I am toasting @audioquattr right now. His contribution has been clearly invaluable behind the scenes and valued by Taiko and by connection all of us. Imagining a serious audio lab at his shop and months of dedicated listening. Thanks to all the behind the scenes passionate contributors named and unnamed.
Taiko Audio SGM Extreme : the Crème de la Crème
- Thread starter CKKeung
- Start date
Wowee Emile. FANTASTIC UPDATE.. Well its really a UPGRADE because its well beyond a update. Thank you for your response on the thread about using basic stuff. Its all about science not voodoo.
Moving to a long basic cat6 terminated with basic full plastic connectors clearly improved SQ over optical. No going back to optical.
I don't post much and so I have a few questions I have saved up.
1. Would it be best to just pull the optical card ? Less interrupts on the buss, less going on, less power drawn, one less driver..
2. Voodoo. Back in the 1980's a new format came out. CD. While its specs were way better then anything known, it sounded horrific. In studios digital audio was also horrible and almost every digital cable used sounded different. There was all sorts of voodoo boxes. YEARS later after a battle with humans doing double blind AB were nearly 100% able to pick out digital and defined it as worse SQ did a investigation reveal Jitter was perceptable by the brain. From that moment on gear, cables, digital mediums were all tested for jitter and the industry cleaned up the voodoo factor. The issue was something we did not know to measure for.
Right now we have a HUGE voodoo problem. Mfgrs claim RF, Jitter and things we cant measure all effect USB/Ethernet/Servers/DACS/Routers/Switches and more. I have seen things inside some of these boxes that really are pure fiction. We can all VERY CLEARLY hear HUGE differences in SQ with all these items, even complete laypeople can hear these changes of a USB cable or Ethernet cable. It can be more dramatic then a speaker wire or line level interconnect.
SO.. What is it that we should be measuring. Is it RF ? I am not sure that explains it fully. Is it packet jitter ? That cant explain it fully. As a industry I think its time to invest in figuring this out. A ethernet interface is a HIGHLY complex signal that uses a huge number of parameters it negotiates on connection including constellation sizes and a list of things. Because it uses levels and not just 1/0 and a D/A and then A/D is used its really a analog medium that conveys digital data. To analyze noise on a Ethernet cable or jitter on the physical layer you need a $150,000 MSO with a load of options for Ethernet testing. https://download.tek.com/datasheet/DPO-DSA-MSO70000-DataSheetEN-US-56W-23446-37.pdf
The question. Is there a push to 100% quantify and define what parameters are causing the issues that your gear is so well removing ?
3. The long run of UTP makes a lot of sense technically. Its a transmission line. A kinda longer run will naturally clean up and filter the highly complex ethernet signal. Is there a standing wave ratio kinda thing going on with the highly complex 1Gbps ethernet signal ? So would exact multiples of a lenght be best to terminate at as long as you know the cable velocity factor ? Is a specific turn radius important ? Should it be coiled or straight lined ?
4. Ground loops. If a system has galvanic isolation from the outside world by fiber to a switch at the audio gear to isolate the whole room ground wise when it has a Torus isolation system and a chemical ground rod system to create a isolated ground, I would think it would still be best to preserve that isolation even tho ethernet has isolation transformers ? The Extreme would be a long run of Ethernet to that switch. Then the feed from the house would be optical. Also if other devices like AppleTV, DirecTV, misc other devices are on that same network in that same room those should be optical isolated from the switch the Taiko is connected to.. All this isolation is to prevent RF from spreading from a device like AppleTV/DirecTV to the Taiko switch. Does this make sense ? Or do you feel the isolation of the ethernet transformers is enough ?
5. RF.. Unshielded ethernet is a truly bad wideband source of RF noise. I would suggest to everybody to try and keep these cables and loops of cables away from analog interconnects like phono cables, preamp outputs and things. Is that sound advice ?
OK.. Too long a post. But I had saved up some questions.
WONDERFUL UPGRADE...
