No Jan - what you are talking about is the _blue_ trace ("Messkoppler" - measurement coupler) the red trace is "Messplatte" (measurement record).
The text explains that the the rumble (blue) on modern turntables is substantially lower (and near to the measurement limit - black trace) than
the noise from "the best measurement records" (red trace).
In short: The _red_ trace is what is possible with (quality) vinyl.
No, I am taking the RIAA curve into account. Plug the Goldring numbers into the handy-dandy thermal noise calculator on my website (which gives both RIAA and flat weighted S/N) and see what you get.
In the meantime, take a look at the spectra of CD players that Stereophile publishes. Noise floors at -130dB are quite typical.
In the meantime, take a look at the spectra of CD players that Stereophile publishes. Noise floors at -130dB are quite typical.
SY, I think the measurement (made with an AP) I have linked to is clear. It´s 100dB above
300 Hz for a vinyl record (red trace). That´s just a fact whatever numbers your calculator
might give.
That the noise floor of a CD player with zero signal is lower is not really relevant here.
(And does not translate to the resolution achievable with 16bit).
It was about PMAs statement that using vinyl would deliberately reduce resolution,
and people might find that more pleasing.
If needed I can translate the german text for you.
300 Hz for a vinyl record (red trace). That´s just a fact whatever numbers your calculator
might give.
That the noise floor of a CD player with zero signal is lower is not really relevant here.
(And does not translate to the resolution achievable with 16bit).
It was about PMAs statement that using vinyl would deliberately reduce resolution,
and people might find that more pleasing.
If needed I can translate the german text for you.
Last edited:
The -130dB noise floor IS with a signal. I think the problem is that you're using a non-standard definition of S/N, then trying to compare it to numbers that are obtained using the standard definition. Remember, the S/N in standard terminology is obtained by integrating the noise floor over the bandwidth.
Again, I urge you to look at actual measured spectra presented in Stereophile reviews.
Again, I urge you to look at actual measured spectra presented in Stereophile reviews.
Now you are shifting to CD Players which was not _my_ intention.
(I´m absolutely aware that you can extend the AD/DA resolution with
dither for _repetitive_ signals.)
What I wanted to show is that vinyl is not really taking away resolution and thus more
pleasing to the ear, as PMA suggested ("... loss of details is preferred. This loss of details is achieved by: - higher background noise (tubes, vinyl records, tape hiss)...").
(I´m absolutely aware that you can extend the AD/DA resolution with
dither for _repetitive_ signals.)
What I wanted to show is that vinyl is not really taking away resolution and thus more
pleasing to the ear, as PMA suggested ("... loss of details is preferred. This loss of details is achieved by: - higher background noise (tubes, vinyl records, tape hiss)...").
But PMA is exactly right- you will ALWAYS have an order or magnitude (or two) more noise in a phono system, and the spectra you linked to (from a site selling phono stuff) confirm that. Remember, you have to use the same criteria for comparisons- if you want to talk about resolution for non-repetitive impulse signals in one, you need to do the same for the other. If you want to talk about resolution for tones in one, you need to do the same for the other. If you want to talk about S/N, you have to have signal, right?
Consistent definitions in comparisons are vital.
Well it only shows up to 500 Hz, I have a similar measurement at home
up to 20kHz (can scan it in the evening) and the noise floor does not
rise to higher frequencies.
And again: It´s _not_ about comparing this to the noise floor of CD,
it´s just that 100dB should be enough for not supressing any detail
on a real world recording.
Last edited:
Wrong, they don´t sell anything except for their publication.
The tests are made by the german "Testfactory" which is a DIN/EN ISO 9001:2008
certified laboratory. I have no reason to doubt their results.
And again and again: It´s _not_ about comparing this to the noise floor of CD,
...
Last edited:
Come on, how often do I have to repeat that ? It was about PMAs post:
http://www.diyaudio.com/forums/anal...rch-preamplifier-part-ii-744.html#post3425755
and that I don´t see how a vinyl record would obliterate detail present on a
real recording.
http://www.diyaudio.com/forums/anal...rch-preamplifier-part-ii-744.html#post3425755
and that I don´t see how a vinyl record would obliterate detail present on a
real recording.
