AUDIO over IP - REDNET 3 & 16 Review. AES67 Sets A New Standard for Computer Audio
Sep 18, 2017 at 7:30 AM Post #3,211 of 3,694
What were the improvements you noticed with the JCAT card?
Bigger soundstage, better imaging and separation, macro/micro details improved. What I probably like the most is the improvements I'm hearing in the clarity of the voices. I'm getting a better sense of natural voice pitch. It would be nice if someone compared the JCAT card with the Mutec MC3+ Reclocker.
 
Sep 20, 2017 at 1:35 PM Post #3,212 of 3,694
Golfnutz,

I never thanked you for your excellent suggestion.

Between your suggestion and the parts that Tommy is graciously sending me, it seems I will be in good shape.

Thanks a lot.

Joel
 
Sep 20, 2017 at 8:15 PM Post #3,213 of 3,694
Golfnutz,

I never thanked you for your excellent suggestion.

Between your suggestion and the parts that Tommy is graciously sending me, it seems I will be in good shape.

Thanks a lot.

Joel

My pleasure, very kind of you, thanks.

Just a couple of additional comments (please disregard in case Tommy covered these with you).

1. If you don't already have a Multimeter, you should purchase one. Required to measure the voltages.
2. Take lots of internal pictures of your D16 after you've removed the lid, and before you start anything else.
3. After taking pictures, I suggest you measure the voltage to get an idea of what the SMPS is putting out. Mine was 4.94v (so I adjusted my LPS to this voltage). You can do this by putting one probe on the solder joints just in front of the JST 5v connector and the other probe on any grounding point.
4. Connect the DC jack to the LPS DC cable and test the voltage at the DC jack before attaching wires from the kit Tommy sent you. Basically, you are confirming 5v from the LPS.
5. Test the modification BEFORE removing/changing anything. By this I mean, connect everything by running the wires over and into the D16 (not through it yet). No point removing/disconnecting everything first if it doesn't work, or you don't like what you're hearing. You can do the cosmetic changes afterwards.
6. If you're going to remove the IEC at some point, you will need to cut the wires attached to the prongs (they are soldered, so you have no choice). Just try to feel were you can cut them as close to the prongs as possible - this will leave you lots of wire to reconnect later if you want. You'll only need a $15 crimping tool to reattach the wires back (female crimp connectors). The prongs will have to be cleaned up as well. You could just take the IEC to a repair shop and ask them to clean it up, should be pretty cheap - 2 minute job (or buy a new IEC plug).

It's pretty simple really. If you have any questions just ask. The attached photo is where to get the voltage reading.

D16 voltage .jpg
 
Last edited:
Oct 13, 2017 at 10:25 AM Post #3,216 of 3,694
I have been wanting to try out the Mutec REF10 reference clock in my RedNet AOIP system after hearing such good things about it, especially after having found the Mutec 3+ USB to be such an excellent product. I contacted the US distributor to arrange for a trial purchase with the idea that I might do a review and post my findings. The distributor was extremely cooperative and only asked that they be allowed to fact check before I posted. They have done that and had no comments to add. I asked for the trial because I would be purchasing a piece of interface equipment that cost more than my DAC and adding it to a front end that was already over the edge. If it did not work out I wanted to be able to return it!

The Short Version:

I purchased the REF10. It is a keeper.

The Long Version:

I was hoping to get an improvement in sound quality but not to the extent that occurred. To add icing on the cake it has also allowed me to drop two pieces of equipment from my chain which had become extreme even by audiophool standards.

My system before adding the REF10 consisted of:

1. “CAPS-like” server running Windows Server 2012R2 and Audiophile Optimizer. Powered by LPS.

2. Dante Virtual Soundcard (DVS) software which allows JRMC to send the audio stream out via an ethernet port instead of USB

3. A PCIe fiber optic ethernet card with fiber running to a Fiber Media Converter(FMC) and then via CAT7 cable into the Focusrite RedNet D16 AES ethernet to AES/SPDIF converter.

4. Antelope LiveClock as external wordclock to the D16

5. AES cable out from D16 to Mutec 3+ USB interface acting as a re-clocker.

6. AES cable out to a second Mutec 3+ USB interface acting as a second re-clocker.

7. AES out to Schiit Audio Yggdrasil DAC.


This is the order in which I tested the REF10 and coincidentally the order in which I preferred the sound. The biggest jumps were step 2 and step 5. Adding the REF10 made an immediate improvement in refinement which increased as I simplified my chain.


1. RedNet D16 using LiveClock as wclk/ M3USB#1 as re-clocker/ M3USB#2 as 2nd re-clocker


2. RedNet D16 using LiveClock as wclk/ REF10 as 10M reference to both M3USBs and to the LiveClock/ M3USB#1 as re-clocker/ M3USB#2 as 2nd re-clocker


3. RedNet D16 using M3USB#2 as wclk (LiveClock removed) / REF10 as 10M reference to both M3USBs / M3USB#1 as re-clocker/ M3USB#2 as 2nd re-clocker/


4. RedNet D16 now using M3USB#1 as wclk / REF10 as 10M reference to both M3USBs/ M3USB#1 as re-clocker/ M3USB#2 as 2nd re-clocker


5. RedNet D16 using M3USB#1 as wclk and with REF10 as 10M reference/ M3USB#1 as the only re-clocker/ (M3USB#2 removed)


Note: When you add an external 10M reference clock source to the Mutec 3+ USB it then allows you to use its wordclock output to clock another device like the D16.

*Warning: Purely subjective observations to follow which may not be appropriate for all viewing audiences.

To my ears and in my system and in my room the overall effect of adding the REF10 was to create a much more natural, refined and realistic sound. Nearly all aspects of music were improved for me even dynamics, soundstage localization, and depth. It is basically a more engaging overall presentation with subtle notes and decay being more apparent with a very realistic tonal balance. Musical threads by individual instruments are now very easily followed. Music is much more like a living, moving tapestry.

I mostly listen to acoustic music, preferring classical and Jazz, from small ensembles to the over the top “big print” Russian Romantics. I now listen more to the music and less to the equipment. This is the highest praise that I can think of.

Notes: Originally adding the second Mutec USB had added a questionable improvement. Adding the LiveClock to the D16 added a clear improvement but adding the LiveClock to the M3USB degraded the sound.

After adding the REF10 to the M3USB I found that the wclk output from the M3USB was now available and sounded much better than that of the LiveClock which I removed. Even by itself the RedNet D16 was an improvement over my previous USB chain which included a Regen run from LPS, the M3USB using its USB input, and AES out.

BTW. To my ears even the D16 by itself using AES out was better than the new GEN5 USB board in my Yggdrasil DAC.

It has been a long and winding road!
 
Oct 13, 2017 at 8:47 PM Post #3,217 of 3,694
I have been wanting to try out the Mutec REF10 reference clock in my RedNet AOIP system after hearing such good things about it, especially after having found the Mutec 3+ USB to be such an excellent product. I contacted the US distributor to arrange for a trial purchase with the idea that I might do a review and post my findings. The distributor was extremely cooperative and only asked that they be allowed to fact check before I posted. They have done that and had no comments to add. I asked for the trial because I would be purchasing a piece of interface equipment that cost more than my DAC and adding it to a front end that was already over the edge. If it did not work out I wanted to be able to return it!

The Short Version:

I purchased the REF10. It is a keeper.

The Long Version:

I was hoping to get an improvement in sound quality but not to the extent that occurred. To add icing on the cake it has also allowed me to drop two pieces of equipment from my chain which had become extreme even by audiophool standards.

My system before adding the REF10 consisted of:

1. “CAPS-like” server running Windows Server 2012R2 and Audiophile Optimizer. Powered by LPS.

2. Dante Virtual Soundcard (DVS) software which allows JRMC to send the audio stream out via an ethernet port instead of USB

3. A PCIe fiber optic ethernet card with fiber running to a Fiber Media Converter(FMC) and then via CAT7 cable into the Focusrite RedNet D16 AES ethernet to AES/SPDIF converter.

4. Antelope LiveClock as external wordclock to the D16

5. AES cable out from D16 to Mutec 3+ USB interface acting as a re-clocker.

6. AES cable out to a second Mutec 3+ USB interface acting as a second re-clocker.

7. AES out to Schiit Audio Yggdrasil DAC.


This is the order in which I tested the REF10 and coincidentally the order in which I preferred the sound. The biggest jumps were step 2 and step 5. Adding the REF10 made an immediate improvement in refinement which increased as I simplified my chain.


1. RedNet D16 using LiveClock as wclk/ M3USB#1 as re-clocker/ M3USB#2 as 2nd re-clocker


2. RedNet D16 using LiveClock as wclk/ REF10 as 10M reference to both M3USBs and to the LiveClock/ M3USB#1 as re-clocker/ M3USB#2 as 2nd re-clocker


3. RedNet D16 using M3USB#2 as wclk (LiveClock removed) / REF10 as 10M reference to both M3USBs / M3USB#1 as re-clocker/ M3USB#2 as 2nd re-clocker/


4. RedNet D16 now using M3USB#1 as wclk / REF10 as 10M reference to both M3USBs/ M3USB#1 as re-clocker/ M3USB#2 as 2nd re-clocker


5. RedNet D16 using M3USB#1 as wclk and with REF10 as 10M reference/ M3USB#1 as the only re-clocker/ (M3USB#2 removed)


Note: When you add an external 10M reference clock source to the Mutec 3+ USB it then allows you to use its wordclock output to clock another device like the D16.

*Warning: Purely subjective observations to follow which may not be appropriate for all viewing audiences.

To my ears and in my system and in my room the overall effect of adding the REF10 was to create a much more natural, refined and realistic sound. Nearly all aspects of music were improved for me even dynamics, soundstage localization, and depth. It is basically a more engaging overall presentation with subtle notes and decay being more apparent with a very realistic tonal balance. Musical threads by individual instruments are now very easily followed. Music is much more like a living, moving tapestry.

I mostly listen to acoustic music, preferring classical and Jazz, from small ensembles to the over the top “big print” Russian Romantics. I now listen more to the music and less to the equipment. This is the highest praise that I can think of.

Notes: Originally adding the second Mutec USB had added a questionable improvement. Adding the LiveClock to the D16 added a clear improvement but adding the LiveClock to the M3USB degraded the sound.

After adding the REF10 to the M3USB I found that the wclk output from the M3USB was now available and sounded much better than that of the LiveClock which I removed. Even by itself the RedNet D16 was an improvement over my previous USB chain which included a Regen run from LPS, the M3USB using its USB input, and AES out.

BTW. To my ears even the D16 by itself using AES out was better than the new GEN5 USB board in my Yggdrasil DAC.

It has been a long and winding road!

One of the things I like most about the D16 (other than it's not USB), is how well it responds to better sources upstream. Sounds like the Mutec combo is a real winner.

It's too bad you went the fiber optic route on your PC, since you wouldn't have any interest in the JCAT Net Femto Card. I've dropped Fiber Optic since I couldn't hear any differences with or w/o FMC's, so it just simplifies the chain (I'm using Belden 10GX Ethernet cable). After installing the JCAT card, I couldn't listen to my system without it - probably comparable about how you feel with the changes the Mutec combo is doing with your system (not comparing JCAT to Mutec, but rather the significant changes we're both hearing).
 
Oct 13, 2017 at 10:06 PM Post #3,218 of 3,694
Thanks. And I thought that I had drained my wallet to the bottom :)

I think that I need to get used to the new sound. Perhaps I will consider it later.

I like the idea that there is no ground connection using fiber to the D16 but to be honest I am not sure that I heard a significant difference. I was going mostly on theory.

I had a Paul Pang USB card for a while when I was still using USB. I never heard much of a difference with that either. The big changes came with the D16, then the M3USB, and then adding the REF10. No guess work there!
 
Oct 16, 2017 at 3:27 AM Post #3,219 of 3,694
Thanks mourip for the REF10 review

After installing the JCAT card, I couldn't listen to my system without it

l've been using an Intel X540-T2 PCIe card in preference to either of two Z270-WS onboard LANs. The price of the X540-T2 is highly variable. I have never seen an audiophile review of a RedNet PCIe card which presumably delivers very low latency. I can see the JCAT NET card FEMTO permits a LPS direct. I'd love to see a reliable audiophile shoot-out between these three PCIe card alternatives.
 
Oct 16, 2017 at 8:10 AM Post #3,220 of 3,694
Thanks mourip for the REF10 review



l've been using an Intel X540-T2 PCIe card in preference to either of two Z270-WS onboard LANs. The price of the X540-T2 is highly variable. I have never seen an audiophile review of a RedNet PCIe card which presumably delivers very low latency. I can see the JCAT NET card FEMTO permits a LPS direct. I'd love to see a reliable audiophile shoot-out between these three PCIe card alternatives.

Yes, primary reasons for going the JCAT Net route was the external LPS option, and Femto clock. I wasn't looking to improve the latency. The only thing being powered on my motherboard is the CPU and memory, everything else is external with LPS.

I highly doubt the Rednet PCIe card can compete with JCAT Net card based sound quality. There might be some improvement with latency, but really who cares, as long as there are no dropouts. Obviously, the Rednet PCIe card was not designed for running a single Rednet box to a single PC as we're doing.

You say your preference is the X540, but you don't describe anything about the differences. If the changes aren't substantially obvious, than you might want to consider the JCAT Net card to improve in this area instead of latency (if that's the only difference).

As you can see, you don't need a lot of processing power to keep your latency under 1ms. After all, Ethernet usage is only 18.7 Mbps. Utilization is higher than my 6700K processer, but it's only running 6W TDP (externally powered LPS) vs 91W TDP (internal 650W power supply).
Task Manager.PNG


Not bad latency considering the lower powered motherboard I'm using. To be honest, the display is a bit skewed because it's normally 1.1ms for both Peak and Average.
Latency.PNG
 
Oct 16, 2017 at 8:30 AM Post #3,221 of 3,694
Yes, primary reasons for going the JCAT Net route was the external LPS option, and Femto clock. I wasn't looking to improve the latency. The only thing being powered on my motherboard is the CPU and memory, everything else is external with LPS.

I highly doubt the Rednet PCIe card can compete with JCAT Net card based sound quality. There might be some improvement with latency, but really who cares, as long as there are no dropouts. Obviously, the Rednet PCIe card was not designed for running a single Rednet box to a single PC as we're doing.

You say your preference is the X540, but you don't describe anything about the differences. If the changes aren't substantially obvious, than you might want to consider the JCAT Net card to improve in this area instead of latency (if that's the only difference).

As you can see, you don't need a lot of processing power to keep your latency under 1ms. After all, Ethernet usage is only 18.7 Mbps. Utilization is higher than my 6700K processer, but it's only running 6W TDP (externally powered LPS) vs 91W TDP (internal 650W power supply).


Not bad latency considering the lower powered motherboard I'm using. To be honest, the display is a bit skewed because it's normally 1.1ms for both Peak and Average.

I don't have any performance issues at all. I get good <1ms latency = 790-810μs whether onboard LAN or PCIe. Latency is neither a problem nor a top issue - the top issue of course being SQ. I'd say the X540-T2 is more stable and there is a slight reduction in unwanted "edge". It's not a great SQ margin over the onboard LANs. Perhaps it is true that the RedNet card may not present a SQ advantage - but who knows. We audiophiles may never turn over that stone (not even one of us to my knowledge?) as it costs somewhere 500-800 GBP. Similarly, one can't evaluate the JCAT without splashing out over 400 Euros say 400 GBP. I got the X540-T2 for barely over 100 GBP and could sell it for more instantly. It's a matter of bang for buck. To evaluate the JCAT I would have to invest 400 GBP on faith that the bang would be large enough - in comparison with the potential bang that cash spent elsewhere in the system would generate ...

(Incidentally do you run your CPU on low power or is it a 1.6 GHz - I run without SpeedStep in BIOS / at 100% / performance "High" in W10 at 3.4 GHz. Your other stats are good - are you using Fidelizer or AO or whatever. You have plenty RAM - I find playing from RAM using fb2k component an improvement.)
 
Last edited:
Oct 16, 2017 at 10:06 AM Post #3,222 of 3,694
I don't have any performance issues at all. I get good <1ms latency = 790-810μs whether onboard LAN or PCIe. Latency is neither a problem nor a top issue - the top issue of course being SQ. I'd say the X540-T2 is more stable and there is a slight reduction in unwanted "edge". It's not a great SQ margin over the onboard LANs. Perhaps it is true that the RedNet card may not present a SQ advantage - but who knows. We audiophiles may never turn over that stone (not even one of us to my knowledge?) as it costs somewhere 500-800 GBP. Similarly, one can't evaluate the JCAT without splashing out over 400 Euros say 400 GBP. I got the X540-T2 for barely over 100 GBP and could sell it for more instantly. It's a matter of bang for buck. To evaluate the JCAT I would have to invest 400 GBP on faith that the bang would be large enough - in comparison with the potential bang that cash spent elsewhere in the system would generate ...

(Incidentally do you run your CPU on low power or is it a 1.6 GHz - I run without SpeedStep in BIOS / at 100% / performance "High" in W10 at 3.4 GHz. Your other stats are good - are you using Fidelizer or AO or whatever. You have plenty RAM - I find playing from RAM using fb2k component an improvement.)
If the JCAT Net card was a waste of money, I would have said so (I spent more money on Fiber Optics, and that to me was a real waste of money. Same as my 6700k CPU system). However, IMO the changes the JCAT Net card brings is just as or even bigger than the D16 itself (in combination with modified LPS for D16, and low powered Celeron 1.6 CPU). I can't say how it would change anything with unmodified D16 and high powered CPU, but I would think there would still be a noticeable difference. There are no SMPS power supplies in my chain.

I have AO, and again, a total waste of money for AOIP. There was zero improvement in sound. Yes, less processes running which probably helps with a more stable system (less likely to have dropouts), but what I found with the low powered CPU is the "CPU Utilization" went up to 45-50%. There are other things about AO that I don't like as well, but this isn't the place for that discussion.

I get better results using these methods than AO, and it's all free - https://www.computeraudiophile.com/...ws-10-optimization-script-a-community-effort/

Edit: Sorry missed this in your post, the Intel Celeron can run up to 2.24GHz, but I have it set to always run at 1.6GHZ.
 
Last edited:
Oct 16, 2017 at 10:16 AM Post #3,223 of 3,694
If the JCAT Net card was a waste of money, I would have said so (I spent more money on Fiber Optics, and that to me was a real waste of money. Same as my 6700k CPU system). However, IMO the changes the JCAT Net card brings is just as or even bigger than the D16 itself (in combination with modified LPS for D16, and low powered Celeron 1.6 CPU). I can't say how it would change anything with unmodified D16 and high powered CPU, but I would think there would still be a noticeable difference. There are no SMPS power supplies in my chain.

I have AO, and again, a total waste of money for AOIP. There was zero improvement in sound. Yes, less processes running which probably helps with a more stable system (less likely to have dropouts), but what I found with the low powered CPU is the "CPU Utilization" went up to 45-50%. There are other things about AO that I don't like as well, but this isn't the place for that discussion.

I get better results using these methods than AO, and it's all free - https://www.computeraudiophile.com/...ws-10-optimization-script-a-community-effort/

Edit: Sorry missed this in your post, the Intel Celeron can run up to 2.24GHz, but I have it set to always run at 1.6GHZ.

I have seen the W10 community effort thread. Even tho' I roll W10 back a great deal I confess I am tempted with Fidelizer having tried the Free version (I've had a chat with Keetakawee about demotion of Dante processes, giving him the necessary info to fix that + he'll update for Merging).

I have an HDPLEX now and could run the JCAT card easily from that I guess - if I get one I shall be sure to let you know what I think!
 
Last edited:
Oct 16, 2017 at 2:06 PM Post #3,224 of 3,694
I have seen the W10 community effort thread. Even tho' I roll W10 back a great deal I confess I am tempted with Fidelizer having tried the Free version (I've had a chat with Keetakawee about demotion of Dante processes, giving him the necessary info to fix that + he'll update for Merging).

I have an HDPLEX now and could run the JCAT card easily from that I guess - if I get one I shall be sure to let you know what I think!

Just did a quick test with the Dante Background Processes. Looks like the first 2 on the list cannot be stopped (Dante stops working). The last 3 can be stopped, but you cannot use Dante Virtual Soundcard (DVS), Dante Controller, or Rednet Controller.

Dante Background Tasks.png
 
Oct 16, 2017 at 4:20 PM Post #3,225 of 3,694
If the JCAT Net card was a waste of money, I would have said so (I spent more money on Fiber Optics, and that to me was a real waste of money.

Glad that you like the JCAT in your system. I checked the price of the JCAT card and it is about $500 before shipping. What was your fiber optic setup?
 

Users who are viewing this thread

Back
Top