SGM Taiko Extreme

[isquirrel]

33 minutes ago, Custodian said:

Interesting piece of equipment. How does the Taiko synchronise the time clock with the rest of the digital system? For me, being able to lock the time signal to my external BVA master clock would be an essential piece of the spec. The Taiko site seems a bit sparse with information.

 

If you use USB your DAC receives the data asynchronously, that is to say that your DAC re-clocks the data, which voids any benefits of using a Word Clock sync, I have experimented with different outputs from the W20 and only the AES 110ohm & the SPDIF outputs are affected by the word clock sync option. As I want to be able to take advantage of high res files it means using the USB output anyway. So the word clock sync is of no value. MSB offer a Pro USB module http://www.msbtechnology.com/accessories/prousb/ which effectively reduces any noise out of the USB output by converting the data to fibre. 

 

SGM who make the Taiko Extreme previously built the SGM 2015 server, which used Roon Server to output data to HQPlayer and upsample everything to DSD 256 or DSD 512 whichever format your DAC could accept. 

 

However honestly, ultimately the PCM bit depth, even if it’s just 16 bits, gives a better dynamic range then DSD. I have been quite a fan of DSD for a long time, the first SGM servers where designed to upsample everything to high DSD rates. This is a very valid approach for “medium range” systems, but people buying the Extreme typically own top flight systems, with upper echelon DACs, and those do just perform better with PCM if the server noise floor is vanishing low. DSD256 or DSD 512 sounds a bit less precise / smoother, which is just what the format sounds like.

 

The new Taiko Extreme does not upsample and does not use any internal clocks as typical high end DAC's re-clock the data anyway. It does play bit perfect as I have verified using MSB test files. The use of internal clocks cause noise which is the antithesis of the way the Taiko Extreme was designed. Rather than disabling services and removing Windows components after installation the OEM version of Windows installed on the Taiko Extreme is massively cut down so that anything unnecessary is not installed in the first place. The playback chain is Roon routed through JPlay. 

 

I was initially skeptical of the Taiko Extreme, I was about to upgrade to the W20SE and I have always been very happy with my W20 and raised the same questions you have! I held off for about 3 months whilst I read every page on the WBF site about the Taiko Extreme.

https://www.whatsbestforum.com/threads/taiko-audio-sgm-extreme-the-crème-de-la-crème.27433/ There is no doubt that it is beautifully built however it is considerably more expensive than the W20 or the W20SE. Here in Australia the W20SE retails for $34,000, the Taiko Extreme for $49,000 before options.

 

I started to hear 1st hand reports from friends who had listened and done the comparison for themselves in Singapore, Hong Kong and Korea and it seemed to me that I had to satisfy myself. I started a dialog with the designers - Emile Bok and his partner and asked pretty much every question I could think of. Frankly I didn't want it to sound better than the W20 or W20SE. Eventually a few weeks ago I was loaned one by the new Australian Distributor - Nirvana Sound. The Extreme weighs 42KG and is even better built in the flesh. However it was the sound that drew my undivided attention, this was not just an upgrade over the Aurenders it represented a Sea change in digital sound. Suddenly I was listening to what I thought were very familiar tracks and they sounded very close to 1/2 inch tape masters and certainly what I have experienced from top flight Analog Turntables, in some cases superior to the best of Vinyl, the variances I was hearing were mostly due to the original recording quality.

 

In summary - I have never heard digital sound anything remotely approaching the sound I was hearing from the Taiko Extreme. 

 

I take delivery of mine in a few weeks time. It would be a good idea to start a new thread on these forums where we can hopefully invite the designers to respond to our questions.

 

 

    

[Custodian]

Having compared various usb cables with top BNC cables from Furutech and Shunyata, I must declare it no contest. Asynchronous usb might be a clock solution but at a high price in sq.

 

for me, any alternative streamer must have facility to synch with my Oscilloquartz bva 8607 masterclock. The improvement of ultra accurate clocking is huge.

[dminches]

9 hours ago, isquirrel said:

 

The new Taiko Extreme does not upsample and does not use any internal clocks as typical high end DAC's re-clock the data anyway. It does play bit perfect as I have verified using MSB test files. The use of internal clocks cause noise which is the antithesis of the way the Taiko Extreme was designed. Rather than disabling services and removing Windows components after installation the OEM version of Windows installed on the Taiko Extreme is massively cut down so that anything unnecessary is not installed in the first place. The playback chain is Roon routed through JPlay. 

