it wasn't at audio frequencies but the reason it caught my attention was that a number of us have found that vocals heard while driving a car often appear easier to understand than in the lounge room and I was just wondering if the extra low frequency noise in the car had something to do with the increased clarity.
Jonathan
Do you have a digital phonostage which plays digitally encoded records? Or does your phonostage have terrible quality components in it with multiple dead bands? If so, then possibly yes. If not, then perhaps reading things in context might be useful.
Hi,
I find this interesting. It would be good if you could find the reference.
My native language is German and especially in my earlier years in the UK I had many problems understanding the lyrics of English songs. Noise never helped. Indeed, I think much of my work in HiFi at the time was spurred by this. It took much work to get to a point where I could listen to CD's/LP's and just understand.
My improving english no doubt helped, but I could easily follow a conversation in a noisy pub between a bunch of people speaking english with irish, jah-meican and sundary other accents but even there less noise helped intelligibility while I still had problems following english lyrics past the basic pop song level when hearing them on crappy sound systems, but non on ones I consider "good". In fact, even now I find english lyrics on "consumer grade" gear at times difficult (say stuff like Marillion, some songs by Kate Bush or Enya etc.).
For want of better word, hifi gear that I dislike has a "noisy" sound, it seems to have a subtle signal dependent noise behind the music (I know some call this "detail", to my ears it is not, I never hear this from a live acoustic instrument), I would not call it distortion as such, that makes it more difficult and more of an effort to follow phrasing, lyrics/vocals and the like...
The real fun part is, music with german lyrics, even stuff hard to understand for many germans I can follow easily on almost any gear, I still would not claim that noise helps... Still, such music I also enjoy more with the kind of HiFi systems I favour.
Ciao T
I find this interesting. It would be good if you could find the reference.
My native language is German and especially in my earlier years in the UK I had many problems understanding the lyrics of English songs. Noise never helped. Indeed, I think much of my work in HiFi at the time was spurred by this. It took much work to get to a point where I could listen to CD's/LP's and just understand.
My improving english no doubt helped, but I could easily follow a conversation in a noisy pub between a bunch of people speaking english with irish, jah-meican and sundary other accents but even there less noise helped intelligibility while I still had problems following english lyrics past the basic pop song level when hearing them on crappy sound systems, but non on ones I consider "good". In fact, even now I find english lyrics on "consumer grade" gear at times difficult (say stuff like Marillion, some songs by Kate Bush or Enya etc.).
For want of better word, hifi gear that I dislike has a "noisy" sound, it seems to have a subtle signal dependent noise behind the music (I know some call this "detail", to my ears it is not, I never hear this from a live acoustic instrument), I would not call it distortion as such, that makes it more difficult and more of an effort to follow phrasing, lyrics/vocals and the like...
The real fun part is, music with german lyrics, even stuff hard to understand for many germans I can follow easily on almost any gear, I still would not claim that noise helps... Still, such music I also enjoy more with the kind of HiFi systems I favour.
Ciao T
Sy,
So, you claim ADDING NOISE to a digital system reduces noise and improves SNR/DNR? And doing the same to an analogue system does the opposite?
Or do you claim that the LOSS of SNR/DNR and thus resolution and ENOB when applying dither has magic properties that places it outside the realm of normal physics?
Or do you claim that there is some hidden digital distortion (not covered by traditional measurements) that dither mitigates, so the loss of SNR is preferable to leaving the hidden distortion?
Or what precisely is your claim? You are not very precise.
Also, do you have some serious references or measurements that show how dither improves RESOLUTION (which I would suggest is best evidenced as Effective Number Of Bits [ENOB])? Or maybe a DBT that demonstrates that dither improves "resolution"?
Ciao T
So, you claim ADDING NOISE to a digital system reduces noise and improves SNR/DNR? And doing the same to an analogue system does the opposite?
Or do you claim that the LOSS of SNR/DNR and thus resolution and ENOB when applying dither has magic properties that places it outside the realm of normal physics?
Or do you claim that there is some hidden digital distortion (not covered by traditional measurements) that dither mitigates, so the loss of SNR is preferable to leaving the hidden distortion?
Or what precisely is your claim? You are not very precise.
Also, do you have some serious references or measurements that show how dither improves RESOLUTION (which I would suggest is best evidenced as Effective Number Of Bits [ENOB])? Or maybe a DBT that demonstrates that dither improves "resolution"?
Ciao T
When you find me saying that dither increases SNR as you falsely attribute to me, I'll apologize for calling you a liar yet again.
If you think quantization distortion is "hidden distortion" then you truly have no understanding whatever of digital systems. I don't believe that's true, so I'll assume you just want to argue for argument's sake, which does not interest me.
If you think quantization distortion is "hidden distortion" then you truly have no understanding whatever of digital systems. I don't believe that's true, so I'll assume you just want to argue for argument's sake, which does not interest me.
- Status
- Not open for further replies.