Taiko Audio SGM Extreme : the Crème de la Crème

[spiritofmusic]

Thanks Emile. I've spoken to a fair number of well heeled audiophiles who wouldn't blink at spending serious sums on a DAC but the bottom line is absolute modular upgradability. They must have the next best from the company witjout being left with an uber pricey doorstop. These very people are highly reluctant to invest in any server beyond Roon level because they do not see this modularity being offered in the same way.

OTOH, it's highly commendable that you're going further than most in yr scheme to trade in an SGM2015 for Evo for the simple price differential. Surely that should assure most of these reluctant and sceptical non adopters.

 

[CKKeung]

Thank you for asking. Indeed very similar.

The original SGM 2015 can be fully upgraded to the current EVO model.

Naturally we will offer the same support on the Extreme.

It is impossible to upgrade an EVO to Extreme, but we do accept SGM 2015 and SGM EVO trade in’s on the Extreme.

Software updates are provided for free. We even install them for you if desired.

Additional Interfaces, USB, AES/EBU, Ethernet are available on both models. Future interfaces can be easily retrofitted.

The main differences from this perspective between the 2 models are:

EVO:
1 CPU
100W passive cpu cooling system
150W linear powersupply
2 expansion slots

Extreme:
2 CPU’s
240W passive cpu cooling system
400W linear powersupply
7 expansion slots

This kind of upgrade policy is very rare amongst the hi-end level components.
And I heard that price-wise it's very reasonable.
My salute to Emile & Edward!

I have friends in Hong Kong who have upgraded their SGMS2015 to EVO happily.
And a friend who has an EVO is considering to trade-in it to Extreme now.

However I think overseas dealers will charge shipping cost/import tax/handling fee accordingly.

BTW please wait for Emile's internal photos of Extreme taken by professional photographers.
Once available, you will understand why I am saying that diyers and other brands have extreme difficulty to catch up with the Extreme!

 

[Audiophile Bill]

Hi All,

Excuse my ignorance as I may have missed this - what is the design aim for omitting the upsampling ability in the extreme and why does it achieve even higher levels of performance by the omission?

 

[EuroDriver]

Hi Bill,

For us the interesting development in DAC's is the emergence of discrete R2R PCM DAC's that achieve a low noise floor by running R2R converter boards in parallel. The Total DAC D1-12 mk 2 has no less than 6 converter boards per channel

When the total system noise floor is lower, you can hear much more detail, but you can also hear processing noise incurred by upsampling code running on the server

When the system noise floor is higher, then there are definite sonic gains to be heard from upsampling to higher PCM rates or DSD

 

[Esotar]

I compared EVO to MP 64-32 Diamond.

8 Korean audiophiles visited at this test event.

6 audiophiles : EVO wins

2 audiophiles : MP wins

I don't post the writing about SGM Extreme quality, because there is no rivals against Extreme.

I'll post Extreme specific contents after one month.

 

[spiritofmusic]

Esotar, how are you guys rating Denafrips Terminator? My audio buddy here had it for a short time, and despite the stellar online reviews that prompted his purchase, he found the promise of SQ was a mirage, inferior to his T&A Dac8 at the time, and his current Aqua Formula XHD.

 

[Esotar]

Esotar, how are you guys rating Denafrips Terminator? My audio buddy here had it for a short time, and despite the stellar online reviews that prompted his purchase, he found the promise of SQ was a mirage, inferior to his T&A Dac8 at the time, and his current Aqua Formula XHD.

Though I choose Terminator, I think this is fair comparison because of same situation.

If Formula xHD, the result will be same~!

I'm importing Terminator.

But Terminator I'm importing is full custom version.

I call Terminator DE. (PSU higher upgrade, Capacitors upgrade)

Only Korean audiophiles can order Terminator DE.

Other country audiophiles never can buy it.

As my thought, Terminator DE is better than Formula xHD.

And its price is 1/2 of Formula xHD price.

^^

 

[spiritofmusic]

Any reason why it's only available in Korea?

 

[Esotar]

Any reason why it's only available in Korea?

Terminator DE is made by only my ideas.

So I have exclusive publication right about it.

Except SGM series, I'm importing only customized products by their brands.

Never custom them in Korea.

 

[Audiophile Bill]

Hi Bill,

For us the interesting development in DAC's is the emergence of discrete R2R PCM DAC's that achieve a low noise floor by running R2R converter boards in parallel. The Total DAC D1-12 mk 2 has no less than 6 converter boards per channel

When the total system noise floor is lower, you can hear much more detail, but you can also hear processing noise incurred by upsampling code running on the server

When the system noise floor is higher, then there are definite sonic gains to be heard from upsampling to higher PCM rates or DSD

Hi Ed,

Thanks - sorry it has taken me so long to reply to this. A couple of points for my understanding:

> Does one need a dac that does no upsampling to derive the advantages of extreme? I ask because I thought the merit of the HQ per SGM EVO (among other things) was to get that work done off board and taking that processing demand away from the dac?

> If the extreme is all about system noise - what is the system signal to noise threshold before Extreme makes sense vs EVO? I presume that any system driven with valves is a no go for Extreme since the noise floor will never be better than 110dB (absolute tops)? To derive the benefit of the extremely low noise doesn’t one need to be really using amps with SNR of 130dB or better meaning we are mainly in class D territory or some Uber solid state amps?

Thanks a lot - sorry for all the questions.

