Have you discovered a digital source, that satisfies you, as much as your Turntable?

[Binely]

After playing around with my settings, closing my eyes... I was happy with higher word length but lower bitrate closer to native. Very intangible with my darlington pair 2nd hand sony and basic drivers.

 

[billshurv]

word length? surely the DAC processes at max wordlength and just adds padding zeros?

 

[YashN]

Sure - some people's ears do get impressed by large multiples of oversampling, placebo works wonders.

So to you, different filters aren't audible? No difference in different filter types? No difference between the sound of a digital filter and the sound of an analogue one?

Some people do prefer NOS, and that's a good way to proceed as well, especially if you have a NOS DAC like Peter Stordiau's Phasure NOS1a.

I'm not following - what is about attack dynamics? The SQ?

What do you think we usually talk about here if it isn't SQ about audio?

In my experiments with filters I've not yet noticed any degradation caused by post-ringing. Not saying it doesn't exist just that as far as I can see its over-hyped. Pre-ringing's another matter though.

Both add something that wasn't in the original signal and smear the transients which affects the soundstage negatively.

But building new filters are the realms tradeoffs: you can minimise both types of ringing, but then aliasing issues crop up.

That's why people have been working on, analysing and listening to a lot of different filter implementations, like the linear phase filters (no phase shift but both pre and post), minimum phase filters ('apodizing' filters) and Charles Hansen at Ayre proposed one which has no pre-ringing and approximately a single cycle of post-ringing.

In my experiments with filters I've not yet noticed any degradation caused by post-ringing. Not saying it doesn't exist just that as far as I can see its over-hyped. Pre-ringing's another matter though.

Good for you if you've tamed post-ringing, well at least if that's the case. What do you listen for when you change parameters for post-ringing? How do you change the parameters for post-ringing in your setup?

Audirvana has Izotope's SRC filter and a settings page where you can change a lot of parameters and even a slider to move between the post and pre-ringing characteristics.

So, now you write that you noticed degradation with pre-ringing. Don't you think that the different filters which deal with pre-ringing differently in the digital chain at different sample rates could have an effect?

 

[YashN]

Memory is DDR3 in all PCs these days, layout is pretty similar from one to another...

DDR3 is a whole subject in itself, as is how you access memory for buffering at the software level. This can have big audible surprises. Case in point: BugHead. Try BugHead Green, that's the simplest version of it and the coder avoided the cache because he says it sounds different if he does it differently. What do you hear?

Why the noise in PCs is a surprise shocks me😱

What gave you the impression that "noise in PCs is a surprise"?

Galvanic isolation, that's the way to do it problem goes away.... I would never use a PC directly without some isolation preferably wireless. As long as the digital data gets to the DAC (a trivial problem these days) then if you are neurotic about pS of jitter re-clock it... Job done.

Well, have you done it? Wireless isn't a panacea: it has its own set of issues. It does help to some extent.

Bit-perfection is a requirement, but far from sufficient: you can set iTunes to play in bit-perfect mode, but Audirvana will still trump it in quality every single time.

You seem to think it's enough that the data gets to the DAC. That doesn't even begin to address several issues that happen at the DAC.

You talk about re-clocking: isn't that what I mentioned above? I even specifically mentioned the Uptone Audio USB Regen which is on my wish list as future improvement.

Putting a linear supply on a PC wont do that much, PCs generate their own noise just by working and most PCs are designed to a cost, even the devices for memory control etc. are designed to minimise the number of layers so PCs by their very nature are quite noisy, but many use sound cards with no reported problems......

On the contrary, a lot of the noise that occurs with switching power supplies is reduced with linear ones. Many people have reported great results with linear PS instead of switching with their Mac Minis for instance. Additionally, what makes you think linear supplies are only useful with PCs...?

We're not talking about using internal sound cards here: we're talking about getting great audiophile results using an external DAC and optimising that digital chain.

 

[YashN]

Riiight. This is at face value a very daft statement. A DAC is a custom chunk of silicon, not a microprocessor.

"Riiight": the daftest I've seen is things 'looking better' than other things 'on paper', and additionally, judging formats or reproduction chains based on compression used in mastering for each. That's a mastering issue, not a format nor a reproduction chain issue.

In many cases it runs at the same speed whatever you feed it. In others it doesn't but you cannot use a PC analogy to think about it.

Incomprehensible.

 

[YashN]

After playing around with my settings, closing my eyes... I was happy with higher word length but lower bitrate closer to native. Very intangible with my darlington pair 2nd hand sony and basic drivers.

What I've found in my own system is similar: I'd rather provide the DAC with the higher number of bits it requires instead of letting it work to do the padding itself. The sweet spot here in my rig is 24-bits.

 

[abraxalito]

Could you please point out where in the PCM1704 datasheet is that data ?

I was writing from memory and had forgotten that the way I figured out the PCM1704's measured performance fell at higher sample rates was by comparing against the PCM1702 which came before it. This earlier DAC was characterized at 8X fs (353kHz) whereas the PCM1704 is tabulated at 768kHz.

Comparing the PCM1704 with the 1702, the 0dBFS figures are better (-102dB vs -100dB) - this tells me that the static performance (bit weighting accuracy) has improved in the more recent part. But then at the lower stimulus levels (-20dB is the only level comparable as for some reason the -60dB data has been omitted in the PCM1704) the PCM1704 doesn't score much, if any advantage.

Take the -K grade - the 1704 @-20dB turns in -84.4dB whereas the 1702 is at -84dB. No significant difference given that no fractions of a dB are being provided in the 1702's data. I think there should be a slight advantage going to the 1704 as its being fed 24bit data, cf 20bits for the 1702. I put this degradation of the 1704's superiority down to running at twice the sample rate and speculate that it would beat the 1702 if compared at the same rate. But I'm open to other interpretations of this comparison.

 

[abraxalito]

So to you, different filters aren't audible?

How are you drawing such a deduction? I haven't made comparisons of different filters myself but I'd tend to think they would be.

Some people do prefer NOS, and that's a good way to proceed as well, especially if you have a NOS DAC like Peter Stordiau's Phasure NOS1a.

I don't consider the NOS1a a NOS DAC myself seeing as in my understanding it runs the chips at a multiple of the 44k1 sample rate with upsampling performed on the PC. Perhaps I have a fundamental misunderstanding of Peter's design though as I've always found his prose almost totally impenetrable.

Good for you if you've tamed post-ringing, well at least if that's the case. What do you listen for when you change parameters for post-ringing? How do you change the parameters for post-ringing in your setup?

Are you assuming I'm using a digital filter? Its analog so perhaps for 'parameters' you mean 'number of inductors/capacitors in the circuit' ? I design the filters in the frequency domain and accept the post ringing that shows up.

So, now you write that you noticed degradation with pre-ringing.

Nope, I didn't write that because I haven't noticed it myself - but then I haven't listened for it either as I have no setup which includes it. But my belief is that its much more likely to be audible than post-ringing.

Don't you think that the different filters which deal with pre-ringing differently in the digital chain at different sample rates could have an effect?

Certainly.

 

[marce]

What creates a lot of noise in a PC, the PC itself working, a PC with a good SMPS is as good as a PC with a linear supply. Simultaneous switching noise is a major cause of noise, hence DDR3 will do spread spectrum.
DDR memory access changing the sound!!!! come on......
Get the data to the DAC, filter or isolate any noise, jitter is ONLY a CONCERN at the point of conversion, at the DAC. These are solvable engineering problems....
I also believe that USB2 cables should be 90 ohms differential mode characteristic impedance and ready made ones are best.
I use squeezeboxes, my PC server is in another room, its internal noise has NO effect on my music replay, the data gets there correctly no problem.


USB isolators are either complex FPGA/CPLD based that can do higher bit rates or Adum4160 based (the simpler ones) these are limited to USB 2 full speed 12Mbps, the better ones will also either isolate and filter the power (a good option) allow for a bespoke PSU or totally isolate the power and feed the 5Vs from an on board supply.

 

[YashN]

What creates a lot of noise in a PC, the PC itself working, a PC with a good SMPS is as good as a PC with a linear supply.

Yes, the PC creates tons of noise - that's what I've been saying isn't it. Now you start to qualify using 'good SMPS', but most off-the-shelf SMPS aren't good at all.

DDR memory access changing the sound!!!! come on......

Everything does when there's RFI/EMI and ground plane noise involved. Have you tried BugHead Green? What do you hear? You can also try changing parameters with XXhighEnd.

Get the data to the DAC, filter or isolate any noise, jitter is ONLY a CONCERN at the point of conversion, at the DAC. These are solvable engineering problems....

Nobody said they are not Engineering problems, nor that they aren't solvable.

What I wrote was precisely that you need to work on your digital chain to make it sound good. Jitter is a concern at the point of conversion, hence why any type of processing near the DAC prior to D/A and which has an effect by inducing jitter has to be avoided or reduced.

This said, there are things which happen from the computer which do induce jitter at the DAC clock and chip, so that needs to be worked on as well.

I also believe that USB2 cables should be 90 ohms differential mode characteristic impedance and ready made ones are best.

Experiment with building your own with a generic donor cable and hear what gives in different configurations. USB was never designed for audiophile purposes in the first place.

I use squeezeboxes, my PC server is in another room, its internal noise has NO effect on my music replay, the data gets there correctly no problem.

The data can be bit-perfect but there can still be issues. Proper digital chain optimisation is not just about getting bit-perfection. iTunes in bit-perfect mode can get the data correctly to your DAC, but will still sound less good than Audirvana.

Having physical distance between the PC and the DAC does help, but your PC is still connected directly to your DAC. If it's by USB there are ground plane issues in that configuration, and additionally, issues with noise through the cables connected to mains.

You probably have Ethernet in your configuration, that helps as well.

In my future configuration, I will have a much smaller device than a computer as a 'client' in client-server mode as a network-attached device using Ethernet, using HQ Player in both. The device will be something similar to a Beaglebone board.

You use a Squeezebox. Do you know John Swenson?

USB isolators are either complex FPGA/CPLD based that can do higher bit rates or Adum4160 based (the simpler ones) these are limited to USB 2 full speed 12Mbps, the better ones will also either isolate and filter the power (a good option) allow for a bespoke PSU or totally isolate the power and feed the 5Vs from an on board supply.

Here again, isolators and re-clockers are necessary, I already said that. I think Swenson doesn't like the Adums. I'd rather get the USB Regen he worked on. One day.

 

[YashN]

How are you drawing such a deduction? I haven't made comparisons of different filters myself but I'd tend to think they would be.

Because I was talking about filters when you replied with 'placebo'. The subject was filters, their implementations and audible characteristics.

I don't consider the NOS1a a NOS DAC myself seeing as in my understanding it runs the chips at a multiple of the 44k1 sample rate with upsampling performed on the PC.

So, the DAC doesn't do the OS. I don't find any naming issues here unless NOS is supposedly defined as restricted to keeping things at 44.1.

Perhaps I have a fundamental misunderstanding of Peter's design though as I've always found his prose almost totally impenetrable.

I believe you're right in how the DAC functions (I don't own one so I don't know the details of it), but the more important question here is why does he and customers think it sounds better this way, i.e. the OS is done computer-side and not near the D/A in the DAC. What does that bring to the whole listening experience.

Yes, he may be hard to read, but if you take the time to actually understand what he says, it makes a lot of sense. He has a lot of knowledge and experience when it comes to computer playback and DAC implementations and everything within that chain.

Are you assuming I'm using a digital filter? Its analog so perhaps for 'parameters' you mean 'number of inductors/capacitors in the circuit' ? I design the filters in the frequency domain and accept the post ringing that shows up.

I made no assumptions, that why I asked questions. Yes, the parameters are anything that you can vary as a designer or as an end-user with the filters. If its your own analogue filters, then it's the components you use to change its characteristics. If it's a digital filter, then you can use different parameters, like Audirvana allows on the Izotope SRC page. Why, was there any other way to interpret that?

I think time-domain is extremely important.

Nope, I didn't write that because I haven't noticed it myself - but then I haven't listened for it either as I have no setup which includes it. But my belief is that its much more likely to be audible than post-ringing.

Worth a lot of experiments and listening sessions.

 

[billshurv]

"Riiight": the daftest I've seen is things 'looking better' than other things 'on paper', and additionally, judging formats or reproduction chains based on compression used in mastering for each. That's a mastering issue, not a format nor a reproduction chain issue.



Incomprehensible.

OK well let me start again. You seem to think that a custom logic block changes in some fundamental operating parameter depending on how much upsampling it is doing. As the logic is clocking away at the same rate whatever is coming in This does seem daft. Where is the mechanism for a change?

I have no idea what the rest of your reply is supposed to mean. But you do seem to have a strong belief in things things and do not wish to discuss or potentially learn from others here?

 

[YashN]

Are you assuming I'm using a digital filter? Its analog so perhaps for 'parameters' you mean 'number of inductors/capacitors in the circuit' ? I design the filters in the frequency domain and accept the post ringing that shows up.

Ok, but "what do you listen for?" is still unanswered.

So, when you design in the frequency domain, what is it your paying attention to aurally for the results?

What are you listening for?

 

[abraxalito]

I'm paying attention to how the reproduced music affects my emotions - i.e. listening satisfaction. Not paying particular attention to any specifics aurally, though if something comes to my attention that'll be significant. Perhaps you're assuming I'm doing analytic listening - I don't do that. I'm listening purely for pleasure.

 

[Binely]

Lively discussion. Suggestion. I defy you to find audible artifacts that show up regularly in a fairly common DC offset output on a DAC, fed into an amplifier that uses a common balancing scheme and adequate notch filtering.

