Do high-end streamers make an audible difference?

[ENIGMA]

1 hour ago, aussievintage said:

 

 

I am thinking of being reactionary, and getting a steam powered DAC.

They did exist, but a bit hard to find now and in mono only.

 

 

Edited January 28 by ENIGMA

[Chigurh]

I have tried various digital sources through my Luxman DAC. Streamers, CD players, BD players, smart phones, tablets and laptops. I have never heard any discernible difference. Just my experience (and opinion I suppose).

[Satanica]

11 hours ago, Assisi said:

I consider that different streamers will deliver a different file to the DAC.  Therefore the outcome from the DAC will be different depending on  what is in or not in the file besides the music information that it receives from different streamers.

 

Unless intentional (e.g. digital attenuation adjustment) which means a streamer/source has deliberately changed the file/bits, there is no reason for this to be at all true. I'll go as far to say your statement is a complete mind fabrication, intentional or not.

Edited January 29 by Satanica

[Keith_W]

Indeed. It's like asking a streamer to play Beethoven and getting Rick Astley instead. I wouldn't put it beyond some of those nefarious designers to deliberately do this though ... 

 

And @sir sanders zingmore well played! 

[ENIGMA]

1 hour ago, Keith_W said:

Indeed. It's like asking a streamer to play Beethoven and getting Rick Astley instead. I wouldn't put it beyond some of those nefarious designers to deliberately do this though ... 

 

And @sir sanders zingmore well played! 

I can't even get Rick Astley - just Kylie Minogue.

[Assisi]

2 hours ago, Satanica said:

 

Unless intentional (e.g. digital attenuation adjustment) which means a streamer/source has deliberately changed the file/bits, there is no reason for this to be at all true. I'll go as far to say your statement is a complete mind fabrication, intentional or not.

I understand completely why you have interpreted what I said.  My statement should have been much more precise as to what I meant by different especially the first sentence that you quoted.  My apologies for being clumsy with what I said.  I should turn my mind on more when I write posts.  I do consider the second sentence to overall be correct.

 

The music information aspect of the digital file is not different at all.  I consider though, that the file including the music signal can depending on network components also include other unwanted information resulting from resonance or interference from various sources.  That additional information can end up as noise along with the music that you hear.  Less resonance etc will ensure a lower noise floor and enhance the music listening experience.  Some streamers/players will deal with and mitigate the resonance better than others.

 

John

[StereoNET]

Just a gentle reminder of our guidelines for this community. Let's be respectful of other opinions please, even if we do not necessarily agree with them. And above all else, ensure you are not making it 'personal'.

[Satanica]

29 minutes ago, Assisi said:

The music information aspect of the digital file is not different at all.  I consider though, that the file including the music signal can depending on network components also include other unwanted information resulting from resonance or interference from various sources.  That additional information can end up as noise along with the music that you hear.  Less resonance etc will ensure a lower noise floor and enhance the music listening experience.  Some streamers/players will deal with and mitigate the resonance better than others.

 

If there was added noise to the digital out signal of a streamer then it would not be bit perfect.

The DAC portion of streamers might vary in noise, but that's a separate topic and not what the thread starter inquired about.

[Stereophilus]

I hope the OP can forgive me here, but I think some disambiguation might assist the discussion further.

The question posed is: do high-end streamers make an audible difference.

 

The problems with answering this question include the following:

1. What is a high-end streamer?

2. are we to generalise across all “high-end” streamers, in all situations? 

3. Is the audible difference in relation to the OP,  in the OP system? 
 

Not to be overly pedantic here, but we have some issues here with definition and perspective.  This does tend to encourage a fairly broad range of opinions as a result.

 

Let’s ignore the “high-end” part of the question, we’ve been there before trying to define this term and it mostly comes down to pejoratives.

 

It also makes sense (I think) to consider the question from a more open minded perspective to help us draw broader conclusions.  So, to paraphrase, could a different streamer sound audibly different in the same system to the same observer? The self evident answer here is yes! Because we aren’t excluding any subjective biases.

 

What if we ask the question from an objective perspective… could a different streamer create a measurably different analog output from a DAC in the same system? Again, the answer is a self evident yes! As has been explained in this thread, some streamers allow for DSP which will measurably (and probably audibly) change the analog output of a DAC.

