|
|
Post by puffin on Nov 12, 2023 20:17:33 GMT
I know nothing about the computer. It was my son's when he worked at Microsoft and I doubt that he will know. I thought that perhaps there was something upsampling the signal, but there is no CD drive and no sound card as far I can see without dismantling it. I just stream from a USB out on the motherboard.
|
|
|
|
Post by puffin on Nov 21, 2023 13:45:33 GMT
Since my last post I have done some digging and found that the stream from Spotify is being processed by ASIO, a computer sound card driver protocol for digital audio.
As I said previously the Dac is telling me that it is handling DSD. Whatever is happening it sounds pretty decent.
|
|
|
|
Post by Slinger on Nov 21, 2023 15:14:34 GMT
I know nothing about the computer. It was my son's when he worked at Microsoft and I doubt that he will know. I thought that perhaps there was something upsampling the signal, but there is no CD drive and no sound card as far I can see without dismantling it. I just stream from a USB out on the motherboard. If there is no sound card, sound would be handled by a chipset on the computer's motherboard itself. |
|
|
|
Post by orange55 on Nov 22, 2023 20:43:38 GMT
Since my last post I have done some digging and found that the stream from Spotify is being processed by ASIO, a computer sound card driver protocol for digital audio. As I said previously the Dac is telling me that it is handling DSD. Whatever is happening it sounds pretty decent. ASIO is a good thing as it bypasses the windows default sound setting that converts everything to 48hz. |
|
|
|
Post by puffin on Nov 26, 2023 18:44:49 GMT
Since my last post I have done some digging and found that the stream from Spotify is being processed by ASIO, a computer sound card driver protocol for digital audio. As I said previously the Dac is telling me that it is handling DSD. Whatever is happening it sounds pretty decent. ASIO is a good thing as it bypasses the windows default sound setting that converts everything to 48hz. I am still getting to grips with how all this streaming lark works and what is doing what. However I am thoroughly enjoying how it all sounds through my ifi Zen Dac v2 and a Topping D30 which a mate gifted me recently. Both handle DSD. |
|
|
|
Post by brettj on May 22, 2024 8:57:50 GMT
I have to use a wifi extender for my system.
Since I upgraded my wifi router, I've been able to play with a couple of settings on the extender.
This extender TPLink RE650 is only used for my system.
1. Incoming signal from the router is 2.4GHz only. I've turned off the 5GHz.
2. Turned off wifi broadcast from the extender.
Will ty this out over the weekend.
|
|
|
|
Post by MartinT on May 22, 2024 10:51:40 GMT
Good luck, Brett. As it's not a powerline sender, you could usefully plug a filter in right next to it to help reduce the switching noise pushed out to mains power.
|
|
|
|
Post by brettj on May 22, 2024 12:37:57 GMT
Good luck, Brett. As it's not a powerline sender, you could usefully plug a filter in right next to it to help reduce the switching noise pushed out to mains power. Thanks Martin. Was going to use a 250mm Meicord Opal from the extender to the second DXE Iso Plus. Then a Pink Faun into the EtherRegen |
|
Tobias
Rank: Quartet
Posts: 320
|
Post by Tobias on Sept 4, 2024 16:37:17 GMT
Could one quantify the digital noise in very rough numbers, in an attempt to visualize the importance of the different noise sources of digital audio? I am thinking something like the below, where i try to put a very rough estimate in percentage. I guess it can be presented in a much better way and the numbers are up to debate and is also an extreme generalization...  Focused streaming-only setup, without NAS, Laptop/Desktop connected: Network: 33,3% - Gear implementation 10% (the sum of network gear physically connected in the signal chain)
- Power Supply 20% (The sum of the PS of all the network gear physically connected in the signal chain)
- Ethernet cable 3% (The last one into the streamer)
Streamer (Digital transport): 33,3% - Gear implementation 10%
- Power Supply 20%
- Software and its CPU usage 3%
DAC: 33,3% - Gear implementation 15%
- Power Supply 15%
- Digital cable from streamer to DAC 3%
Note that if you target one area then the other numbers will go up (the sum is always 100%) since they are now the remaining sources of noise.
