duckworp
pfm Member
Since going down a Roon-based digital streaming music path I have been exploring the world of digital audio. I work professionally in music and one of my systems is a home listening studio room for playback of audio masters. So to help me do my job my aim was to get ripped and Qobuz files to sound as good as the same CD (ie same master), played through the same DAC. To test this I play a CD and an identical ripped file and an identical Qobuz file through the same DAC and compare via switching 'source' on a remote. When I first installed my streaming system, CD gave me much better SQ than streaming. Through adding bits and tweaking things I have reached a point where streaming is now sounding better, both from a ripped file and Qobuz.
Given that I have tried a lot of things I thought that I would report findings to see whether this concurs with other listeners' experiences, or possibly to help other people new to this world. Obviously these are my findings, in my system, listening with my ears, and so others may have come to very different conclusions.
Ideally the ASR disciples on PF will agree to disagree and resist chipping in about confirmation bias and the laws of physics.
These changes I describe were consistent across three different DACs, all of which are decent DACs, so differences heard are unlikely because of a poor DAC.
It is amazing that in digital audio it seems that a lot of things make a difference, sometimes only marginal, but added together with other small changes and you have a big change. If you undo all the small things you have added then you hear a big difference. It is what makes the evolution of a digital system so interesting in that it is easy to change and improve and successfully tinker. Often at low cost too. It's a very exciting time to be creating a digital system: experimentation can be unexpectedly rewarded and ASR-style science is not fully explaining the differences being experienced.
I have attempted to categorise the changes by level of SQ improvement that I hear in my system. I am ignoring the DAC - the DAC is more significant than any other single item here, so I’ll assume a decent quality DAC. The component that feeds the DAC I have called a 'network bridge'.
Things that made the largest difference:
- Network bridge settings - If relevant, I suggest turning off unused settings on the network bridge (I am specifically thinking of the Pro-Ject Stream Box S2 Ultra here). If you have the option to do so then turn off WiFi and hdmi and anything you don’t use. Leaving them on seems to degrade the performance.
- Grounding Box - Using a grounding box to attach to the DAC and Network bridge. I use the CAD GC1. It won the Stereophile and HiFi+ award for best accessory this year and I can hear why. Differences are not subtle.
- A decent network switch. This makes a significant improvement. I have tested audiophile switches, cheap £20 switches, but the best in value terms for my ears is a second hand Cisco 2960. These are top quality switches which years ago were made in the millions and, even with the benefits of mass production, cost hundreds of pounds back in the day as they are so well made. For £50 you can pick one up on eBay. Aside from the Chromecast Audio, these must be the biggest bargain in digital audio. I cannot hear a difference between my Cisco and an audiophile switch costing £1000+. But I can hear a clear difference between the Cisco and a cheap £20 Netgear switch.
- EtherREGEN- Adding an EtherREGEN just before the network bridge. This is a remarkable little switch-type device made by UpTone Audio that takes an Ethernet or optical input and outputs Ethernet to the network bridge. I have no idea how it works, but it really does. Reading the ASR review of this was the final nail in the coffin re ASR - they measured it doing nothing, but there are so many people vouching for the effect this little unit has that it is obvious their measurements are missing something.
- I would recommend not plugging any devices with a switch mode power supply into a dedicated Hi-Fi power circuit if you have one. I plug those devices into my regular house mains circuit. If I plug the smps power supplies into the Hi-Fi circuit then the Hi-Fi circuit loses a lot of its increased performance.
Things that made a significant difference:
- USB cable from network bridge to DAC. The results of changing this are variable in my systems. Many sound identical, even pricey ones, but some sound significantly different. And it is not always cost-related. A friend of mine makes them for himself using a technique he developed, just for a few pounds, and his are amongst the best I tried. The Supra USB cable is the best sounding low price one I tried. Of the pricey ones, I have only tried a few but not many sounded any better than the Supra. The only ones that did were a cable by CAD and one by Phasure, both really took the sound up a level. Though my mates home-made one was nearly as good as these.
- Ethernet cable - I found cheap as chips cat-8 is best for long runs. Though I have not tested audiophile quality on my long stretch as it is too pricey for the 10m I need. Unfortunately my house was wired up with cat-5e so I was hoping that the type of Ethernet made no difference. But unfortunately it made a big enough difference that now I have thick cat-8 cable over the carpet.
For the short run of cable between the EtherREGEN and the network bridge I have experimented with audiophile cable vs cheap patch cable. As with USB cables some Ethernet cable made a positive difference, some a negative difference, some no difference. I tried a few and the cable by Phasure is excellent with its options to try different shielding on the one cable, but after that I found basic shielded cat 8 sounded as good or better than any other unshielded or shielded audiophile or non-audiophile cable. I have only tried 4 audiophile cables though. On the Naim forum they report a blind test of short final stretches of Ethernet cables and the SOtM cat 7 cable won easily, but I have not heard that.