 

I was initially skeptical of the Taiko Extreme, I was about to upgrade to the W20SE and I have always been very happy with my W20 and raised the same questions you have! I held off for about 3 months whilst I read every page on the WBF site about the Taiko Extreme.

https://www.whatsbestforum.com/threads/taiko-audio-sgm-extreme-the-crème-de-la-crème.27433/ There is no doubt that it is beautifully built however it is considerably more expensive than the W20 or the W20SE. Here in Australia the W20SE retails for $34,000, the Taiko Extreme for $49,000 before options.

 

I started to hear 1st hand reports from friends who had listened and done the comparison for themselves in Singapore, Hong Kong and Korea and it seemed to me that I had to satisfy myself. I started a dialog with the designers - Emile Bok and his partner and asked pretty much every question I could think of. Frankly I didn't want it to sound better than the W20 or W20SE. Eventually a few weeks ago I was loaned one by the new Australian Distributor - Nirvana Sound. The Extreme weighs 42KG and is even better built in the flesh. However it was the sound that drew my undivided attention, this was not just an upgrade over the Aurenders it represented a Sea change in digital sound. Suddenly I was listening to what I thought were very familiar tracks and they sounded very close to 1/2 inch tape masters and certainly what I have experienced from top flight Analog Turntables, in some cases superior to the best of Vinyl, the variances I was hearing were mostly due to the original recording quality.

 

In summary - I have never heard digital sound anything remotely approaching the sound I was hearing from the Taiko Extreme. 

 

I take delivery of mine in a few weeks time. It would be a good idea to start a new thread on these forums where we can hopefully invite the designers to respond to our questions.

 

 

    

 

Thanks for the information.

 

Based on your conversations and listening do you have a sense for what it is about the Extreme that puts it at or near the top of the server best of category?

 

 

[Nenon]

11 hours ago, isquirrel said:

The new Taiko Extreme does not upsample and does not use any internal clocks

 

What do you mean by "does not use any internal clocks". Surely, you need an internal clock to generate the USB output. 

I remember reading comments from Emile that they have found a solution that works better than OCXO clocks, but he is very secretive about it.

[bmichels]

I had last year the chance to have a home demo of the SGM 2015 server, which indeed use HQPlayer to upsample everything to high DSD rates.  I Was impressed by the sound, but since I am not fan of upsampling I did not purchased it and got a W20 who does play bit perfect.  

 

Now I am very interested by isquirrel's writing "The new Taiko Extreme does not upsample".  This is a major strategic change to the SGM !   Therefore I am again interested in this new server... expecially after reading isquirrel's listening impressions.

 

question: is HQplayer not used anymore ?  HQplayer was a major foundation of the SGM package... and also a source of complexity of set-up/installation to optimize the SGM because of the myriad of choices/parameters/filters available through  HQplayer. 

[marslo]

10 hours ago, bmichels said:

I had last year the chance to have a home demo of the SGM 2015 server, which indeed use HQPlayer to upsample everything to high DSD rates.  I Was impressed by the sound, but since I am not fan of upsampling I did not purchased it and got a W20 who does play bit perfect.  

 

Now I am very interested by isquirrel's writing "The new Taiko Extreme does not upsample".  This is a major strategic change to the SGM !   Therefore I am again interested in this new server... expecially after reading isquirrel's listening impressions.

 

question: is HQplayer not used anymore ?  HQplayer was a major foundation of the SGM package... and also a source of complexity of set-up/installation to optimize the SGM because of the myriad of choices/parameters/filters available through  HQplayer. 

I have had the same impression after the audition of  SGM 2015 with HQ Player and T+A DAC 8 DSD at Geoffrey's Sound Gallery a few years ago .

I am also in favor of native files be it dsd or SM pcm so I keep my W20 for the sound and functionality.

I had also the home demo by Lukasz Domanski of LDMS which was very promissing, both in usb and RJ 45 connection.

I like very much the idea implemented in Taiko Extreme to transmit to the DAC the files in their native form to reduce the noise due the heavy transcoding and upsampling by HQ Player. The idea to make the transport " as simple as possible but not simpler " is also behind Jplay and W20 solutions , I started my computer playback journey with Jplay which excelled in  hibernate mode .

It would be very interessting to compare LDMS and Taiko Extreme as the price level of Taiko is rather prohibitive.

