How can the same down load sound so different?

[cpt_paranoia]

I thought I would look up the dCS Network Bridge (which is a digital rendering engine), to see if they were claiming to be using any unique techniques. I didn't find much, but scroll to the bottom of this page, https://www.dcsltd.co.uk/products/network-bridge/

where you will find this statement:

"Multi-stage power regulation is employed to isolate the sensitive clock circuitry from digital processing noise."

This is exactly what I would expect, and is entirely consistent with what I have said on this thread.

Switching supply where appropriate (for the digits), and low-dropout, low noise, linear PSU where appropriate (for the analogue). Multiple linear supplies where necessary in the analogue, for instance, the DAC clock will have its own supply. But this has to be part of a well-designed power supply and system design, with appropriate isolation between the various stages of the processing (e.g. network interface, digital rendering engine, DAC digital interface, DAC clock, DAC analogue output, analogue filtering & amplification). I really don't care too much about the first two. I start to take care with the DAC digital interface onwards.

the high speed digital switching being performed by that processor will generate a huge spectrum of noise

I'm confident that if I popped over to Cambridge and had a chat with the engineers at dCS, we'd understand each other perfectly. I'm sure we'd learn from each other. I will have some tricks I've learned from fields outside audio, and they'll have some tricks from having concentrated on the audio field.

 

[DaveVoorhis]

Why do virtually all the manufacturers of the best digital players - Aurender, Bryston, dCS, PS Audio, Sonore, etc. ALL use linear power supplies in their players? Hint: the reason is not marketing.

There may be various non-marketing reasons -- and there are some marketing reasons; you can't escape that -- for choosing a linear power supply over a switching supply in a purely digital device. There is not necessarily one single reason, either.

Player, Dave. Player. Like the dCS Network Bridge. Remember that thing with the linear power supply? Is all of this coming back again?

Your snarkism is unnecessary, juvenile, a waste of energy, and may lead to having the thread shut down. I appreciate that our views are upsetting to you, but please try to be professional, respectful, and polite in spite of it.

Contrary to Dave's lame assertion, the engineers at dCS do know what they're doing after twenty five years of development in this field.

I'm sure they do -- noting that commercial product development often involves motivations that go beyond pure technical concerns -- but that doesn't refute anything I wrote. If you believe anything I've written is a "lame assertion", then please provide appropriately scientific and/or technical counter-evidence.

By the way, counter-evidence means technically-rigorous explanations from appropriate sources. It does not include marketing bumf like the following...

"Because of our unique architecture dCS products are not constrained by conventional design approaches. As a result we are able to focus on creating innovative technologies and products that transcend the ordinary in terms of both technical performance and subjective listening.

The result? A uniquely musical experience of unrivalled emotional clarity and intensity."​

...nor does it include pointing out that there's a purely digital device that uses a linear power supply instead of an SMPS. That is not technical evidence that a linear power supply is a superior choice. It's only evidence that there exists a purely digital device with a linear power supply in it.

My "qualification" is lifetime passion for music appreciation - especially with live, unamplified music that I hear on a regular basis - and long term exposure (>30 years) to state-of-the-art gear via a couple of mentors who were audio reviewers and to my own systems with much better than average capability. Did you follow my link in #103? One doesn't have to have a degree in digital systems to discern audible differences in audio gear - analog or digital - any more than world champion Formula One drivers need a degree in mechanical engineering.

That qualifies you to tell us what you hear. It doesn't qualify you to tell us why you hear it.

 

[cpt_paranoia]

So, that makes you an expert when dealing with music reproduction?

Audio is no different to any other sort of analogue signal processing, It's not special. It's just a wiggly signal.

The same techniques, and the same care, is needed for video, sonar, radar, etc. The fact that you think it's special again shows your lack of technical understanding; it's all just analogue signal processing.

 

[E-Stat]

Audio is no different to any other sort of analogue signal processing, It's not special. It's just a wiggly signal.

Whose variations are decidedly audible. YMMV.

 

[cpt_paranoia]

and there are some marketing reasons; you can't escape that -- for choosing a linear power supply over a switching supply in a purely digital device.

Designing quiet switched mode power supplies is a skill in itself. They may not have the experience, and so may have opted for a simpler, linear supply. They will still need to isolate the digital processing noise from the other supplies (as per the quote above).