 

Again, to aid our discussion we might go a bit further… could different streamers with measurably identical outputs sound audibly different in the same hifi system if we remove (or discount) subjective biases?

 

I think this last question is what a lot of posters are answering, or trying to answer.  I’m not aware of any studies or tests that definitively answer this question directly.  The theory does suggest a vehement NO! as the answer.  However I do have some thoughts around the difficulty in answering this last paraphrased question.

 

Firstly, in my opinion, it is difficult and/or expensive to design and conduct studies that reliably exclude subjective biases in listening tests.  There is a whole SNA thread on this subject.

 

Secondly, as mentioned in this thread previously, streamer output format could play a role in any perceived audible differences.  That is, a synchronous spdif output from one streamer could (in my opinion) sound audibly different from an asynchronous USB output on another streamer.  These 2 different outputs could also sound audibly different from the same streamer. Why? Well, the synchronous output derives the timing from the streamer, whereas asynchronous USB is clocked at the DAC end.  Could different clocking (in spite of buffering) make an audible difference? My experience suggests that it does, so I think it’s worth a mention here.

 

Then there is also the question of streamer software.  Even if we ignore DSP, there are a huge number of forum reports indicating different streamer software making a (subjectively) audible difference when everything else is kept the same.  Not proof by any stretch, but a valid line of inquiry.  An example could be an antipodes device which allows the user to select different server and player software.  Yes, I have tried this too, and I also consider that software can make an audible difference, but I’m not sure why it could or should.  In my situation I can run Roon Core as server and switch between a different software endpoint on the same hardware - Roon endpoint or HQplayer endpoint. No DSP is engaged with either, but the differences are audible to me when switching quickly between the two.

 

Finally, in passing I will suggest that aside from the above, power supplies and isolation in streamers may make a contribution to any audible differences.  Ground loops and ground noise entering a DAC from a streamer could potentially result in audible differences.

[Addicted to music]

1 hour ago, Assisi said:

.

 

The music information aspect of the digital file is not different at all.  I consider though, that the file including the music signal can depending on network components also include other unwanted information resulting from resonance or interference from various sources.  That additional information can end up as noise along with the music that you hear.  Less resonance etc will ensure a lower noise floor and enhance the music listening experience.  Some streamers/players will deal with and mitigate the resonance better than others.

 

John

 

Digital was designed to be immune to noise or anything that marketing can convinced you to spend more.  The zeros and ones even at there DC defined levels,  has little bearing on what is required at the receiving end.  Nearly all digital receivers require an active positive/negative transition to shift from zero to one or a one to zero state to be recognized, from this requirement its literally immune to any analog noise you can think off.  As long as all network components are utilized within there specifications,  its literally bit perfect.

An example:  if i asked a device to send the digital word 255 255 255 for RGB.   The receiver its going to get 255 255 255.  No questions asked, regardless of jitter, EMI RF noise that exist in the whole digital chain.  its then up to the receiving device especially the output how that accurately generate 255 255 255 for RGB.  In your buffer or memory storage its going to store the values in digital 255 255 255 for RGB.    If this is not the case there are error correction built in to either correct the data and asked for the data to be resent, no noise or jitter is ever processed or stored!  Edited:  its unlike a analog tape recording.

 

Edited January 29 by Addicted to music

[bob_m_54]

21 hours ago, Assisi said:

I consider that different streamers will deliver a different file to the DAC.  Therefore the outcome from the DAC will be different depending on  what is in or not in the file besides the music information that it receives from different streamers.

If the streamer delivers something other than what is input to it, then it is defective. A streamer that can't deliver an un-adulterated file is absolutely worthless.

[bob_m_54]

7 hours ago, Assisi said:

The music information aspect of the digital file is not different at all.  I consider though, that the file including the music signal can depending on network components also include other unwanted information resulting from resonance or interference from various sources.  That additional information can end up as noise along with the music that you hear.  Less resonance etc will ensure a lower noise floor and enhance the music listening experience.  Some streamers/players will deal with and mitigate the resonance better than others.

By what you call "The music information aspect of the digital file" I'm assuming you mean the data. If the device receiving the data cannot discriminate between signal pollution (noise) and the data, then the either the induced noise is so high it's unusable, or the receiving device is defective. "Resonance" plays no part in digital data transmission. Where did resonance even come into the discussion?