I cannot detect any discernible difference using USB vs. SPDIF to the same dac using the same bitrate/word length settings.

Your mileage will vary, certainly but if your ears can discern bit perfection versus common component mismatches then I wonder if I'm talking to the first documentable case of AI, or Canines with the ability to type.

I say this with all respect, such a difference is in the same category as the clock fluctuations impacting playback. The main reason I have is that it is not discernible unless your reproduction equipment is extraordinarily poor.

To be able to detect this or the former is to say you are able to detect the difference in timing/effect down to .007 seconds or less which while some contend is possible I would love to see that statistical proof.

It becomes sympathetic angular deflection whilst observing rotational movement at a distance.

 

[YashN]

I'm paying attention to how the reproduced music affects my emotions - i.e. listening satisfaction. Not paying particular attention to any specifics aurally, though if something comes to my attention that'll be significant. Perhaps you're assuming I'm doing analytic listening - I don't do that. I'm listening purely for pleasure.

I don't see why listening for specifics can prevent the enjoyment or satisfaction of the whole.

Can you describe post-emotion what the characteristics are that make the listening satisfying?

I think therein is a lot of information, and not just from you, but also from those people who say they prefer their turntable rig, and those people who say they prefer their digital rig:

"What are they listening for?"

This IMO is a key question where delving into the details, although they may be system-dependent and subject-dependent, can bring a lot of good.

Some people prefer vinyl because of some characteristics that can be formulated precisely rather than a vague reference to emotions or pleasure. These people are focusing on characteristics in digital rigs that decreases their aural pleasure, to such an extent that they can live with the loud clicks and pops with vinyl.

Some others would rather try to reproduce the benefits of analogue like great dynamics, soundstage and rhythm, but cannot stand the clicks and pops (I find the clicks and pops annoying - they bring me out of the music, making my enjoyment less: when they occur, they make me think of the quality of the playback chain rather than focusing solely on the musical performance).

What is it in the sound that makes it pleasurable for you? Frequency reponse bandwidth? What are the details?

 

[YashN]

OK well let me start again. You seem to think that a custom logic block changes in some fundamental operating parameter depending on how much upsampling it is doing. As the logic is clocking away at the same rate whatever is coming in This does seem daft. Where is the mechanism for a change?

You're becoming increasingly difficult to understand and I do understand Peter Stordiau when he writes (well most of it anyway).

I have no idea what the rest of your reply is supposed to mean. But you do seem to have a strong belief in things things and do not wish to discuss or potentially learn from others here?

Maybe you should ask yourself this question? And answer it to yourself?

 

[YashN]

Spot on. Offline upsampling of files helped this part of the spectrum very noticeably in a very basic desktop PC configuration, for me.

I found a lot of positives in that area as well. I found the high-end detailed and clear. I also paid attention to reverb tails, and generally how the sounds decay. Overall dynamics and the effect on soundstage height, width and depth are also interesting to focus on.

Currently, after working on my digital rig, the dynamics are very good: listening can make you want to get up and dance, or if you remain seated, you still either tap the toes or mark the rhythm with the hands and so on. 😀

 

[abraxalito]

I don't see why listening for specifics can prevent the enjoyment of satisfaction of the whole.

Since I've not listened out for anything specific about the sound, I don't know if it does prevent enjoyment of the whole. I might listen more closely to a particular instrument at a particular time, but that's because my attention's been drawn to that instrument, not that I made a conscious decision to listen to it a priori.

Can you describe post-emotion what the characteristics are that make the listening satisfying?

I don't know what 'post-emotion' means here. I doubt that I could - its like asking, when presented with a beautiful girl, what are the characteristics which make her beautiful. Beauty is holistic appreciation as far as I'm concerned.

What is it in the sound that makes it pleasurable for you? Frequency reponse bandwidth? What are the details?

From (only) my personal perspective this does seem like an inappropriate question to ask. Its about being transported by the performance - does the system do the same job as going to the original concert? When listening to live musicians 'bandwidth' and 'frequency response' don't enter into the picture so why should they when there's decent reproduction taking place? I don't make a point to listen out for particular details of the sound at a concert, its the overall effect I go for.

 

[billshurv]

You're becoming increasingly difficult to understand and I do understand Peter Stordiau when he writes (well most of it anyway).



Maybe you should ask yourself this question? And answer it to yourself?

3rd attempt before I add you to the ignore list. why should the DAC behave any differently depending upon the level of oversampling it is operating at? You claim it has 'less work' to do and sounds better. I do not see a mechanism for this. If you 'prefer' to upsample and don't claim any technical advantage we can park this and forget about it.