What aspects have a missed? How would you do the weighting, if you agreed on the very simplified "model"? |
|
|
|
Post by MartinT on Sept 4, 2024 16:43:06 GMT
I am not sure where you are including air gapping (moats etc.), filters and grounding boxes, all are important for digital noise reduction.
Need to think about it more before replying further.
|
|
Tobias
Rank: Quartet
Posts: 320
|
Post by Tobias on Sept 4, 2024 16:47:11 GMT
I am not sure where you are including air gapping (moats etc.), filters and grounding boxes, all are important for digital noise reduction. Need to think about it more before replying further. My thinking is that air gapping and moat is a decoupling, meaning that what is before that doesn´t need to be accounted for, since it is no-longer on the physical network and not connected to your HiFi. Having said that, i am not sure exactly what moats is... |
|
|
|
Post by MartinT on Sept 4, 2024 16:51:44 GMT
A moat (as in a castle surrounded by water) is an isolation mechanism inside some components like my EtherREGEN and Gustard U18. Somewhere I published the circuit board of an EtherREGEN where you can clearly see the physical moat separating 'dirty' from 'clean' side with only the optical coupler IC straddling it.
Remember: no air-gapping is perfect, noise can even go down fibre cable. It is only capable of noise reduction. The before in air gapping still needs to be accounted for (otherwise, how did I clearly hear the effect of a change of power supply for my router with two isolation moats further along the path?)
|
|
Tobias
Rank: Quartet
Posts: 320
|
Post by Tobias on Sept 4, 2024 17:02:50 GMT
Right, it gets complicated when we stack several noise reducing methods on top of each other to gradually reduce the noise on the network.
I was sort of thinking that a moat is part of the "gear implementation"...(even if that is one gear or 30 stacked gears).
At the end of the day it is the noise that goes out of that last ethernet cable into the streamer that matters, from a network noise perspective. (how you got there is sort of irrelevant for the outcome)
But i do agree it is a topic of its own, how you can lower network domain noise.
|
|
|
|
Post by MartinT on Sept 4, 2024 17:42:57 GMT
Ok, but counting the moats and grounding boxes in the path from router to streamer to DAC, as well as the inherent noise in each component including its power supply, is going to give you a decent appreciation of what the sound quality may be like. That's before even considering the SQ of the DAC itself.
It's very tricky, in reality, to know what a digital system will sound like. However, given a system I know how I would go about trying to quieten it.
|
|
Tobias
Rank: Quartet
Posts: 320
|
Post by Tobias on Sept 4, 2024 18:33:28 GMT
ok, i see what you are saying.
I think my model was maybe a very rough way of introducing the noise concept to someone that is trying to understand the concept, as a whole.
Ethernet noise could be extremely important to target but also maybe not important at all, if you buy a streamer that target "all" of it, like the Grimm Audio Mu1 or Mu2. (if we say that it does...?)
In addition, you could buy a very expensive filter solution, like the Muon Acoustic Streaming System, to target the bulk of the ethernet sourced noise. That might lead to the same result, or better/worse, than using a very complicated methods involving a lot of gear.
I agree, the more you think about this the trickier it gets to try to explain it an easy way.
I now feel that the concept is actually very simple and straight forward in theory but hard to achieve fully in practice, which has always been the issue with digital...
I feel that there must be some sort of analogy, table or similar, that one can use to try to visualize the noise reduction concept, on a high level, to someone.
|
|
|
|
Post by MartinT on Sept 4, 2024 19:49:24 GMT
Honestly, compared with, say, a high end analogue system for record playback, a digital system is orders of magnitude more complex.
I simply didn't appreciate that until after I had started down the road of building a system for streaming.
|
|
Tobias
Rank: Quartet
Posts: 320
|
Post by Tobias on Sept 4, 2024 20:17:57 GMT
Yes, i agree.
