Why would the phase noise have a different impact at the DAC output than any other noise? Wait, here comes the "non stationary" nonsense, correct?
Until you care to put some numbers around your statements, they remain technobabble. Meantime, check your hearing: if you can discriminate between two signals separated by 0.01Hz it would certainly qualify as supernatural.
1Hz different should be maskable, according to a sim done on it by HpW. I do not know exact details of how HpW did the sim, but I trust him and his result. Others may feel differently. A link to that discussion was added to my original post, and is repeated here: ESS Sabre Reference DAC (8-channel) ...Sim result and HpW comments follow a few posts later in #2656.
Only if you could explain what has a simulation to do with the audibility of such effects. Last time I've checked, nobody claimed that clock phase noise doesn't translate into artifacts in the DAC output time domain.
What Leeson's empirical equation shows is generally known as "excess [phase] noise". 1/f noise with 10dB/decade is the minimum allowed by physics.
It is interesting that jitter can easily be calculated from the phase noise, but not the other way around. So translating a jitter requirement to a phase noise requirement is far from being trivial.
I applaud Andrea's drive to chase phase noise down into the utterly inaudible region. It really doesn't matter if he surpasses the audible limit by x10 or x100, just that it is achieved and we can then consider this aspect of playback solved.
Nothing exceeds like excess.
Nothing exceeds like excess.
Just take a look at the phase noise plots I have published, the phase noise below 10 Hz from the carrier has a slope of 30 dB/decade.
And it's one of the best oscillator available nowadays, for a lower grade oscillator the 1 / f3 noise starts at greater frequency.
So below 1Hz from the carrier and below we are in the 1 / f3 region and not in the 1 /f region.
This is your subjective opinion, as neither you nor I are able to measure the effect.
Just speculations.
Sorry, I was unclear... "minimum allowed" in the sense that noise cannot get lower that this limit. Excess noise is always measured as the noise coming on top of the flicker noise (1/f).
But fact is, we are talking here effects that are below any audibility threshold. Those claiming they can hear at these levels are required to provide proof (not that I'm holding my breath waiting).
Phase noise is rather easy to measure, if you have access to the right equipment (I know I don't). The Agilent N5511 has a noise floor of -200dBc and can do offsets as low as 0.01Hz, and that for base signals from 50KHz to 40GHz (more with external mixers). In general, phase is among the easiest things to measure in EE.
The 1 / f3 noise is imposed by the quality of the crystal and it's usually due to its random walk.
I still don't know if it is possible to perceive the difference between an oscillator with a phase noise of -150dBc at 10 Hz compared to one with a phase noise of -127 dBc at the same distance from the carrier (I will know soon).
What I can assure is that the difference between a Crystek and an oscillator with a better phase noise around 30 dB is clearly audible.
What is the relationship between the distance a sound can travel, its duration and how loud it is ?
Which Crystek and which "oscillator with a better phase noise"? And care to share the data regarding audibility?
Most entities don't publish research details these days. No more Philips technical notes, little from other companies. The world has changed.
In the world of audio, Bruno Putzey's does not publish listening test details, nor does ESS.
For example, ESS says they have shown audiophiles can audibly distinguish SD dacs from conventional dacs. For more info see: Noise Shaping Sigma Delta DACs - ESS Technology, Inc. ...Starting at page 28.
EDIT: ESS also claims audiophiles rate as inferior systems that have variable excess phase. Starts on page 39 of the above presentation.
They also say that standard FFT measurements (like some people seem to think is all that matters) will not show an audible non-steady state noise problems, see page 31.
In the world of audio, Bruno Putzey's does not publish listening test details, nor does ESS.
For example, ESS says they have shown audiophiles can audibly distinguish SD dacs from conventional dacs. For more info see: Noise Shaping Sigma Delta DACs - ESS Technology, Inc. ...Starting at page 28.
EDIT: ESS also claims audiophiles rate as inferior systems that have variable excess phase. Starts on page 39 of the above presentation.
They also say that standard FFT measurements (like some people seem to think is all that matters) will not show an audible non-steady state noise problems, see page 31.
Last edited:
Crystek CCHD-957 at 22.5792/45.1584 MHz replaced with emitter coupled oscillators with SC-Cut crystal at 11.2896/22.5792/45.1584 MHz.
The Crystek is -97 dBc at 10 Hz from the carrier while the emitter coupled at 11.2896 MHz is -127 dBc.
Regarding audibility I have no data to share, as I said I don't own an instrument to measure the effect at DAC output.
I can only share my listening impressions and the audibility impressions from other members on this forum.
A few links:
The Well Tempered Master Clock - Building a low phase noise/jitter crystal oscillator
The Well Tempered Master Clock - Building a low phase noise/jitter crystal oscillator
The Well Tempered Master Clock - Building a low phase noise/jitter crystal oscillator
The Well Tempered Master Clock - Building a low phase noise/jitter crystal oscillator
The Well Tempered Master Clock - Building a low phase noise/jitter crystal oscillator
I think the point is exactly this: trying is believing.
Just try yourself.
I have a technical mentality, but where the measures do not reach I rely on my ears.
As an example, the measurements do not explain why a tube amplifier with 1% distortion sounds much better than a solid state amplifier with 0.0001% distortion.
Yet the difference is clearly audible.
We can speculate on the reasons but not prove it with measurements.
Understood, no data to share. You seem very quick to extrapolate your subjective findings, though. We call this "anecdotical proof".
Are you sure I have not tried? Perhaps I did, in a controlled way, hence my skepticism?
Psycho acoustics explains very well why some prefer a SE tube amplifier with 1% distortion over a 0.0001% distortion solid state amplifier. Otherwise, the Wavac amplifier won't have a market and a Stereophile great review by Mr. Fremer Wavac SH-833 monoblock power amplifier | Stereophile.com
Such a preference is perfectly fine, my only objection is about calling these amplifiers "HiFi" or "transparent". Me, I prefer to build an effect box rather than relying the amplifier to do it for me.
Just a simple thread search with keywords "impressions", "sound" and so on.
It looks like the diyaudio thread search performs very well and quick.
I can't afford a gear to measure and share data (if one exists), but you can buy one yourself if you are looking for such that measurements.
But if you have tried, please share your measurements data, or your impressions if you have no data to share.
There is a beautiful thread on comparing measurements vs sound quality.
Scott Wurcer and John Curl have opposite opinions.
I'm with John Curl.
I prefer to use a linear device like the vacuum tube rather than a non-linear device like the transistor or the mosfet and then apply tons of negative feedback.
Last edited:
That is so funny. I guess you have not seen syn08s lab? Or his threads that are full of measurements.
Sorry, I haven't seen any of this.
But exactly what measurements are you talking about?