Minor note, maybe its on purpose. The USB1 powershell script seems to have a double entry. Maybe trying to hit the registry twice, sometimes windows does need a good slap.. hahaha..
"HKLM:\SYSTEM\CurrentControlSet\Services\TaikoUsbAudio\ParametersDriver\Settings\AsioInstance0\PredefinedBufferSizesDS\DSD128" -Name "PredefinedAsioBufferSize" -value 128
Moving to a long basic cat6 terminated with basic full plastic connectors clearly improved SQ over optical. No going back to optical.
I don't post much and so I have a few questions I have saved up.
1. Would it be best to just pull the optical card ? Less interrupts on the buss, less going on, less power drawn, one less driver..
2. Voodoo. Back in the 1980's a new format came out. CD. While its specs were way better then anything known, it sounded horrific. In studios digital audio was also horrible and almost every digital cable used sounded different. There was all sorts of voodoo boxes. YEARS later after a battle with humans doing double blind AB were nearly 100% able to pick out digital and defined it as worse SQ did a investigation reveal Jitter was perceptable by the brain. From that moment on gear, cables, digital mediums were all tested for jitter and the industry cleaned up the voodoo factor. The issue was something we did not know to measure for.
Right now we have a HUGE voodoo problem. Mfgrs claim RF, Jitter and things we cant measure all effect USB/Ethernet/Servers/DACS/Routers/Switches and more. I have seen things inside some of these boxes that really are pure fiction. We can all VERY CLEARLY hear HUGE differences in SQ with all these items, even complete laypeople can hear these changes of a USB cable or Ethernet cable. It can be more dramatic then a speaker wire or line level interconnect.
SO.. What is it that we should be measuring. Is it RF ? I am not sure that explains it fully. Is it packet jitter ? That cant explain it fully. As a industry I think its time to invest in figuring this out. A ethernet interface is a HIGHLY complex signal that uses a huge number of parameters it negotiates on connection including constellation sizes and a list of things. Because it uses levels and not just 1/0 and a D/A and then A/D is used its really a analog medium that conveys digital data. To analyze noise on a Ethernet cable or jitter on the physical layer you need a $150,000 MSO with a load of options for Ethernet testing. https://download.tek.com/datasheet/DPO-DSA-MSO70000-DataSheetEN-US-56W-23446-37.pdf
The question. Is there a push to 100% quantify and define what parameters are causing the issues that your gear is so well removing ?
3. The long run of UTP makes a lot of sense technically. Its a transmission line. A kinda longer run will naturally clean up and filter the highly complex ethernet signal. Is there a standing wave ratio kinda thing going on with the highly complex 1Gbps ethernet signal ? So would exact multiples of a lenght be best to terminate at as long as you know the cable velocity factor ? Is a specific turn radius important ? Should it be coiled or straight lined ?
4. Ground loops. If a system has galvanic isolation from the outside world by fiber to a switch at the audio gear to isolate the whole room ground wise when it has a Torus isolation system and a chemical ground rod system to create a isolated ground, I would think it would still be best to preserve that isolation even tho ethernet has isolation transformers ? The Extreme would be a long run of Ethernet to that switch. Then the feed from the house would be optical. Also if other devices like AppleTV, DirecTV, misc other devices are on that same network in that same room those should be optical isolated from the switch the Taiko is connected to.. All this isolation is to prevent RF from spreading from a device like AppleTV/DirecTV to the Taiko switch. Does this make sense ? Or do you feel the isolation of the ethernet transformers is enough ?
5. RF.. Unshielded ethernet is a truly bad wideband source of RF noise. I would suggest to everybody to try and keep these cables and loops of cables away from analog interconnects like phono cables, preamp outputs and things. Is that sound advice ?
OK.. Too long a post. But I had saved up some questions.
WONDERFUL UPGRADE...
Minor note, maybe its on purpose. The USB1 powershell script seems to have a double entry. Maybe trying to hit the registry twice, sometimes windows does need a good slap.. hahaha..
