Taiko Audio SGM Extreme : the Crème de la Crème
- Thread starter CKKeung
- Start date
So if you stream music and you want bit-perfect only, there is no point(reason) to use JPlay classic or femto? Or JRivers or HQplay for that matter?
Another thing I don’t get is the computer’s clock. If you use USB asynchronous dac. Why does the clock matter in the computer?
Thanks ??
Another thing I don’t get is the computer’s clock. If you use USB asynchronous dac. Why does the clock matter in the computer?
Thanks ??
Back in February, there was quite a SQ gap between Roon doing the track serving and JPlay Femto Server doing the track serving so there was a SQ reason to use JPlay
HQ Player is unsurpassed for PCM to DSD upsampling and conversion, so if you have a DSD DAC HQ Player is definitely worth trying out
JRiver is a jack of all trades, audio and video, requires a bigger Windows OS and is consequently SQ handicapped
The frequency stability and the low frequency phase noise of the clock plays a significant role in shaping the RF emissions coming from the clock oscillator
These RF emissions get everywhere and have an audible effect on the sound quality of the playback.
So USB DAC’s with asynchronous receivers are still very sensitive to the RF emissions coming from clocks / oscillators even though the bitstream is error free and accurately timed
So your saying it is the quality of the computers(servers) clock that matter? Low RF causing? Not the accuracy’s?
The clocks in digital equipment are fixed frequency carrier wave radio transmitters with carrier frequencies in the megahertz to gigahertz RF band.
The carrier wave, when it's steady, is benign, but when it's not steady with low frequency (audio band) phase noise or rippling power supply current and voltage modulating the carrier wave, then you are getting audio band noise / disturbances transmitted by the carrier wave that becomes audible when it is detected by a transistor, tube or coil any where in the signal path or power supplies
With clocking, everything matters. Frequency stability, phase noise, steady power supply voltages, vibration isolation. The bit integrity and bit timing is all good, but the RF soup has audible side effects
How the above helps in seeing how non digital RF/EMF has a significant SQ influence
Hi! I really like the design of SGM Extreme Server and I think it is a gorgeous piece of art. To learn more about it I found some pictures of the inside structure of this unit, and there are a few things which make me feel a little curious.
I found that on the main part of the server, the motherboard, there is no USB/AES/network bridge or card of any kind. Instead, you chose to use the USB output on the original motherboard. Is there any specific reason to not design and use a separate output for the server? Also, I noticed that there is no clock installed in the server. I have heard that the clock on the motherboard can be a big problem when it comes to SQ, and some server manufacture chose to install a separate clock to fix it. I wonder if there are some other ways to solve this problem? My last problem is that is there any specific reason to choose a XEON motherboard instead of a Z390 or X299 motherboard (I suppose the SQ of XEON motherboard is superior, but not sure why)?
These questions are just from curiosity and I mean no offense at all. I will be grateful if you would like to answer my questions. Thank you!
Best regards,
Xiao
Last edited:
Hi Xiao,
We have been there, done that LoL
In the SGM and SGM Evo, we had a fancy OCXO, with a very low noise power supply with good results for 2016 - 2018. We had some theories about why the sound improved, but it turned out that our speculative theories were wrong. Now that we have a better understanding about what is going on, Emile has been able to implement a motherboard clock / timing solution that sonically beats what was implemented in the SGM and SGM Evo
Similarly what Emile has been able to achieve with dual CPU Xeon is way ahead of any thing he could achieve with Z 370 / 390.
We tried Xnn a long time ago, and it was seriously disappointing
Thanks for your detailed reply! I am sorry that I didn't check previous SGM servers. I think it is a great achievement to be able to solve the clocking problem on the motherboard!
Best regards,
Xiao Liu
Just got a pair of 1G StarTech SFPs yesterday. Two Clones Audion audio graded switches are used in the system and two SFPs are used to connect these 2 switches. And the performance of these 2 SFPs is almost exactly the same what have described. It is much much better than the one previously used.
Thanks for the sharing.
Hello.
Just wanted to say kudos to Emile @Taiko Audio on the SGM Extreme. I've just read through all 58 pages of this thread, and I've got to say that his approach resonates with my own far more modest efforts. I hope some day to experience this box in my system, but until then, I'll continue to lurk and learn.
Just wanted to say kudos to Emile @Taiko Audio on the SGM Extreme. I've just read through all 58 pages of this thread, and I've got to say that his approach resonates with my own far more modest efforts. I hope some day to experience this box in my system, but until then, I'll continue to lurk and learn.
Thank you and welcome to the What's Best Forum. If you are the same Austinpop as on the Audiophilestyle Forum I would characterize your efforts as beyond modest. One of the driving forces moving the technology forward would be more appropriate.
From my experience so far, the SFP modules are the single most influential components of a fiber network setup. Second is the server side SFP network card, third place goes to the switch / fmc.
The Startech 1G 1550nm SM 80km modules continue to give me the best midrange density and least high frequency glare. 10G SFP+ modules are so far all considerable brighter sounding. You can use these at 1G speeds and they still have this brighter signature. I have ordered several more (exotic) 10G SFP+ modules in an attempt to find a module with a satisfying tonal balance. There is a benefit to be had from running 10G network speeds, resolution is higher, and interestingly it reduces the delta between Qobuz and Tidal playback quality.
In Korea, W20SE vs SGM Extreme comparison test was finished.
10 Korean audiophiles all said SGM Extreme quality is absolutely higher.
W20SE is much more than W20, but there is big difference with two music servers.
By this movies, interest about W20SE was decreased critically
and SGM Extreme awareness is much more higher.
I think to make W30 will be very difficult.
These members final opinions are that Aurender's reputation will be lost more and more.
Actually this comparison is meaningless I think.
Enjoy it~!
https://cafe.naver.com/hifikorea/98121
10 Korean audiophiles all said SGM Extreme quality is absolutely higher.
W20SE is much more than W20, but there is big difference with two music servers.
By this movies, interest about W20SE was decreased critically
and SGM Extreme awareness is much more higher.
I think to make W30 will be very difficult.
These members final opinions are that Aurender's reputation will be lost more and more.
Actually this comparison is meaningless I think.
Enjoy it~!
https://cafe.naver.com/hifikorea/98121
 | Steve Williams Site Founder | Site Owner | Administrator |  | Ron Resnick Site Co-Owner | Administrator |  | Julian (The Fixer) Website Build | Marketing Managersing |