- Network bridge. My experiments have determined that the sweet spot is a sub £1000 bridge from someone like Pro-Ject (Stream box S2 Ultra) or SOtM (SMS-200 Neo Ultra) or Stack Audio (Link 2). These all sounded identical to me and identical to a £3k Auralic Aries G1. But they all did sound significantly better than a cheap home-built Pi or a Chromecast Audio (which, I would add, at £25 is still the biggest bargain in Hi-Fi ever).
- Using Optical instead of copper Ethernet from Cisco switch to EtherREGEN. This is better than the cat 8 Ethernet but the difference is less marked than going from cat 5 to cat 8. I am currently testing this but my wife is not keen on the bright orange optical cables on the carpet! Apparently you can roll your SFP modules to alter the sound. I have not tried this. With the Cisco switch you are fairly limited to the modules which will work.
Things that made smaller improvements:
- Adding an external power supply to the network bridge. An improvement but I didn’t get the large effect that some listeners describe.
- Adding an external clock to the network bridge. Again this made a small improvement but not as much in my system as some people report in theirs. Maybe my DACs are not so sensitive to an upstream clock?
- Having Roon on its own NUC server. I ran it on my home PC at first. But it is worth giving it a dedicated NUC or MacMini to run on. Not only does Roon work quicker but it sounds better too.
- LPS on DAC - One of my DACs has a switch mode power supply (the Chord Qutest). Putting an external Linear PS on this made a small difference, but not enough for me to actually invest in one.
- Length of USB cable - using two lengths of identical cable my friend made it was obvious shorter distances sounded better.
- a ripped CD sounds a little better than the same mastering played via Qobuz
Things that made no difference:
- Adding an external PS to my Roon NUC.
- Adding an external PS to my router
- The quality of the Ethernet patch cable between router and Cisco switch
- The quality of the Ethernet cable between Roon NUC and switch
- The quality of the bnc cable between clock and network bridge
- The type of physical isolation given to any device in the chain prior to the DAC made no difference. They are as happy on the carpet as on expensive feet. The DAC, however, definitely likes being isolated.
- Inserting an Ethernet to optical converter instead of using the optical 'out' on the Cisco switch.
- Playing files from an external hard drive attached to the Roon core vs playing from a drive directly attached to the network bridge, bypassing Roon.
- Playing files from the internal drive of the roon core vs playing files from an attached usb hard drive.
- Ripping to WAV vs ripping FLAC vs ALAC
- Using an audiophile CD drive to rip vs using a cheap CD drive
Given that I have tried a lot of things I thought that I would report findings to see whether this concurs with other listeners' experiences, or possibly to help other people new to this world. Obviously these are my findings, in my system, listening with my ears, and so others may have come to very different conclusions.
Ideally the ASR disciples on PF will agree to disagree and resist chipping in about confirmation bias and the laws of physics.
These changes I describe were consistent across three different DACs, all of which are decent DACs, so differences heard are unlikely because of a poor DAC.
It is amazing that in digital audio it seems that a lot of things make a difference, sometimes only marginal, but added together with other small changes and you have a big change. If you undo all the small things you have added then you hear a big difference. It is what makes the evolution of a digital system so interesting in that it is easy to change and improve and successfully tinker. Often at low cost too. It's a very exciting time to be creating a digital system: experimentation can be unexpectedly rewarded and ASR-style science is not fully explaining the differences being experienced.
I have attempted to categorise the changes by level of SQ improvement that I hear in my system. I am ignoring the DAC - the DAC is more significant than any other single item here, so I’ll assume a decent quality DAC. The component that feeds the DAC I have called a 'network bridge'.
Things that made the largest difference:
- Network bridge settings - If relevant, I suggest turning off unused settings on the network bridge (I am specifically thinking of the Pro-Ject Stream Box S2 Ultra here). If you have the option to do so then turn off WiFi and hdmi and anything you don’t use. Leaving them on seems to degrade the performance.
- Grounding Box - Using a grounding box to attach to the DAC and Network bridge. I use the CAD GC1. It won the Stereophile and HiFi+ award for best accessory this year and I can hear why. Differences are not subtle.
- A decent network switch. This makes a significant improvement. I have tested audiophile switches, cheap £20 switches, but the best in value terms for my ears is a second hand Cisco 2960. These are top quality switches which years ago were made in the millions and, even with the benefits of mass production, cost hundreds of pounds back in the day as they are so well made. For £50 you can pick one up on eBay. Aside from the Chromecast Audio, these must be the biggest bargain in digital audio. I cannot hear a difference between my Cisco and an audiophile switch costing £1000+. But I can hear a clear difference between the Cisco and a cheap £20 Netgear switch.