Peter Walker did that in his Quad adverts several decades ago. He used a power amp (303?). This unit is now frowned on by 'audiophiles' (not sufficiently 'musical'?) yet 10 in series sound indistinguishable from one, except for a rise in noise floor.fas42 said:
Two possible conclusions:
1. it doesn't noticeably damage the signal, but audiophiles prefer some damage
2. it damages the signal in such a way that doing it 10 times does not add any more damage (it is difficult to imagine what type of damage this could be, as any smooth non-linear distortion or frequency filtering would be cumulative)
GMAFB ( http://www.diyaudio.com/forums/soli...s-blowtorch-preamplifier-176.html#post1600857 )
Same crummy arguement as the 100's of opamps during the recording routine.
Whether one jams 1 cocktail-stick through his eyeball, or 10, both proceedings lead to blindness.
Anyone who is not able to differentiate between a 33-303-esl63 and the same with a decent power amp swap, is also totally deaf.
Yes, and similarly true.jacco vermeulen said:
Your argument appears to be that once removed, something cannot be removed again. What is removed? And how can it be 'replaced' again further down the chain by a 'musical' amplifier? To do this would require detailed knowledge of what has been removed, which is difficult to arrange when nobody seems to know even approximately what it is.
Thank you for your kind words.
Not meant as unkind, but do you read books only or sometimes actually listen ?
I've had numerous hours with the 33/303/63 combo, despite still admiring Mr Walker for his amplifier handywork, glad to have skipped Quad amps for the last quarter of a century.
Doesn't require an 'audiophile' for that.
I've had numerous hours with the 33/303/63 combo, despite still admiring Mr Walker for his amplifier handywork, glad to have skipped Quad amps for the last quarter of a century.
Doesn't require an 'audiophile' for that.
99.9% of audiophiles all over the world do not understand noise, do not understand FFT and confuse noise with noise spectral density (FFT bottom). See explained on attached white noise example.
Another example:
you measure -150dB noise floor bottom with 48KHz sampling and 65536 FFT points. FFT resolution is 0.73Hz and difference between "noise bottom" and integrated noise over 24000 Hz is:
20 log(sqrt(24000/0.73)) = 45dB. Noise over 24kHz band is -105dB then and SNR is 105dB, not 150dB. SNR 150dB is only after filtering with 0.73Hz bandwidth. Imagine FFT as tuned narrow band filter, this is what you see as FFT noise bottom.
Another example:
you measure -150dB noise floor bottom with 48KHz sampling and 65536 FFT points. FFT resolution is 0.73Hz and difference between "noise bottom" and integrated noise over 24000 Hz is:
20 log(sqrt(24000/0.73)) = 45dB. Noise over 24kHz band is -105dB then and SNR is 105dB, not 150dB. SNR 150dB is only after filtering with 0.73Hz bandwidth. Imagine FFT as tuned narrow band filter, this is what you see as FFT noise bottom.
Attachments
Originally Posted by fas42
Okay, string a set of 10 of such devices in series, with suitable, "transparent", attenuation separating them as necessary. Would it still be "reliably transparent"? If not, why not, and would it be straightforward to pick up this reason by measurement?
There an article from Stuart in Linear Audio Vol 2, where he did a controlled listening test with opamps in series. He found that the audibility limit was somewhere at 5 or 6 in series.
This was in direct comparisons so in normal listening situations the audibility will be much less.
jan
Okay, string a set of 10 of such devices in series, with suitable, "transparent", attenuation separating them as necessary. Would it still be "reliably transparent"? If not, why not, and would it be straightforward to pick up this reason by measurement?
There an article from Stuart in Linear Audio Vol 2, where he did a controlled listening test with opamps in series. He found that the audibility limit was somewhere at 5 or 6 in series.
This was in direct comparisons so in normal listening situations the audibility will be much less.
jan
I won't be purchasing the magazine just to peruse that one article, so in simple terms was the reason because of an audible increase in noise floor; this is one of the typical reasons given ...
Frank
- Status
- Not open for further replies.