[bob_m_54]

6 hours ago, Satanica said:

 

If there was added noise to the digital out signal of a streamer then it would not be bit perfect.

The DAC portion of streamers might vary in noise, but that's a separate topic and not what the thread starter inquired about.

Why wouldn't it be? Unless the noise is so great that it obscures the data completely?

[Assisi]

6 hours ago, Addicted to music said:

Digital was designed to be immune to noise or anything that marketing can convinced you to spend more.  The zeros and ones even...

 

I have a several questions about aspects of the above with the digital and it being immune to noise.  Is the Immunity only with USB or network or both?  When was the design for immunity from noise first implemented in the context of digital?  Is it possible that things have changed from the design implementation so that the immunity is no longer appropriate in the context of improved or more sophisticated audio components?  Does there need to be a redesign or understanding of how the digital signal is transmitted and treated?

 

My limited understanding is that the digital audio signal in the network is more than just 1s and 0s. The digital signal is not electrical.  The signal is always carried through the network as an analogue waveform representing the digital data.  Many things such as noise can become part of the analogue representation until the signal is converted to an analogue signal in the DAC.  Partial or mitigation of the noise in the signal can be appreciated through a lower noise floor.  To me lower noise floor is the Holy Grail of audio playback using a network and a DAC.  Once experienced there is no going back to what was.

 

Some posters may agree with the above whilst I am sure others will not. 

John

[Assisi]

13 minutes ago, bob_m_54 said:

If the streamer delivers something other than what is input to it, then it is defective. A streamer that can't deliver an un-adulterated file is absolutely worthless.

Pleas read post above when I acknowledge my clumsy wording.

John

[Assisi]

35 minutes ago, bob_m_54 said:

By what you call "The music information aspect of the digital file" I'm assuming you mean the data. If the device receiving the data cannot discriminate between signal pollution (noise) and the data, then the either the induced noise is so high it's unusable, or the receiving device is defective. "Resonance" plays no part in digital data transmission. Where did resonance even come into the discussion?

Resonance can unfortunately be all over digital.  The important thing is to avoid it by various strategies if you can.  It might be useful to listen to a system where the resonance is minimal and does not prevail so that you can appreciate the benefits.  It will be wonderful.

John

[Satanica]

40 minutes ago, bob_m_54 said:

Why wouldn't it be? Unless the noise is so great that it obscures the data completely?

 

Any change to the data interpreted is considered non bit perfect, that includes noise, audible or not. 

[bob_m_54]

The definition of the difference between digital signals and analog signals is that (in a perfect world) digital signals will be voltage levels with instantaneous voltage level changes (ie zero rise and fall times), whereas analog signals have constantly variable voltage levels.

 

Digital data in it's perfect form is either differential voltage levels, or voltage transitions from one level to another (+ to -, or - to +). And yes it's electrical.. But due to the physical imperfections of data transmission media, you won't have perfect leading and trailing edges on the pulses, so you get mishapen square waves (rounded corners or ringing).

 

The thing is, modern data transmission systems are very robust, so that you have to have a very distorted signal, before it gets to the stage where you start losing data. When this happens, the different protocols will have various methods of data correction. So you either have a successful data transfer or you don't. You can't get changes in data, you either get a successful packet, or a packet fail and a resend.

 

The acoustic properties of the file are completely contained within the data of the file. The only difference in how the file sounds is how the DAC handles the file, and the analog section of the DAC

 

long story short.. it works or it doesn't.

 

 

 

[bob_m_54]

2 minutes ago, Assisi said:

Resonance can unfortunately be all over digital.  The important thing is to avoid it by various strategies if you can.  It might be useful to listen to a system where the resonance is minimal and does not prevail so that you can appreciate the benefits.  It will be wonderful.

John

Explain resonance in the context of digital data transmission please.

[bob_m_54]

4 minutes ago, Satanica said:

 

Any change to the data interpreted is considered non bit perfect, that includes noise, audible or not. 

Yes, any change to the data would preclude it being bit perfect, but noise, audible or not, wouldn't necessarily change the data. And if it did, error correction would reject the packet and call a resend.

