Oppo`s BDP105 - discussions, upgrading, mods...

[qusp]

@ zenelectro: yes, I believe its not going to be great either, nothing like a close in phase noise plot in sight, if it had good close in phase noise, they would be selling on that point and it would be priced in a different bracket. that is what I said that that got misrepresented as saying it cant sound good because its cheap, I simply said it cannot measure well or it would not be this cheap, Epson arent a charity and clocks have a pretty narrow task that is fairly well understood by them.

Remember
Sabre has internal re sampling engine to reduce jitter.

indeed, but it cant touch/improve the master clock jitter of course, since thats what it uses to reduce the incoming jitter.

its harmonics will be different part to part and interact with the digital filter selected, the input jitter of the BCK, which is quite important for the ESS, the input sample rate etc and then we have the harmonics of the gear used to play it, the jitter encoded into the mix by the ADC etc etc.

to predict that with any type of reliability... good luck. i'll take a predictable low harmonic clock any day of the week. clocks are meant to be predictable.

but you are right, these subjects are difficult to breach, Coris would get offended each time too when I said pretty much the same thing and then me being me, I escalate when people start making up reasons for technical superiority instead of accepting its a subjective quality. its equivalent to suggesting someone who likes simple low feedback circuits likes the distortion profile, only thats a bit more predictable in that even harmonics are generally more pleasing and its more Gaussian

[Joe Rasmussen]

Originally Posted by zenelectro

...but the
jitter spectrum is spread such that subjectively they sound good.

I have an open mind, at least I hope I do, but does this also explain the increased resolution?

Originally Posted by zenelectro

Unfortunately Joe took this personally.

I did?

Originally Posted by zenelectro

As I have stated previously, there also may be something in this jitter
spectrum spread that works well with Sabres internal ASRC. Remember
Sabre has internal re sampling engine to reduce jitter.

That is an interesting thought.

But again where does that increase in resolution come from. The SAW does not sound like it is adding niceness, in fact it seems to do the exact opposite. It sounds less romantic - and the initial reaction can be a little puzzling - and the longer you listen, the realisation that it a different presentation. Certainly what it does for the bass, and Coris heard this too, I started to hearing textures that I had not been aware before. Better presentation of rhythm and pace, tighter imaging but also natural as the correct size is better perceived, more air around voices and instruments, better separation of the instrument or voice from the soundspace - slap echoes from corners of the soundstage not clearly defined before. (Soundspace and soundstage is not the same thing.)

"A clock is a clock" I was told, yet what you Terry is saying is at least of interest, that the sound we hear can be affected by clocks with different spectrum spreads, does indeed have an affect. So I was ridiculed by this person (actually two of them in concert), yet I find what you say more worthwhile and I am not ignoring anything you have said. And I am not offended.

In the Sabre DAC's ACR, could this be some kind of dither effect triggered?

There is also something else, and this is the trickiest part of all, but when extreme ULF filtering is done, what I hear from that and what I hear from SAWs seems rather compatible - like it is going further down the same path. I can only report what I perceive - so I still wonder whether some further power supply isolation is achieved.

Cheers, Joe

[abraxalito]

Hi Joe - I'm fascinated by the SAW oscillators sounding better than the more expensive xtals. I have a question on this - has anyone done measurements of jitter when the xtal oscillators are in-situ in a player? Seems to me that power supplies might be playing a big part here because presumably the impressive specification for jitter with xtals is made with the best possible power supply in the lab? As a quick stab in the dark here, maybe the SAW oscillators have better PSRR so work better than xtals in typical setups where the power supply to the clock isn't ultra-clean.

[Coris]

Originally Posted by Joe Rasmussen

Hi Coris

From Epson Toyocom website:

XG-1000CA/CB (Crystal Oscillator Low-Jitter SAW Oscillator)

I have tried 100MHz, 125MHz and 50MHz from the same range - all good.

I really have no reason to doubt they will not follow the same specs - makes no sense why the shouldn't.

Specs:

XG-1000CA/CB | Real time Clock Module | Products | Epson Quartz Crystal Devices

Note: Made in Japan

Cheers, Joe

Is OK. We may just trust Epson. I use actually their 125Mhz SAW oscillator and is very good indeed.
I meant that when about large production batches sold to the big (well known) dealers companies for reselling/distribution, the factories have to follow an quite restrictive contract, when about the specified parameters and the deviation allowed for an batch components.
When about this (quite unknown custom order) 1000 components batch, is a little bit different...
But let`s have trust that Japanese Epson will hold its high level reputation for high quality components producer.

[Coris]

You sceptical guys! What for so much theories, and never ending discussion? It is not more easy and reasonable to just try out this SAW, and have those "conversations" after?
I can not understand what is so difficult tu solder an such oscillator inside your device and hear what is coming out? I did. It is different sound in the favour to this SAW type oscillator. Why? It may be found it out. But if we struggle about theoretical/abstract aspects, when the most involved people in those discussions have never ever had/heard/measured this oscillator, we do not go further with this.

