Why not add the -60 or -80dB 9th harmonic to music and see if anyone can hear that?
If a secondary signal is 65 db down from the peak level of music and the music has dynamic range of 60 db, I expect the secondary to show up. I am not clear as to how the demo was done. If the secondary is always 65 db below the music and is not seriously distorted I would expect that to be masked. I suspect for most folks the masking margin can actually be as low as 15 db!
Now if we have a mechanism that causes a ninth harmonic to appear late when there is less masking from the fundamental that should be more objectionable than one tightly linked.
In live music a forte in performance is 10 db louder than a forte in practice. But all of the musicians will tell you a forte is always a forte. Now how you play a violin 10 db louder is a mystery to me. That just shows perceptions shift due to environment, nothing new.
So when I hear something wrong, I take measurements to see what it might be. That way I can adjust the parameter under consideration to "fix" the problem. It is a related, but independent experiment to create test tones with fixed harmonic distortion, phase shift, etc. to see what levels I or others around here can detect.
Yes I have tried switching cables from the worst I have measured to the best. Some folks can tell them apart, most can't. Fully scientific... no, because I can't find a way to change cables with out a person swapping them on the spot without having far greater distortion in the switch.
Almost everyone here can tell that my rotary switch has a different sounding position #1. It is a Lorlin silver plated plastic cased switch.
On my list of testing I would like to try is wideband noise levels. Exactly how to measure it is an open question.
I'm not sure what you're trying to say about 24-bit recording equipment. There are A/D options available that are truly 24-bit in terms of digital bit depth and quantization noise. The catch is that the analog stage in front of any A/D is generally going to be 111 dB at best. Sure, there are a rare few better analog stages, but my point is that it is the analog stage that is not really there yet, not the digital converter. Granted, there are 24-bit A/D chips with much worse than 24-bit quantization noise - those cases are truly a failure to live up to the "24-bit" claim, but not all 24-bit converters fall into this.
I don't. There is no evidence in any context for this exaggerated claim. At some point you have to stop sending random resistors to folks and asking for their uncontrolled anecdotal comments and presenting it as evidence.
> At some point you have to stop sending random resistors to folks
> and asking for their uncontrolled anecdotal comments and presenting
> it as evidence.
And how do you define that 'point' ..........
I'll take some of your samples .....
__________________
You make up some test where some resistors measure -140dB vs -150dB distortion and then you send those resitors to a third party and ask them to use them in some circuit and they line them up in your empirical model of what matters. And then you claim this means that stuff at the -150dB level matters.
Over the decades, we hi end designers have discussed differences in resistors. Audio designers used to use carbon resistors, sometimes 'lousy' carbon resistors as standard. Allen Bradley in the USA seemed to make a pretty good carbon resistor, and Ampex used them for everything except precision attenuators. Back in those days, a 1% precision resistor might cost the equivalent of a cafeteria lunch.
Still, exceptions were found in resistors that added excess noise or distortion. As our measurements got better, we noticed this more. Richard Heyser pointed this out to me, over 40 years ago. In the 1980's 'HFN' came up with an article that measured distortion in both caps and resistors, AND they even tried to evaluate them by listening. They did a very good article.
In any case, all my best designs include specifying certain resistor brands and size. It is part of the design process for me, and I will stick with it. By some coincidence, one of my favorite resistor brands, the German made 'Resista' the standard component in my Vendetta Research phono stage and the Blowtorch, MEASURES really well, compared to some other Resistors, and they only cost me $.05 in limited quantity of 100 or so. Today I rely on Ed Simon to guide me on what is still available and how well it works.
Still, exceptions were found in resistors that added excess noise or distortion. As our measurements got better, we noticed this more. Richard Heyser pointed this out to me, over 40 years ago. In the 1980's 'HFN' came up with an article that measured distortion in both caps and resistors, AND they even tried to evaluate them by listening. They did a very good article.
In any case, all my best designs include specifying certain resistor brands and size. It is part of the design process for me, and I will stick with it. By some coincidence, one of my favorite resistor brands, the German made 'Resista' the standard component in my Vendetta Research phono stage and the Blowtorch, MEASURES really well, compared to some other Resistors, and they only cost me $.05 in limited quantity of 100 or so. Today I rely on Ed Simon to guide me on what is still available and how well it works.
The best high-end mastering converters have 127dB SNR that is equivalent to 21 real bits of audio. The remaining 3 bits are noise. This is the maximum that is possible to obtain considering the difference between the Johnson noise and the maximum levels used in balanced proaudio (0Vu = +4dBU vith 24-26db headroom) which is way more than the -10dBV used in unbalanced hifi. So there is no real chance to improve this 127dB, except if the whole level standard would be changed to a much higher one (no chance) or to imerse the whole converter into liquid nitrogen to reduce Johnson noise.
Anyway even if it would be possible it wouldn't matter because the SNR situation is already compromised long before any AD or DA conversion at the microphone and mic preamp.
chrissugar
- Status
- Not open for further replies.