 

[EuroDriver]

The Taiko Extreme does indeed bring digital streaming and saved digital files to SQ performance levels that Emile and I never imagined could be possible even 18 months ago.

 

I have had the experience of listen to an decent quality 16/44 track played back on the Extreme and it sounded like a master tape.  How and when is level of digital SQ playback possible ?

 

The major factors are

 - RF Emissions shaping

 - Vibration control

 - reduction of music correlated distortion / noise

 

These advances have been made possible by developing and engineering the Extreme on a system which has the Total DAC D12 Mk2 with its incredible low noise floor, and the Audionet Stern and Heisenberg Pre and Monoblocs.  This ultra low noise combination has allowed Emile to clearly detect small sonic differences which on other systems could hardly be detectable and then one would not be sure which was sounding better.  When you sum up 30 to 40 of these small differences then the SQ Delta becomes a big jump.

 

Some of the variations are actually counterintuitive, but repeated testing consistently delivered positive results in the Extreme Server, low noise electronics, high transparency system context.  An example of where shaped random noise helps dramatically to improve SQ is adding shaped dither to a digital mix of 2 digital PCM streams.  You can find dither demo's on YouTube

 

With vibration, the surprise takeaway is that digital electronics are just as vibration sensitive as analog electronics, so attention and engineering applied to vibration control is essential to achieve "extreme performance"

 

The industrial grade RAM is extremely expensive and has to be custom ordered but it has very audibly better performance which we believe is due to lower vibration sensitivity than standard RAM.  Before you rush out to order industrial RAM, the optimization of RAM speed, refresh rates is comparable to tuning a concert grand, only possible by ear, but when you hit the sweet spot, the music sings.

 

I need to stop here, before i give away too much of Emile's IP.  Emile has left pretty much no stone un-turned in chassis design, server board data channel architecture, OS stripping /optimization, CPU core process allocation, latency optimization and vibration management

 

The other good news as Emile has delivered many Extremes World wide which then have been hooked up to many different DAC's and systems, the elevation of the customer's system sonics has been consistent in character and difference, a lot of happy customers

[Nenon]

6 minutes ago, EuroDriver said:

I have had the experience of listen to an decent quality 16/44 track played back on the Extreme and it sounded like a master tape. 

 

Well, it was a big pleasure to read the report of your visit to Mike Lavigne's listening room. But after reading Mike's comments, I was not left with the impression that the Extreme has reached the level of a master tape.

 

https://www.whatsbestforum.com/threads/taiko-audio-sgm-extreme-the-crème-de-la-crème.27433/post-603088

Quote

it was interesting seeing the reactions to my 2 turntables; they caused some quite strong preferences. and the grins and muttering after hearing the tape were priceless.

"how can we ever get digital to sound like 1/2" tape?" was said by someone.

how indeed?

 

Having said that, I have no doubts to believe that the Extreme is the best digital source on the market at the moment. The level of attention to details is OUTSTANDING! 

[EuroDriver]

On Mike's system, his Studer Master tape running 1/2 inch tapes at 15 ips was the clear sonic champ.  The sound at Mike's is glorious and gorgeous.

 

In Hengelo we have great electronics low noise electronics, but mikey mouse speakers in comparison to Mike's MM7's.  In Hengelo, Redbook sounds like a Master Tape in it's information density 

[justubes]

On 11/3/2019 at 9:49 PM, Custodian said:

Having compared various usb cables with top BNC cables from Furutech and Shunyata, I must declare it no contest. Asynchronous usb might be a clock solution but at a high price in sq.

Isquirell,

 

How does USB connection in your comparisons with AES you were using previously with your Select and the clock connection? Any particular usb cable you demoed with?

 

By any chance did you compare Tidal/Qobuz streaming quality?

 

It would be expected that the especially high resolution music played from internal pcie drives would be superior.

 

Thanks

 

 

[Custodian]

I tried a number of usb cables with several twin AES and BNC cables. For all cables, usb was worst. 
 

twin reference Siltech Aes was, in my system, slower and less detailed than my best BNC reference cables from Shunyata and Furutech. Comparison using both my current Aurender w20 and my previous Aurender s10 into DCS (both Vivaldi and Scarlatti)

 

between Qobus and Tidal, I prefer sound quality from Qobus, particularly with their HiRez files compared to somewhat coloured Tidal.

 

hope that helps, but remember as always, your own mileage may differ!