[The Mad Scientist]

27 minutes ago, bob_m_54 said:

Yes, any change to the data would preclude it being bit perfect, but noise, audible or not, wouldn't necessarily change the data. And if it did, error correction would reject the packet and call a resend.

When a well-designed DAC FFT demonstrates Signal to Noise Ratio in the output of >130 dB, why are we even concerned about noise?

 

I also think misunderstandings are being made regarding digital. Noise can't be carried with the data. The data is reconstructed and buffered as the original 1s and 0s in memory at each network interface. It can't carry analogue noise within it.

 

If it's being claimed that the physical  connecting wires allow noise to travel between the digital and analogue components via the grounding arrangements, then firstly, ground continuity would be required between devices. Use of the correct U/UTP ethernet cables and no higher than Cat6 architecture would prevent this.

Secondly, for noise to affect DAC conversion/output at an audible level, it would have to either be so high as to corrupt the interpretation of bits at the DAC input (unlikely, as that would result in TCP/IP checksum failures and packet loss upstream), or it would have to be on the DAC ground rail at such a level to be audible. SNR would have to dip to ~ 50 to 60 dB for this to occur in a typical listening room. Noise floor of ~ 30 dB(A) and listening levels of ~80 - 90 dB(A).

Edited January 29 by The Mad Scientist

[Anton Mullet]

I was amazed at the jump in sound quality going from a Volumio Primo( 1st generation) streamer to an Eversolo DMP A8. 

And this is running both direct to my dac. 

Before the Eversolo I just thought streaming was to blame as my CD player( Redgum) always sounded better. Streaming sounded a bit crappy. Now there is no gap in sound quality.  It just comes down to the quality of the digital media or version of what's being streamed. 

I'm imagining the cheaper power supply and other circuit and isolation differences are to blame. 

A huge bonus is a much better UI. The Volumio one was always clunky. 

I imagine Volumio has improved as my unit is the first Primo. 

But damn now the Fiio S15 has come out at half the RRP price of the Eversolo. 

I got mine for $2300 on sale just before Xmas but that's still $700 more.  

 

[Keith_W]

1 hour ago, Assisi said:

 

I have a several questions about aspects of the above with the digital and it being immune to noise.  Is the Immunity only with USB or network or both?  When was the design for immunity from noise first implemented in the context of digital?  Is it possible that things have changed from the design implementation so that the immunity is no longer appropriate in the context of improved or more sophisticated audio components?  Does there need to be a redesign or understanding of how the digital signal is transmitted and treated?

 

My limited understanding is that the digital audio signal in the network is more than just 1s and 0s. The digital signal is not electrical.  The signal is always carried through the network as an analogue waveform representing the digital data.  Many things such as noise can become part of the analogue representation until the signal is converted to an analogue signal in the DAC.  Partial or mitigation of the noise in the signal can be appreciated through a lower noise floor.  To me lower noise floor is the Holy Grail of audio playback using a network and a DAC.  Once experienced there is no going back to what was.

 

Some posters may agree with the above whilst I am sure others will not. 

John

 

OK, time for another analogy. Let's say you have a light switch. It can be either on or off. Let's say that you wiggle the light switch. This is noise. As long as the wiggling does not close the circuit and turn the light on, it does not matter how much you wiggle it. While the light switch itself is analog, in that there is a continuous state in the position of the "up" and "down" position of the physical switch, the outcome is digital - the circuit is either closed, or it isn't. So while the digital signal can be thought of as "analog" pulses going down the wire, the outcome is not analog. The receiver (in this case a DAC) either sees a 1 or a 0, and nothing in between. 

 

There is a certain threshold where noise is tolerated - if the light switch wiggling becomes so large that it inadvertently closes the circuit, then we have data corruption. Data corruption is most certainly audible, and it is NOT subtle. You will hear a buzz, or click, or the music might stop. It does not manifest as "wider soundstage" or "blacker blacks" or anything like that. It is music interruption. 

In this way, digital tolerates noise to a certain extent. If there is a lot of noise, then it will stop working. You may have noticed this if you listen to digital FM radio. Drive into a basement and the signal weakens, then you start hearing digital artefacts. 

Streamers and wi-fi is not the same as digital FM radio though. For one, there is buffering, and there is 2-way communication. If there is packet loss, it requests the packet again. If the buffer becomes starved, then you will start hearing music interruption. 