But don´t you agree that it is largely because the D to A conversion step is so ridiculously sensitive to any modulation of noise in the digital domain.
I mean, for ages no-one really understood what was going on in digital audio and people where guessing all over the place. This caused a lot of confusion and made digital a mystery that seemed more complex than necessary.
Only fairly recently (after the pandemic) the magnitude of the digital domain noise challenge has been logically explained publicly, by people like Hans Beekhuyzen, and by manufacturers of digital audio that has spent years trying to understand this themselves.
The concept of noise reduction is very simple, as such, just as making sure the needle can track the grooves of a vinyl record. Still it isn´t easy to achieve to perfection. (I agree that digital is even more sensitive)
I am not saying that digital noise is everything but it is a never ending quest to remove something that can´t by nature be fully removed and as the tolerances becomes smaller (as you remove even more digital domain noise) it gets even more obvious that you can target even more of it... :-D.
As you say, the sad part is that we don´t tend to understand this until we have ventured fairly deep into this subject and sometimes in the "wrong" direction, since we didn´t understand the logic. I now feel that I understand (without having tried) that buying a very very good (expensive) power supply will likely make things much better, even if my current power supply was already very expensive (for me).
I also understand that draining noise (grouding) is very important, even if i haven´t tried it yet. It just makes sense, and is actually "simple" logic, when you realize how much impact the noise has in the digital domain.
|
|
|
|
Post by NigelB on Sept 4, 2024 21:05:45 GMT
A moat (as in a castle surrounded by water) is an isolation mechanism inside some components like my EtherREGEN and Gustard U18. Somewhere I published the circuit board of an EtherREGEN where you can clearly see the physical moat separating 'dirty' from 'clean' side with only the optical couple IC straddling it. Remember: no air-gapping is perfect, noise can even go down fibre cable. It is only capable of noise reduction. The before in air gapping still needs to be accounted for (otherwise, how did I clearly hear the effect of a change of power supply for my router with two isolation moats further along the path?) I think it was ER who coined the term/concept moat and it seems to have stuck, which is a stroke of marketing genius I have to admire but it is not a term any electrician or electronics expert would recognise as far as I know. Whatever a moat is supposed to be, there needs to be a bridge over it to allow the incoming signal to reach the outgoing side so it can't by definition provide complete isolation. A physically visible separation on a circuit board doesn't really tell anyone anything about the degree of noise separation it provides.
All switches provide galvanic isolation (conceptually a moat), it's a merely a question of degrees, and the only way to reliably determine the effectiveness of this galvanic isolation (including a "moat") would be to measure the noise on the input signal and the noise on the output signal. This would not only represent a subtraction of the noise on the outgoing cable from the noise on the incoming cable, it would also take into account any noise generated by the switch's circuitry itself (eg, through the inclusion of typically noisy LEDs) and of radiated noise entering the case through ventilation holes, unused ports, LED holes etc.
Right, it gets complicated when we stack several noise reducing methods on top of each other to gradually reduce the noise on the network. I was sort of thinking that a moat is part of the "gear implementation"...(even if that is one gear or 30 stacked gears). At the end of the day it is the noise that goes out of that last ethernet cable into the streamer that matters, from a network noise perspective. (how you got there is sort of irrelevant for the outcome) But i do agree it is a topic of its own, how you can lower network domain noise. Agreed 100% about the final connection.
The more effective the final noise reduction device before the streamer is (let's say it's a switch) the less of a contribution upstream devices make.
|
|
|
|
Post by John on Sept 5, 2024 6:41:25 GMT
Reading this again it reminds me how good and practical Martin post is. For anyone wanting to get the best out of a streaming system this is great advice.
|
|
Tobias
Rank: Quartet
Posts: 320
|
Post by Tobias on Sept 5, 2024 7:55:35 GMT
I agree the post is great and probably important for many.
I believe it was several of Martin´s post here that got me thinking that digital audio maybe isn´t perfect, after all. I was the one laughing about this craziness prior to trying things out... :-D
|
|
