|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
In Reply to: if the S/N ratio reduces from 130dB to 120dB (for example) ... posted by Christine Tham on January 30, 2003 at 11:51:24:
Maybe you don't read Stereophile.
Follow Ups:
have you ever measured the SPL of "ambient" noise in a quiet room? the figure is something like 30-40dB! even in my room (and i live in a quiet street) it can rise to 50-60dB with the fridge turned on (from a different room behind closed doors)let's say you are listening to music at rock concert like levels of 110dB. you are getting an overall S/N ratio of effectively 60-70dB at best.
in a worst case scenario (air-con, projector turned on) the S/N ratio may well drop to 30-40dB
this is all old news to mixing engineers which is why you will seldom hear content mixed at below 30-40dB (an exception is pat metheny who says he deliberately introduces content at -40-50dB for headphone listeners)
the S/N ratio on my old amp was only around 80dB. yet i have never heard the noise floor on this amp, even when cranking the volume all the way to 11 :-) (of course, there is noise if my ears are right next to the speakers, but not at normal listening position.)
so i very much doubt (if the player is performing to "spec") that the high frequency noise in sacd is detectable. of course, this is not to say that the noise may not induce artefacts (amplifier circuits oscillating, speaker colouration, phase shifts, whatever) that are detectable. but these are all equipment limitations.
Once the signal is analog you're correct Christine, because you can hear well below the noise floor. However, a 90 dB S/N ratio is not as good as CD - and S/N ratio is critical with digital sources..
sony/phillips deliberately contoured the noise shaping curve of sacd so that it matches or equal cd s/n ratio in the worst possible scenario, which means typically the s/n ratio is better than cd.
SACD players have a lower S/N ratio than CD above 10 kHz or so, or JA is misleading us. How can you possibly misundertand that?
as i recall (which may be incorrect), the cd digital audio format was designed with a goal of 90dB S/N ratio.the fact that many cd players exceed the design goal is due to techniques such as oversampling.
what i stated was that i believe sony/phillips designed dsd to be no worse than the cd spec (ie. 90dB) in the worst case.
in which case, any s/n ratio better than 90dB is "per spec" in my opinion.
i could of course be wrong and will be happy for someone to point this out. but until then, i am not concerned with the results in stereophile since i don't believe they subjectively impact the listening experience (as stereophile themselves point out).
It shows that DSD has a lower noise/resolution above 10kHz or so than a 16 bit pcm signal.Redbook players with 1 bit dac's and noise shaping show the same noisespectrum but with a rising curve in a higher frequency range.
That's because the 1 bit dac's usually oversample at 128 or 256 times.
in the stereophile review of the sony scd-xa777es, jan 2002 edition, page 129-130, figs 3, 4 and 5 clearly show that the noise level is at around -110dB around 10kHz.yes - it rises sharply after that - to round -92dB at 20kHz in fig 4
which proves my point. these are all "within spec" as far as i am concerned, and at these levels the resolution capability of the dacs and analog circuitry are limiting factors.
what i do concur is that sony/philips marketing blurbs about frequency response to 50 kHz are clearly misleading - the signal is pretty junky above 20kHz. but that's fine, i'm not a bat.
Measurements from the same player but a rise in noise of ~6dB at 10kHz when a signal is played.Clear evidence that the noise gets modulated. And so will musical content in the upper frequency range.
This post is made possible by the generous support of people like you and our sponsors: