I have two PCIe NVMe cards. My C: drive is the SSD card; the two NVMe cards are configured as the D: drive for music storage.
Steve Z
Taiko Audio SGM Extreme : the Crème de la Crème
- Thread starter CKKeung
- Start date
Steve Z
Not to interrupt all this tech talk and cruelty to woodchucks with music, but here is another (partial) batch of music I've been loving via the Taiko USB card-equipped Extreme:
Part 2 to follow.
Steve Z
Part 2 of music appreciation:
Hope there are some selections folks like.
Steve Z
Hope there are some selections folks like.
Steve Z
I can't do that comparison but spent last night doing the more obvious one- an SoTM switch vs no switch.
The switch I use is an upgraded sNH-10G with the upgraded sCLK clock/7N silver/EVOX caps.
The comparison involved an assessment of both the switch and a companion LPS. The stock SoTM LPS is the sPS 500.
To cut to the chase, The Extreme/USB board with the switch as per above was a miniscule benefit that I'm not sure I could pick out reliably in a blinded comparison. But the more interesting comparator was using the switch with a modest, custom LPS that I've found quite a bit better than the stock SoTM LPS. Its a very inexpensive Chinese LPS that appears to be OEM'd by one manufacturer but might come with several names on the box. Mine says "BreezHiFi" but there are others the key is to order these with an upgraded Talema transformer. This will set you back all of $118 (the stock version is about $68!). I use a few of these LPS units. They're cheap but appear to be nicely made and have been quite reliable. (I discovered a while ago that they provide a nice benefit when they replaced any wall wart power supply plugged in my system. I was surprised just how pernicious those little things really are even if the are only used to control peripheral devices i.e. lighting. infrared repeaters, etc. Who knew?).
Bottom line- the sNH-10G plus "Talema" BreezHiFi LPS stays. It's a small but welcome uptick for me vs. no switch at all. YMMV. One take away from this exercise is appreciating that the conclusions others may have for their switch preferences are in fact based on their switches AND the LPS that comes with it.
Here's one of several similar ads on Ebay that all appear mostly identical:
Upgraded HIFI 50W Talema Ultra-low Noise Audio Linear Power Supply DC5V-24V PSU | eBay
(Use all aluminum chassis. we also can provide 5.5 2.1mm or other size) NOTE: The AC input cable is not included by default.
www.ebay.com
Last edited:
Steve williams
Site Founder, Site Owner, Administrator
Emile, only because you asked, I revisited this.
In short, the USB card has done nothing to change the relative SQ of switches and switch chains, since my last analysis of the subject here:
A novel way to massively improve the SQ of computer audio streaming
audiophilestyle.com
Here is the ranking of SQ, from best to worst:
1. No network. Last switch to Extreme powered off
2. Full monty chain of 3 switches
3. Modded Buffalo
4. SOtM and eR (tie)
5. opticalModule
6. direct attach from router (copper. 2-5 above are fiber to Extreme).
Now we need to re - run this evaluation with a XMOS 2 USB DAC ;-)Emile, only because you asked, I revisited this.
In short, the USB card has done nothing to change the relative SQ of switches and switch chains, since my last analysis of the subject here:
A novel way to massively improve the SQ of computer audio streaming
Here is the ranking of SQ, from best to worst:
1. No network. Last switch to Extreme powered off
2. Full monty chain of 3 switches
3. Modded Buffalo
4. SOtM and eR (tie)
5. opticalModule
6. direct attach from router (copper. 2-5 above are fiber to Extreme).
austinpop,Emile, only because you asked, I revisited this.
In short, the USB card has done nothing to change the relative SQ of switches and switch chains, since my last analysis of the subject here:
A novel way to massively improve the SQ of computer audio streaming
Here is the ranking of SQ, from best to worst:
1. No network. Last switch to Extreme powered off
2. Full monty chain of 3 switches
3. Modded Buffalo
4. SOtM and eR (tie)
5. opticalModule
6. direct attach from router (copper. 2-5 above are fiber to Extreme).
I'm confused. When you say " the USB card has done nothing to change the relative SQ of switches and switch chains,".....
are you saying that since receiving your USB board (~2 weeks ago?) you've done all these tests and your rankings are unchanged?
Or are your rankings conjecture only and have not been validated with the new USB board in the Extreme?
Marty
Last edited:
Emile asked me to revisit these tests based on his own experience that the new USB card changed the impact of his network topology. So this set of tests was done with the new USB card deployed. Yes, a fair bit of work, but I am always happy to help when Emile asks for it.