Or they may have decided that, since they need a linear supply for the analogue side, that it was simpler to use one linear supply, and use filtering to prevent the digital processor noise from coupling in to the analogue side.

Both are perfectly reasonable choices.

 

[cpt_paranoia]

Whose variations are decidedly audible

And the variations in video, radar, sonar, etc, are also decidedly discernable by the 'listener'.

 

[E-Stat]

Or they may have decided that, since they need a linear supply for the analogue side, that it was simpler to use one linear supply, and use filtering to prevent the digital processor noise from coupling in to the analogue side.

In the case of the dCS "Network Bridge" I posted, there is no analog section.

Where you find analog and digital sections, however, like that of my old GamuT CDP or my current Audio Research DAC8, they employed separate power supplies to each - appropriate for each type.

John Swenson of Sonore spends a lot of time with the specific challenges of SMPS induced leakage current and that of memory induced noise by the software itself. While you may disagree that such is important, the software in both my streamers shares concepts that support his findings: they use very low footprint Linux derived operating systems and playback software that minimizes memory and buffer activity. Bits may be bits, but there are multiple ways of introducing noise to the signal chain.

 

[DaveVoorhis]

In the case of the dCS "Network Bridge" I posted, there is no analog section.
...
John Swenson of Sonore spends a lot of time with the specific challenges of SMPS induced leakage current and that of memory induced noise by the software itself. While you may disagree that such is important, the software in both my streamers shares concepts that support his findings: they use very low footprint Linux derived operating systems and playback software that minimizes memory and buffer activity. Bits may be bits, but there are multiple ways of introducing noise to the signal chain.

On a DAC host it's desirable to limit noise, because it's easy to inject noise into the analog side of the DAC.

It's essentially meaningless on a pure digital device using buffered, error-corrected protocols, like a NAS box, Plex server, or the distant array of boxes that act as streaming sources for Amazon Music, Tidal, Google Play, Spotify, Apple Music, etc. That's because error correction and buffering eliminate any errors caused by analog noise. At worst, significant noise will result in network slowdown as damaged packets are retransmitted. That's one of the reasons we use such digital systems -- they're noise-free. Data in = data out.

It's hard not to inject noise into analog circuitry, but something has to be outright broken to inject noise into an error-corrected, buffered digital transmission.

 

[GChief]

sonar

I have been told that I am clueless about audio equipment. I guess all the sound processing, recording and reproduction Sonar stuff I operated, maintained and repaired doesn’t seem to count towards my understanding of how digital electronics work

And its not just this site.

 

[E-Stat]

It's essentially meaningless on a pure digital device using buffered, error-corrected protocols, like a NAS box, Plex server, or the distant array of boxes that act as streaming sources for Amazon Music, Tidal, Google Play, Spotify, Apple Music, etc.

My experimentation with the Sonore microRendu and RPi/DigiOne over the past two years using different power supplies doesn't support that sweeping assertion. Even the DC cord made a difference in low level dynamics with the LPS-1 power supply.

Clearly audible improvements were found. My experience is not isolated.

 

[elcoholic]

That appears to be notionally an Apple equivalent to Microsoft's RDP. If so, (and almost certainly if not), it's purely digital. Video is intercepted at or before the framebuffer, audio is intercepted before the source device's audio DAC, and transmitted to the Apple TV.

Aside from whatever device differences might affect network reliability, latency, speed, and perhaps selection of audio file format to play, the source device power supplies, network interfaces, OS types or versions, and source DACs will have no effect.

Dave, Intuitively I agree with your first paragraph. I think your second boils down to all the digital processing and hardware before the DAC doesn’t effect SQ. If so, then all digital transports with the same DAC should sound identical but clearly the marketplace and users says otherwise.