[RickyV]

On 11/3/2019 at 11:16 PM, Nenon said:

I remember reading comments from Emile that they have found a solution that works better than OCXO clocks, but he is very secretive about it.

I believe I read on the whatsbestforum in one of the posts from @austinpop that he had an idea about what that might be. 
Distance to the chip is critical as well as temperature stability. My guess is the supply voltage or a specific voltage to the chip. For instance 

This is the network chip on the nuc7i7dn.

Just speculating here.

[austinpop]

2 hours ago, RickyV said:

I believe I read on the whatsbestforum in one of the posts from @austinpop that he had an idea about what that might be. 

 

For the record, I have no clue what Emile has found to render the need for OCXO clocks unnecessary. When I made that comment, my best guess was a comprehensive overhaul of PSUs, all the way down to the VRMs on the motherboard.

 

From subsequent comments, it appears it is more than that. I am of course intensely curious, but respect Emile's interest in protecting his IP.

[dminches]

Could someone explain to me how vibrations can get into the digital stream and cause issues?  I assume it is actually affecting the electronic stream of bits?

 

I am not saying it can't affect things; I would just like to understand it better.

 

[Nenon]

43 minutes ago, dminches said:

Could someone explain to me how vibrations can get into the digital stream and cause issues?  I assume it is actually affecting the electronic stream of bits?

 

I am not saying it can't affect things; I would just like to understand it better.

 

 

Vibrations impact the oscillator / clock operation and create phase noise, i.e. audible jitter. There are many studies on this if you search the web. Think about where the clocks are in your system and which clocks are most important - that's the place where you need better care for vibrations. 

[marce]

On 11/4/2019 at 10:35 PM, EuroDriver said:

 - RF Emissions shaping

What is RF emission shaping?

[marce]

20 hours ago, dminches said:

Could someone explain to me how vibrations can get into the digital stream and cause issues?  I assume it is actually affecting the electronic stream of bits?

 

I am not saying it can't affect things; I would just like to understand it better.

 

Depends where the kit is being operated, domestic homes are low vibration environment... a helicopter is not.. To affect the bit stream would take some serious vibration, clocks can be effected, but even then you'd need a lot of vibration to make a difference that is audible IMO.

For the record, the main difference between Commercial and Industrial components is temperature range, most other things are equal.

[rando]

On 11/3/2019 at 4:41 PM, bmichels said:

question: is HQplayer not used anymore ?  HQplayer was a major foundation of the SGM package... and also a source of complexity of set-up/installation to optimize the SGM because of the myriad of choices/parameters/filters available through  HQplayer. 

 

I owe @Miska a small apology for the casual construction of this exact question over the Summer.  A visit to SGM's site made no mention of, and still listed the Evo, Taiko Extreme.  Business relationships and directions change.  Far from trolling him, I was exploring every option known to me saying "I'll be watching the next move they make very closely."  🤐

 

If I found myself in NY a listen to this new server would be scheduled.  If it could be had given the immense popularity.  

[EuroDriver]

On 11/7/2019 at 1:27 PM, marce said:

What is RF emission shaping?

 

Each and every clock in a digital device is a source of RF emissions.  The frequency and amplitude of these clock emissions are modulated by

 - power supply fluctuations

 - vibrations

 

These RF emissions can be attenuated but they still get everywhere, through the air, piggy backing off the AC supply cables, piggy backing on network cabling

 

Random modulation of clock emissions are not so harmful, tape noise does not kill the sound of a master tape.  However music signal correlated noise is harmful and needs to be managed to be less harmful.

 

Why do CD transports have a hard time competing with SOTA Music Servers ? It's the vibrations coming off the CD drive.  Really good CD transports look like turntables for a reason, vibration control

 

You can do a test with a perfectly optically isolated DAC, you do things to the music PC, and you can hear it

[marce]

16 minutes ago, EuroDriver said:

 

Each and every clock in a digital device is a source of RF emissions.  The frequency and amplitude of these clock emissions are modulated by

 - power supply fluctuations

 - vibrations

 

These RF emissions can be attenuated but they still get everywhere, through the air, piggy backing off the AC supply cables, piggy backing on network cabling

 

Random modulation of clock emissions are not so harmful, tape noise does not kill the sound of a master tape.  However music signal correlated noise is harmful and needs to be managed to be less harmful.