[Addicted to music]

2 hours ago, Assisi said:

 

I have a several questions about aspects of the above with the digital and it being immune to noise.  Is the Immunity only with USB or network or both? 

 

Digital fundamental was introduced and used in way back in the 1930-40.  It was used to perform basic operations.   

The Networking that you refer TCP/IP  was developed in 1970 by a division of the US Defence.  

USB was introduced 1996.

 

Because 2 states: zero and ones are used its known as binary operation.  The 2 state zero and one disregards any levels in between, so it doesn't care what is between unless its a zero or a one value.  By implementing this it disregards any values in between to a certain specification.   

 

 

2 hours ago, Assisi said:

 

  Is it possible that things have changed from the design implementation so that the immunity is no longer appropriate in the context of improved or more sophisticated audio components?  Does there need to be a redesign or understanding of how the digital signal is transmitted and treated?

 

 

 

Digital Fundamentals apply to everything that uses and involve binary,  digital audio falls into this category,  it's never changed.  Zeros are zeros and ones are ones,  anything in between is disregarded.  Boolean Algebra that's used for binary calculations also used to design/ developed digital circuits.  If you want better audio via a digital system,  you need to increase the bit length:  Going from 16 to 32bits etc, this will also increase resolution and lower noise floor.   If you look at the ESS Delta Sigma marketing it explained the benefits of 32bit to 16 bit resolution.

 

2 hours ago, Assisi said:

 

 

My limited understanding is that the digital audio signal in the network is more than just 1s and 0s. The digital signal is not electrical.  The signal is always carried through the network as an analogue waveform representing the digital data.  Many things such as noise can become part of the analogue representation until the signal is converted to an analogue signal in the DAC.  Partial or mitigation of the noise in the signal can be appreciated through a lower noise floor.  To me lower noise floor is the Holy Grail of audio playback using a network and a DAC.  Once experienced there is no going back to what was.

 

 

 

This is how its market to you to justify there products.   

 

Digital receivers are immune to most noises such as RF, EMI etc  Unless the noise falls into the specification where the receiver identifies and responds  is rare.   And if it does to interrupt and cause errors, the receiver will know as the checksum indicate it's in correct.   As in previous posts, your DAC doesn't respond to EMI RF and even if the noise floor is increased it doesn't care as long as it sees legible data is what you will hear.

Edited January 29 by Addicted to music

[Addicted to music]

1 hour ago, Keith_W said:

 

OK, time for another analogy. Let's say you have a light switch. It can be either on or off. Let's say that you wiggle the light switch. This is noise. As long as the wiggling does not close the circuit and turn the light on, it does not matter how much you wiggle it. While the light switch itself is analog, in that there is a continuous state in the position of the "up" and "down" position of the physical switch, the outcome is digital - the circuit is either closed, or it isn't. So while the digital signal can be thought of as "analog" pulses going down the wire, the outcome is not analog. The receiver (in this case a DAC) either sees a 1 or a 0, and nothing in between. 

 

There is a certain threshold where noise is tolerated - if the light switch wiggling becomes so large that it inadvertently closes the circuit, then we have data corruption. Data corruption is most certainly audible, and it is NOT subtle. You will hear a buzz, or click, or the music might stop. It does not manifest as "wider soundstage" or "blacker blacks" or anything like that. It is music interruption. 

In this way, digital tolerates noise to a certain extent. If there is a lot of noise, then it will stop working. You may have noticed this if you listen to digital FM radio. Drive into a basement and the signal weakens, then you start hearing digital artefacts. 

Streamers and wi-fi is not the same as digital FM radio though. For one, there is buffering, and there is 2-way communication. If there is packet loss, it requests the packet again. If the buffer becomes starved, then you will start hearing music interruption. 

 

Another way to look at it is compare the old analog PAL TV transmission to current digital HD TV transmission.  In the old analog TV transmission the noise is very evident, you see ghosting,  snowing  and color bleeding if  interference was present.  In digital HDTV,  these effects are non existent.  if the signal is low youll get pixelation and finally  stops working.   digital is precise,  i want 255 255 255 RGB on pixel 1028 at XX time, You'll get exactly 255 255 255 RGB  on pixel 1028 at that XX moment and nothing else.