I have another example. If digital error correction is absolute then it follows that live NFL broadcasts should also be identical up to the HDMI port on my TV. I stream all TV using Roku. I have 2 ways to stream any given game - either my live TV streaming package or the individual network's Roku channel. Eg. Local Fox ch 11 or FoxNow, NBC ch 4 or NBCsports, CBS ch 2 or CBS All Access, etc. On any given game on any of the 4 networks the picture quality is always superior on the dedicated streaming channel, say FoxNow vs. Fox Ch 11. Furthermore there's a time differential of almost 30 seconds between them when using 2 TV's and 2 Roku's in adjacent rooms. So we have a digital feed from the truck at the stadium and consistently repeatable differences in received picture quality that can only occur in the digital processing and transmission realm. Furthermore this is observed over all 4 of my available networks. Clearly end to end digital error correction is not infallible. Even if there are multiple links and technologies in the path, satellite, terrestrial RF and broadband cable from my ISP each link should correct the degradation that occurs within that link and pass a bit perfect stream identical to the received applestream to the next link. Why is that not happening? If the song file = the live game feed, the Roku = AppleTV and Fox 11 = the iPhone and FoxNow = the MacBook Pro then how is that different?

 

[DaveVoorhis]

Dave, Intuitively I agree with your first paragraph. I think your second boils down to all the digital processing and hardware before the DAC doesn’t effect SQ. If so, then all digital transports with the same DAC should sound identical but clearly the marketplace and users says otherwise.

I have another example. If digital error correction is absolute then it follows that live NFL broadcasts should also be identical up to the HDMI port on my TV. I stream all TV using Roku. I have 2 ways to stream any given game - either my live TV streaming package or the individual network's Roku channel. Eg. Local Fox ch 11 or FoxNow, NBC ch 4 or NBCsports, CBS ch 2 or CBS All Access, etc. On any given game on any of the 4 networks the picture quality is always superior on the dedicated streaming channel, say FoxNow vs. Fox Ch 11. Furthermore there's a time differential of almost 30 seconds between them when using 2 TV's and 2 Roku's in adjacent rooms. So we have a digital feed from the truck at the stadium and consistently repeatable differences in received picture quality that can only occur in the digital processing and transmission realm. Furthermore this is observed over all 4 of my available networks. Clearly end to end digital error correction is not infallible. Even if there are multiple links and technologies in the path, satellite, terrestrial RF and broadband cable from my ISP each link should correct the degradation that occurs within that link and pass a bit perfect stream identical to the received applestream to the next link. Why is that not happening?

Digital error correction is infallible, and the transmission is bit-perfect, but that infallibility is achieved by consuming bandwidth: every time bad data is identified, it has to be retransmitted to obtain "good" data. That means a given packet of data may be transmitted twice, three times, or more before it is successfully transmitted. Therefore, if there are significant network problems or inherent bandwidth limits, the error correction rate may be high enough -- and the available bandwidth low enough -- that full-resolution data causes the receiver buffers to empty faster (by playing data) than they can fill (by data transmission).

When that happens, most streaming transmission protocols recognise that the buffer will empty and instruct the sender to switch to send low-resolution data -- which takes less bandwidth but at a reduction in quality -- in order to accommodate the available bandwidth. It's still bit-perfect, but it's bit-perfect transmission of low-resolution, low-quality data.

When there are outright network failures or the available bandwidth falls to a point that even low-resolution data can't be reliably transmitted, some systems play the invalid data and cause digital distortion (e.g., digital satellite, which has no re-send capability and minimal redundancy for error correction), others pop up a "buffering" message or similar display.

 

[DaveVoorhis]

My experimentation with the Sonore microRendu and RPi/DigiOne over the past two years using different power supplies doesn't support that sweeping assertion. Even the DC cord made a difference in low level dynamics with the LPS-1 power supply.

Clearly audible improvements were found. My experience is not isolated.

What effect do you think changing the DC cord, or using a linear power supply instead of an SMPS, has on a purely digital device?

Is a DAC connected directly to the microRendu or RPi/DigiOne, say via USB or mounted onboard?

They both appear to be DAC hosts, for which it is desirable to limit noise that may be injected into the analog side of the DAC as I mentioned in post #148.

 

[E-Stat]

What effect do you think changing the DC cord, or using a linear power supply instead of an SMPS, had on the purely digital device?

Good question. I report the results. Regarding the cord, John Swenson (designer of both microRendu and LPS-1) posted his experience here about using low inductance starquad and I experimented to see if I could tell any difference. I could, but it came from an unexpectedly different direction than I've found with AC based cords - more precise rendering of dynamics. I've commented on many occasions, two in this thread alone - about the audible difference of replacing the switchers commonly used. Better linears or the Uptone ultraconductor based unit reduce a false brightness or "tizz" imparted to the top end of the spectrum by switchers. When removed, you hear more innate musical detail and transparency.