- EtherREGEN- Adding an EtherREGEN just before the network bridge. This is a remarkable little switch-type device made by UpTone Audio that takes an Ethernet or optical input and outputs Ethernet to the network bridge. I have no idea how it works, but it really does. Reading the ASR review of this was the final nail in the coffin re ASR - they measured it doing nothing, but there are so many people vouching for the effect this little unit has that it is obvious their measurements are missing something.
- I would recommend not plugging any devices with a switch mode power supply into a dedicated Hi-Fi power circuit if you have one. I plug those devices into my regular house mains circuit. If I plug the smps power supplies into the Hi-Fi circuit then the Hi-Fi circuit loses a lot of its increased performance.
Things that made a significant difference:
- USB cable from network bridge to DAC. The results of changing this are variable in my systems. Many sound identical, even pricey ones, but some sound significantly different. And it is not always cost-related. A friend of mine makes them for himself using a technique he developed, just for a few pounds, and his are amongst the best I tried. The Supra USB cable is the best sounding low price one I tried. Of the pricey ones, I have only tried a few but not many sounded any better than the Supra. The only ones that did were a cable by CAD and one by Phasure, both really took the sound up a level. Though my mates home-made one was nearly as good as these.
- Ethernet cable - I found cheap as chips cat-8 is best for long runs. Though I have not tested audiophile quality on my long stretch as it is too pricey for the 10m I need. Unfortunately my house was wired up with cat-5e so I was hoping that the type of Ethernet made no difference. But unfortunately it made a big enough difference that now I have thick cat-8 cable over the carpet.
For the short run of cable between the EtherREGEN and the network bridge I have experimented with audiophile cable vs cheap patch cable. As with USB cables some Ethernet cable made a positive difference, some a negative difference, some no difference. I tried a few and the cable by Phasure is excellent with its options to try different shielding on the one cable, but after that I found basic shielded cat 8 sounded as good or better than any other unshielded or shielded audiophile or non-audiophile cable. I have only tried 4 audiophile cables though. On the Naim forum they report a blind test of short final stretches of Ethernet cables and the SOtM cat 7 cable won easily, but I have not heard that.
- Network bridge. My experiments have determined that the sweet spot is a sub £1000 bridge from someone like Pro-Ject (Stream box S2 Ultra) or SOtM (SMS-200 Neo Ultra) or Stack Audio (Link 2). These all sounded identical to me and identical to a £3k Auralic Aries G1. But they all did sound significantly better than a cheap home-built Pi or a Chromecast Audio (which, I would add, at £25 is still the biggest bargain in Hi-Fi ever).
- Using Optical instead of copper Ethernet from Cisco switch to EtherREGEN. This is better than the cat 8 Ethernet but the difference is less marked than going from cat 5 to cat 8. I am currently testing this but my wife is not keen on the bright orange optical cables on the carpet! Apparently you can roll your SFP modules to alter the sound. I have not tried this. With the Cisco switch you are fairly limited to the modules which will work.
Things that made smaller improvements:
- Adding an external power supply to the network bridge. An improvement but I didn’t get the large effect that some listeners describe.
- Adding an external clock to the network bridge. Again this made a small improvement but not as much in my system as some people report in theirs. Maybe my DACs are not so sensitive to an upstream clock?
- Having Roon on its own NUC server. I ran it on my home PC at first. But it is worth giving it a dedicated NUC or MacMini to run on. Not only does Roon work quicker but it sounds better too.
- LPS on DAC - One of my DACs has a switch mode power supply (the Chord Qutest). Putting an external Linear PS on this made a small difference, but not enough for me to actually invest in one.
- Length of USB cable - using two lengths of identical cable my friend made it was obvious shorter distances sounded better.
- a ripped CD sounds a little better than the same mastering played via Qobuz
Things that made no difference:
- Adding an external PS to my Roon NUC.
- Adding an external PS to my router
- The quality of the Ethernet patch cable between router and Cisco switch
- The quality of the Ethernet cable between Roon NUC and switch
- The quality of the bnc cable between clock and network bridge
- The type of physical isolation given to any device in the chain prior to the DAC made no difference. They are as happy on the carpet as on expensive feet. The DAC, however, definitely likes being isolated.
- Inserting an Ethernet to optical converter instead of using the optical 'out' on the Cisco switch.
- Playing files from an external hard drive attached to the Roon core vs playing from a drive directly attached to the network bridge, bypassing Roon.
- Playing files from the internal drive of the roon core vs playing files from an attached usb hard drive.
- Ripping to WAV vs ripping FLAC vs ALAC
- Using an audiophile CD drive to rip vs using a cheap CD drive