When I said " the USB card has done nothing to change the relative SQ of switches and switch chains," I meant relative to the last time I did these tests in September (my AS post), when I was using the builtin USB port.
So this must be system specific, as on my chain, the relative SQ of different switch topologies has not changed with the installation of the new USB card.
Thank you @austinpop
Still very limited feedback:
Impact of network configuration unchanged:
1x Chord Dave - Amanero - Chord driver - HQPlayer - Fiber
Impact of network configuration reduced:
2x Lampizator Pacific - XMOS 2 - Taiko driver - TAS - copper
1x Lampizator Pacific - XMOS 2 - Taiko driver - Roon - copper
1x Pilium Elektra - XMOS 2 - Taiko driver - TAS - copper
1x MSB Select - XMOS 2 - Taiko driver - TAS - copper
If anybody else is willing to take the effort to add their experience, it would be much appreciated.
Still very limited feedback:
Impact of network configuration unchanged:
1x Chord Dave - Amanero - Chord driver - HQPlayer - Fiber
Impact of network configuration reduced:
2x Lampizator Pacific - XMOS 2 - Taiko driver - TAS - copper
1x Lampizator Pacific - XMOS 2 - Taiko driver - Roon - copper
1x Pilium Elektra - XMOS 2 - Taiko driver - TAS - copper
1x MSB Select - XMOS 2 - Taiko driver - TAS - copper
If anybody else is willing to take the effort to add their experience, it would be much appreciated.
Last edited:
Thank you, austinpop for your yeoman's work in testing all these network permutations.
My initial takeaway is that simply turning off the last switch and so mostly ( or completely?) eliminating the influence of the network is better (in your Chord DAC system).
That is, the full monty of modded switches and expensive power supplies is made superfluous by "simply" turning off the last switch.
As I understand, at present, the only way play network free files through Extreme/TAS is from local files, not streamed? Do I have that right, Taiko team?
My initial takeaway is that simply turning off the last switch and so mostly ( or completely?) eliminating the influence of the network is better (in your Chord DAC system).
That is, the full monty of modded switches and expensive power supplies is made superfluous by "simply" turning off the last switch.
As I understand, at present, the only way play network free files through Extreme/TAS is from local files, not streamed? Do I have that right, Taiko team?
hopefully i'm not being too pedantic, but i'm not sure how to interpret the above.
it seems to me that austinpop is saying the relative ordering of the networking options has not changed in his system. that does not necessarily contradict the statement that the impact of network configuration is reduced.
@austinpop: do i have this correct? in other words, is the benefit of your networking of a similar magnitude as with the old usb connection? has the new usb card reduced the delta between the various network options? lastly, can you place your old reference (full monty with old usb) in comparison with any of the new usb card options?
For the group that falls under “Impact...reduced” does that mean that people had switches in place prior to installing the USB card and now they feel that removing the switches does not result in any or very little degradation of sound? Also, are some of these situations for people who don’t post here?
I will be able to add a data point for the “Lampizator Pacific - XMOS - taiko driver - TAS - copper” by the end of the weekend.
austinpop, Thank You for taking your time and sharing all these data. I've noticed above that you mixed SFP: Planet on the Buffalo and Finisar on the Sotm, Is that correct? Which SFP are you using on the Extreme?
For me, the best results are Finisar on the Extreme and Planet on the OM.
Finally, I could installed the USB Card two days ago, Emile assisted me recreating a new pool due to the removal of Intel drives in slot 5. Honestly, I did not expect my system could sound any better when I had TAS installed, but yes, the New USB Card add layers of improvement I did not hear before, a much more relaxed music with better dynamics. much noticeable on the high band.
During the setup, I could play some fave tracks with and without the switches connected and did not noticed any sound degradation. I will revisited after 200 of burn in.
During the setup, I could play some fave tracks with and without the switches connected and did not noticed any sound degradation. I will revisited after 200 of burn in.
Please don't read too much into my conclusions!
As I've tried to make clear several times on this thread, I'm done with network tweaks. If you've followed the "Novel" thread over on AS, I've spent years chasing network optimizations with the goal of achieving SQ parity of streaming music with local content. Switches, clocks, PSUs — you name it, I've tried them all. Every time I've gotten close, someone moved my cheese.
I've decided to abandon this Sisyphean endeavor, and just do the simplest thing possible: turn off the network. It gives me the best SQ every time!As for streaming, I use it now only for discovery, and for that Roon/HQP is just fine. Anything I like enough to enage me, I buy. The stack of switches I have now is just a legacy.