 

Why do CD transports have a hard time competing with SOTA Music Servers ? It's the vibrations coming off the CD drive.  Really good CD transports look like turntables for a reason, vibration control

 

You can do a test with a perfectly optically isolated DAC, you do things to the music PC, and you can hear it

So what you are saying is basically EMC .  The term you use implies more.

Analogue noise such as a tape adds to the analogue signal though and does have an effect on the resulting signal.

I would not have thought the levels of vibration would have such an effect on the clock stability.

As to your last comment, you have problems, big problems if you have an optically isolated DAC and you are still getting EMC issues...

 

[Blackmorec]

5 hours ago, marce said:

So what you are saying is basically EMC .  The term you use implies more.

Analogue noise such as a tape adds to the analogue signal though and does have an effect on the resulting signal.

I would not have thought the levels of vibration would have such an effect on the clock stability.

As to your last comment, you have problems, big problems if you have an optically isolated DAC and you are still getting EMC issues...

 

Optical isolation, assuming its actually on the input to the DAC will only protect against conducted EMI. If your room has sources of main’s born or radiated EMI or RFI, optical isolation does nothing. If your optical isolation is further upstream than your DAC, before your server for example, then any noise generated by the server or picked up by its cables may be problematic. The DAC itself can pick up radiated EMI and RFI and can also self-generate problems if its not thoroughly well designed .

The difference between analog and digital is that with analog, any noise that pervades the system cannot be removed without damaging the signal and SQ. In digital the data stream can be cleaned and reformatted before conversion to analog with a positive impact on the SQ. Once cleaned and formatted however its essential that the digital stream be well protected from conducted and radiated EMI and RFI. 

[Nenon]

On 11/7/2019 at 6:38 AM, marce said:

Depends where the kit is being operated, domestic homes are low vibration environment... a helicopter is not.. To affect the bit stream would take some serious vibration, clocks can be effected, but even then you'd need a lot of vibration to make a difference that is audible IMO.

 

9 hours ago, marce said:

I would not have thought the levels of vibration would have such an effect on the clock stability.

 

I am not a scientist, and it's hard for me to interpret some of the studies. Here is a chart table I have saved on my computer (can't remember where I got it from, but I am sure I can find it if I look for it).

It shows oscillators phase noise without vibration and under vibration.  The problem is I have no idea what 0.015 g2/Hz correlates to. Is it an earthquake with magnitude of 7+? Or is it vibration that can be caused by sound waves in our listening room, or transformers, or from our AC mains, etc. If someone can give a real-world example of that value, I would appreciate it. 

 

One thing I can say for sure, though. Vibration treatment on my computer and my DAC is clearly audible in my system. There is no question about it. I was looking for a logical explanation why, and vibrations impacting oscillators made the most sense to me. That does not explain why vibration treatment helps my phono stage, which does not have any oscillators, so there must be more to it. 

 

BTW, that Taiko Extreme motherboard looks a lot like this one - https://www.asus.com/us/Motherboards/WS-C621E-SAGE/overview/

 

 

 

[Nenon]

2 minutes ago, Nenon said:

BTW, that Taiko Extreme motherboard looks a lot like this one - https://www.asus.com/us/Motherboards/WS-C621E-SAGE/overview/

 

I guess that's not news. It has been publicly announced already.

Quote

It’s the Asus WS C621E Sage with 2X Silver 4114.

Source: https://www.whatsbestforum.com/threads/taiko-audio-sgm-extreme-the-crème-de-la-crème.27433/post-557810

 

 

[Miska]

1 hour ago, Nenon said:

One thing I can say for sure, though. Vibration treatment on my computer and my DAC is clearly audible in my system. There is no question about it. I was looking for a logical explanation why, and vibrations impacting oscillators made the most sense to me. That does not explain why vibration treatment helps my phono stage, which does not have any oscillators, so there must be more to it. 

 

With asynchronous clocking, only clock involved in audio playback is the one inside DAC.

 

Electrolytic capacitors are massively more microphonic than clocks. That would probably explain effect on DACs and phono stage.

 

[Nenon]

9 minutes ago, Miska said:

With asynchronous clocking, only clock involved in audio playback is the one inside DAC.

That's the theory but not what I hear. Upstream clock changes still had some impact in my system when I was using asynchronous data transfer. I use I2S now, so it matters even more.

 

13 minutes ago, Miska said:

Electrolytic capacitors are massively more microphonic than clocks. That would probably explain effect on DACs and phono stage.

That makes perfect sense.