 

[Audiophile Bill]

Hi Ed,

Thanks - sorry it has taken me so long to reply to this. A couple of points for my understanding:

> Does one need a dac that does no upsampling to derive the advantages of extreme? I ask because I thought the merit of the HQ per SGM EVO (among other things) was to get that work done off board and taking that processing demand away from the dac?

> If the extreme is all about system noise - what is the system signal to noise threshold before Extreme makes sense vs EVO? I presume that any system driven with valves is a no go for Extreme since the noise floor will never be better than 110dB (absolute tops)? To derive the benefit of the extremely low noise doesn’t one need to be really using amps with SNR of 130dB or better meaning we are mainly in class D territory or some Uber solid state amps?

Thanks a lot - sorry for all the questions.

Sorry Ed.

One last question - why is the power of the extreme much higher than the EVO when it doesn’t do the upsampling heavy lifting. I would have imagined that you would have wanted the lowest possible power for noise reduction?

 

[EuroDriver]

Hi Ed,

> Does one need a dac that does no upsampling to derive the advantages of extreme? I ask because I thought the merit of the HQ per SGM EVO (among other things) was to get that work done off board and taking that processing demand away from the dac?

As they say it all depends on the implementation ! There are some DAC's such as the MSB Select II that do a very good job of upsampling / internal processing and have a low noise floor. However with these DAC's you may find you need to pay a lot of attention to vibration control otherwise the sonic benefits will be masked.

The DAC's which Emile has been working with are the TotalDAC D1-12 with its spectacularly low noise floor and the Aqua Formula XHD. Both of these units do no upsampling

 

[Kingsrule]

The MSB doesn't upsample, it plays the native rate of the file

Dacs like the dCS Vivaldi/upsampler do upsample .....

 

[EuroDriver]

Hi Ed,

> If the extreme is all about system noise - what is the system signal to noise threshold before Extreme makes sense vs EVO? I presume that any system driven with valves is a no go for Extreme since the noise floor will never be better than 110dB (absolute tops)? To derive the benefit of the extremely low noise doesn’t one need to be really using amps with SNR of 130dB or better meaning we are mainly in class D territory or some Uber solid state amps?

In the demo room in Hengelo Emile used to have and was very happy with a pair of VTL mono blocs. Then we heard what the Audionet Stern Preamp and Heisenberg mono blocs could deliver. Besides ruler flat frequency response and extremely low distortion the SNR of these units is as low as we have seen and wow can you hear it. In our humble opinion the reviews by David Robinson and others about this amplification combo is spot on.

What the better than -150 db noise floor of the D1-12 and the - 123 db SNR of the Stern and Heisenberg fed by the Taiko Extreme delivering 16/44 music from PCI storage is simply stunning. B- grade recordings sound like Master Tapes. I have only heard Master Tapes 5 times in my life, but it leaves such a deep impression. I get this sonic wow with this again with this system.

I think it's important to point out that there is a lot of vibration control and vibration shaping going on, and that's what brings the image depth, breadth and the dimensional stability of the sound stage. After a discussion with a hearing aid technologist, we learnt what improves intelligibility and spatial perception and fits with the results we get from the materials and design of the vibration control in the SGM Extreme

 

[EuroDriver]

The MSB doesn't upsample, it plays the native rate of the file

Dacs like the dCS Vivaldi/upsampler do upsample .....

That's why I said "internal processing" The MSB Select II has a very powerful 80 bit processor that is not sleeping, but doing a great job

 

[microstrip]

Does any one know about a good article explaining how noise produced in computers reaches the DACs through asynchronous links such as USB or Ethernet packets?

In other words, can any one provide a detailed technical explanation, not voodoo words, why the SGM EVO sounds different from the Statement when feeding the same DAC?

 

[microstrip]

The MSB doesn't upsample, it plays the native rate of the file

Dacs like the dCS Vivaldi/upsampler do upsample .....

Do you have details about the MSB Sellect DAC that you can share with us?

 

[EuroDriver]

Sorry Ed.

One last question - why is the power of the extreme much higher than the EVO when it doesn’t do the upsampling heavy lifting. I would have imagined that you would have wanted the lowest possible power for noise reduction?

What's important and audible, is how quickly the OS and playback software executes its processes. We have known for a longtime that low latency improves the sound of a server. Low latency will more often than not shorten the execution time of code blocks, less waiting around and so delivers a SQ uplift

 

[EuroDriver]

Does any one know about a good article explaining how noise produced in computers reaches the DACs through asynchronous links such as USB or Ethernet packets?

In other words, can any one provide a detailed technical explanation, not voodoo words, why the SGM EVO sounds different from the Statement when feeding the same DAC?

The RF emissions from these two servers are different. Where are these emissions picked up ? Our best guess is anywhere and everywhere !
- out the power cord for sure


We have done a test where a server was hooked up to a DAC using an Adnaco fibre optic USB connection. You make changes on the server and you can hear it in the DAC output, so there are other pathways for the RF than just the async USB

 

[microstrip]

The RF emissions from these two servers are different. Where are this emissions picked up ? Our best guess is anywhere and everywhere !
- out the power cord for sure

We have done a test where a server was hooked up to a DAC using an Adnaco fibre optic USB connection. You make changes on the server and you can hear it in the DAC output, so there are other pathways for the RF than just the async USB

Thanks for chiming on this mysterious subject. So you are assuming that the difference are due to the RF (electromagnetic emissions ) of the server, not to the timing of packets or noise going through the data links? BTW, an optical link is not completely noise free, the noise can can modulate the signals being transmitted through it and be recovered at the receiver. It only limits the bandwidth of the noise.