Here's the innards of the LPS-1:



With the garage system, I use a 12000 mAh Jockery phone charger battery.

Also, is a DAC connected directly to the microRendu or RPi/DigiOne, say via USB or mounted onboard?

Two different scenarios. With the microRendu, it is attached to the DAC8 via a USB to S/PDIF converter since the Audio Research DAC does not support Linux.

The RPi / DigiOne in the garage connects to a Music Hall DAC 25.3 via S/PDIF using BNC with 1.5M cable:

They both appear to be DAC hosts, for which it is desirable to limit noise that may be injected into the analog side of the DAC.

Ok.

There's nothing special about the Dell Studio server located distantly from either system where the music library is housed. It is connected to a UPS and to a Tripp Lite conditioner to prevent its noise from back washing to the AC. Connectivity is via a 50' long CAT6 to main system upstairs and WiFi to the semi-detached garage where a wired connection is impractical.

 

[cpt_paranoia]

In the case of the dCS "Network Bridge" I posted, there is no analog section.

Again, you betray your lack of understanding.

The timing clock that they're relying on to remove digital noise and jitter must be treated as an analogue signal, with great care taken with its power supply, just as you would for the DAC. It is absolutely critical to the good performance of a DAC. As I said earlier, the point at which we need to start taking care of the power supply noise is the DAC interface. The DAC interface is included in that device. They even specifically mention the care they take with the clock.

 

[DaveVoorhis]

There's nothing special about the Dell Studio server located distantly from either system where the music library is housed. It is connected to a UPS and to a Tripp Lite conditioner to prevent its noise from back washing to the AC. Connectivity is via a 50' long CAT6 to main system upstairs and WiFi to the semi-detached garage where a wired connection is impractical.

I presume it's a pure digital device? No closely-coupled (e.g., USB) analog devices like a DAC, or analog audio connection from a built-in sound card?

If so, replacing its switching power supply with a linear supply will have no effect, unless (say) the switching power supply is sufficiently defective as to cause network transmission problems.

 

[E-Stat]

The timing clock that they're relying on to remove digital noise and jitter must be treated as an analogue signal

The digital clocks must be treated as analog. Ok!

As I said earlier, the point at which we need to start taking care of the power supply noise is the DAC interface.

Which is part of the compact players fed by the linear/ultracapacitor power supplies.

I presume it's a pure digital device? No closely-coupled (e.g., USB) analog devices like a DAC, or analog audio connection from a built-in sound card?

There is a slot mounted M-Audio 2496 sound card used for the office system where the Dell is the "player" via Windows applications: XM Radio/misc internet radio on browser, Tidal application, Squeezelite player for digital library. I use a DIY run of Belden 1505F from it to an Emotiva X-100 amplifier that drives in ceiling speakers located around the house and out by the pool. The music systems are fed by LMS streaming the raw FLAC/MP3 content via network. Pad/phone/watch used as remote controls for LMS. There's no "casting" or "senders". You can turn 'em off at any time and not affect the playback queue. The exception is when I use the playback functionality of the iPeng app on the pad for a bedroom system using Shure 535 IEMs. It can access the server like any other LMS client.

I pIf so, replacing its switching power supply with a linear supply will have no effect, unless (say) the switching power supply is sufficiently defective as to cause network transmission problems.

On that we agree. Just filter its output with the Tripp-Lite from the AC.

 

[DaveVoorhis]

The digital clocks must be treated as analog. Ok!

The clock is a periodic analog pulse, typically of specific wave shape. Noise can affect the apparent period and wave shape, causing jitter.

 

[E-Stat]

The clock is a periodic analog pulse, typically of specific wave shape. Noise can affect the apparent period and wave shape, causing jitter.

Very well. Another reason to ensure lowest noise.

 

[DaveVoorhis]

Very well. Another reason to ensure lowest noise.

Yes, in a digital device that has an analog component like a DAC or ADC.

Low noise is critical for analog circuitry, because analog circuitry is sensitive to noise. Circuitry at the border of digital and analog is sensitive to noise.

Pure digital circuitry isn't.

For pure network devices without analog components -- NAS boxes, Plex servers, streaming servers, Web servers -- it makes no difference at all, up to the point where the power supply is broken and network connectivity fails outright.