[qusp]

its not theory Coris, worse jitter that is.

Oh haha sorry I see another has come up with the same explanation while grasping at straws

yes abraxalito has come up with the same idea, yes mate that was one of the theories I came up with for the irrational idea that worse could be better in some applications (easily tested by sending it to Ian). it relies on the power supply for the crystek or other clock inflicting more jitter than the difference between the basic close in phase noise.

[Coris]

Originally Posted by zenelectro

Hi Coris,

The data sheet Joe linked to doesn't really give enough information to
adequately characterize the clocks performance. Much more is required.
The reason is they are a comparatively low spec clock.

Manufacturers generally don't bother supplying phase noise and Allan
Variance plots for low spec oscillators, there's no point.

If you want to guarantee consistency, this is the way they do it and has
always been done. If they are supplying super low phase noise clocks for
space applications the customers will demand it.

The more I think about it the more I am convinced these SAW oscillators
have a fairly high jitter compared to the best crystal oscillators but the
jitter spectrum is spread such that subjectively they sound good.

I gave an example previously of comparisons between high spec oscillators
where often the lowest jitter didn't win subjectively. Unfortunately Joe
took this personally. My research on this doesn't cover just a few people
but many peoples experience over many years.

As I have stated previously, there also may be something in this jitter
spectrum spread that works well with Sabres internal ASRC. Remember
Sabre has internal re sampling engine to reduce jitter.

Hi zenelectro

I understand very well, and agree too, that SAW oscillator is specified lower than another type ones. That because is no any reason for the producers of such oscillators to specify it for another special parameters.
But over all those thing, this SAW it sounds better. I agree too with (your) theory that spreading the jitter spectrum which it may happen with this type oscillator, it may explain the subjective appreciation of an better sound.
At least one it may be interested to have a better sound out of the involved device, than having an lowest jitter level and another very high performance parameters out of an oscillator. You may agree that we do not listen/enjoy the music out of an gear with an scope or something beside, monitoring all the time the jitter, the noises, or the system`s parameters. Such situation is not normal at all, but is used in some cases when one it may be interested to find out one or another about how is working the system.
But anyway it is an challenge and interesting in the same time, to find out why an low spec oscillator it may lead to a better sound out of an DAC system. We can further work on this...

[Joe Rasmussen]

Coris, don't pay much attention to him - qusp.

Take a look at his own description on his left column: "Mean-spirited shatterer of illusions".

Rather arrogant, wouldn't you say? A self-anointed Pontifex Maximus.

Here in Australia we have a name for somebody like that, a "crap-stirrer."

In Internet Parlance, that also means he is a Troll.

[Joe Rasmussen]

Originally Posted by qusp

yes abraxalito has come up with the same idea, yes mate that was one of the theories I came up with for the irrational idea that worse could be better in some applications (easily tested by sending it to Ian). it relies on the power supply for the crystek or other clock inflicting more jitter than the difference between the basic close in phase noise.

I suppose YOU think came up with that first?

WRONG

Been saying that for years. Join the queue behind me.

See, you are not the only one who can play your ridiculous game. Now just behave yourself, make yourself useful, we shall be forgiving and forgetful... as they say, become a useful member of society.

A narrow mind starts with a narrow view, his own.

[zenelectro]

Originally Posted by qusp

@ zenelectro: yes, I believe its not going to be great either, nothing like a close in phase noise plot in sight, if it had good close in phase noise, they would be selling on that point and it would be priced in a different bracket. that is what I said that that got misrepresented as saying it cant sound good because its cheap, I simply said it cannot measure well or it would not be this cheap, Epson arent a charity and clocks have a pretty narrow task that is fairly well understood by them.

Understood qsp.

This is a tricky discussion because we have peoples products (upgrades)
being (indirectly) scrutinized as a result responses will be affected. I try
to avoid conflict but it's not so easy.

I have to say I have a certain degree of frustration, technically speaking,
trying to make sense of this whole scenario. On one hand Joe is promoting
power supplies that have super low LF noise and very good 'stability' , but
then we have a clock which is poor in this area of performance. From a
purely technical perspective it is hard to join the dots so to speak.

The second reason, which up till now I wasn't going to get in to was the
results I have had with my own clocks. However, I think it is relevant to the
discussion. Refer to the phase noise plot of one of my 'zenclocks'. As you
can see these -do- have very good close in phase noise. That is -95dBc
at 1Hz offset.

These are based on custom made 11.2896MHz OCXO. Sine wave OP, the
crystals are custom made by batch. The supplier gives me complete
set of measurements with each and every clock. I do my own squaring /
level shift, separate PS for every element. So far these have sonically beat
every other aftermarket clock on the systems tried. They are expensive.

So as you can see my results are at odds with results from SAW but there
are other issues WRT 100MHz versus 10MHz, SAW vs crystal, also power
supply 'Q' factors etc etc. I will talk later gotta scoot.