"HKLM:\SYSTEM\CurrentControlSet\Services\TaikoUsbAudio\ParametersDriver\Settings\AsioInstance0\PredefinedBufferSizesDS\DSD128" -Name "PredefinedAsioBufferSize" -value 128
Last edited:
Indeed, but I feel like a rabbit watching holes pop up all around me.
Attachments
Based on your photo, you are using the Nagra HD DAC X, is that correct?
I have the Tube DAC and plan to add the HD DAC X to my system late this year/early next year.
well, last night i replaced my 50m fiber run with the 50m cat6 that runs in parallel and has been in my system from the earliest days of my streaming experiments (logitech touch, remember those?)
all i can do is quote the pillow on my sister-in-law's couch: "Too much of a good thing is WONDERFUL"
can't wait for xmds!
Yes. Latest build is 903. Don't forget to update your Roon Remote App to 903, as well.
Well with Emile's magic, no more kung fu required
In fact I gotta dekung-fu things. Taiko made me have to rethink things. And I am very happy about it.
Last edited:
Taiko clearly has the best understanding of the mechanisms at work, and clearly no voodoo in your boxes, just pure science. I sent ya a email. I would love a chat. I want to find a way to try and actually measure things that correlate to what we all hear with these digital streams.
Thank you!
1. It'll enter a low power state when unused (sleep). I'd leave it there for now.
2. I think we'll leave that to disclose when we launch the network card and switch if you don't mind
3. Turns do degrade signal integrity, google "belden bonded pair" as an example, but signal integrity is not really an issue. There are some differences in sonics between shorter and longer lengths, especially apparent when you use a nearby switch, shorter is then more "incisive", longer more "relaxed".
4. I'd say the noise generated by the SFP receiver is larger then whatever residual "bleeds" through over copper, a cable shield can form a direct unhindered connection though.
5. Avoiding loops should be a priority imho. 2 interconnected components with ground leads to the same ground can be a loop too btw, perhaps interesting food for thought!
Could someone tell me what XDMS is or point me to some information?
Thanks!
Our upcoming music playback software suite.
Post #9,227
https://www.whatsbestforum.com/threads/taiko-audio-sgm-extreme-the-crème-de-la-crème.27433/page-462#post-778608
Last edited:
Many of us have shielded LAN cable installed in-wall. It may be impractical to run an external unshielded cable. If it is the shielding that conduct noise, can we just use a RJ45 connector to connect a cat 6 unshielded cable just before the extreme?
OMG.. You KNOW the answer.. Your going to make me crazy not knowing till then..
But WHY !! hahahaha..
OMG what do I do with a straight 15 foot cable, hehehe..
Congratulations on a voodoo killer. The network card and switch should be great fun.
Side note. I did a quick test. Best to keep a unshielded cable reasonably away from other wires. Most likely very unlikely to cause a issue, but shielded cables have none of this. So don't loop it and try and stay away from other cables.
Taiko was updated this morning, it's being done on a first come first serve basis, and since they had room on the schedule it was completed in all of 10 minutes.
The MSB / Taiko system I know extremely well, it is very holographic and dimensional and easily capable of filling the room in this manner.
This software update forcing me now to use Roon brought a further level of control, detail and depth on top of tighter imaging then previous.
There is no doubt the servers choice of hardware, assembly and software is some of the best on the planet, and when software can increase performance, it's a comforting fact that a product like Taiko is worth the expense by every stretch!
Great job to Emile and Company!
The MSB / Taiko system I know extremely well, it is very holographic and dimensional and easily capable of filling the room in this manner.
This software update forcing me now to use Roon brought a further level of control, detail and depth on top of tighter imaging then previous.
There is no doubt the servers choice of hardware, assembly and software is some of the best on the planet, and when software can increase performance, it's a comforting fact that a product like Taiko is worth the expense by every stretch!
Great job to Emile and Company!