It is against this backdrop that Emile introduced the idea that his changes on the USB side of the chain were inoculating him from the impact of the network side of the chain. Moreover, he thought the relative benefit of specific switches had shifted. That was intriguing enough for me to test on my system, and this is what led to this round of switch tests.
Bottom line: in my system, I don't experience either of these effects. The improvements due to the USB card, considerable as they are, neither increase nor reduce the impact of changes on the network side. And the relative SQ of the various switches is the same as before.
Emile thinks this may be related to the USB chipset in the DAVE (Amanero) vs. XMOS. It's an intriguing hypothesis, and I look forward to any further insights he uncovers.
The other wrinkle here is whether the Extreme is isolated on a subnet. I have not yet revisited this with the new USB card, but will try this weekend.
Previously, I heard a huge impact with the Extreme on its own subnet (not exposed to traffic from the rest of the network). With TAS downloading streaming content prior to playback, this had a huge impact on SQ for me. There was VERY slight difference with network on or off, but it could have been wishful thinking.
I too have an opticalModule on a remote switchable outlet. This weekend I'll give a listen to local music with network, and local music with oM powered down.
That being said, @austinpop setup is a better indication of how sensitive the Extreme + USB card is to network traffic. I suspect the 5 examples Emile cites above probably aren't isolating subnets. My subnet strategy basically sidesteps the network question by isolating my Extreme from all network traffic that is not routed to it (it is the only device on the subnet), while in Rajiv's set up, his system is be exposed to all the network traffic on his home network (even though it is passed through a stack of switches).
Previously, I heard a huge impact with the Extreme on its own subnet (not exposed to traffic from the rest of the network). With TAS downloading streaming content prior to playback, this had a huge impact on SQ for me. There was VERY slight difference with network on or off, but it could have been wishful thinking.
I too have an opticalModule on a remote switchable outlet. This weekend I'll give a listen to local music with network, and local music with oM powered down.
That being said, @austinpop setup is a better indication of how sensitive the Extreme + USB card is to network traffic. I suspect the 5 examples Emile cites above probably aren't isolating subnets. My subnet strategy basically sidesteps the network question by isolating my Extreme from all network traffic that is not routed to it (it is the only device on the subnet), while in Rajiv's set up, his system is be exposed to all the network traffic on his home network (even though it is passed through a stack of switches).
Apologies if I confused you there! I have a pair of each transceiver deployed. So when I say:
Upstream > Supra Cat 8 > (A) eR (A) > Sablon Eth > modded Buffalo > Planet SFP > sNH-10g > Finisar SFP > Extreme
That means:
- I have a pair of Planet Tech MGB-TLX transceivers, one in the Buffalo's SFP slot, the other in the SOtM's SFP slot, and
- a pair of Finisar FTLF1324 transceivers, one in the other SOtM SFP slot, and the other in the Extreme fiber card's SFP slot.
Hope that clarifies matters?
My USB card is not fully burned in . 4 days only and it is still improving.
It is a true game changer for my system. I love the SQ improvement.
Thanks to Emile, Edward and all from Taiko Team!
In regards to Buffalows 2016 I tested it yesterday and I must say they still do the job !
yes my 2 Buffalows are fully moded with Pink Faun ultra clocks , 2 are Nenon moded with Mundorf siver gold DC wire.
Aall is powered by SJ DC4 rails (Nenon mode with Mundorf siver gold internal wireing).
I also have Router that I moded with Pink Faun ultra clock and modem that has PF ulta clock Installed.
Every single thing is still making noticable positive difference .
As Austinpop mentioned full stack of switches ( 4 Buffalo 2016 in my case) is the best .
The only better thing is to disconnect the last switch after starting the music.
But be aware your Pink Faiun ultra clock should never be turned off !
It takes upto 2 days to have full performance once disconnected.
So if you test it , please disconnect the Buffalo 2016 swich power off and keep its clock powered at all times. This is just to make sure the others are not using wrong methodology when testing.
As a side not not only the streaming is improved . The local files playback is also much better with stack of switches and good power supplys.
So as few of us already mentioned including Emile it will alwasy be noticable change once you add
someting to the chain.
As I dont trust myself when it comes to „first impressions” and I love testing ,
I am planning to redo this switch tests in a week when usb card is fully broken.
I predict it will still be important to keep them. I would prefer to say they are not important
and sell all Buffalows but it is impossible.
I thik this will be possible when Emile decides to create the Taiko Network card that will be as good as the USB is.
