Introducing Olympus & Olympus I/O - A new perspective on modern music playback

[Taiko Audio]

Thanks for being so clear - this means that people currently owning or considering top DACs will have to use AES or SPDIF lines - implying that clocks must be extracted from these lines that carry the digital data. The limitations and problems of these lines, when used asynchronously have been addressed extensively for decades, can we know why you think that this a preferred mode of data transmission to a DAC?

I’m sorry but I wrote so much about this already that I would kindly request you to use the “search this thread” for posts by member “Taiko Audio” function on this particular subject. If that leaves any lingering questions, which I strongly doubt with your level of understanding, I’m more then happy to answer those

 

[Taiko Audio]

I enjoy Stax headphones. Will Olympus be compatible with them too? Thanks

If that deviates significantly from common headphones (which it appears to do) then we’d need to design something for it specifically, there will need to be a demand!

 

[Taiko Audio]

Can the new storage modules be used together with the old modules migrated from the Extreme? For instance, if an Extreme owner with 8TB of storage in his Extreme wished to transfer those drives to his new Olympus and also purchase an additional, say, 8TB of storage space, would you recommend the new generation drives for the expanded storage? Or would that create compatibility issues?

Yes that’s possible. You can use both simultaneously. I’d be inclined when it comes to expansion to use a new drive where 8TB is pretty small.

 

[Taiko Audio]

Seconded - would love to have XLR analog outputs on mine as my amps only take XLR inputs. Would like to avoid an adapter if possible. Goal is ultimately to interface digital XDMI to the Horizon with whatever Emile/Lukasz come up with, though.

@Taiko Audio would ripping out these three ports give enough room?
Merry Christmas!
View attachment 122322

I’m sure we can come up with something “fitting”.

 

[Taiko Audio]

Nice. Had asked a few days ago about the potential of xlr out from Olympus to mono balanced amps in the future/or current.

understood only rca currently available. Is xlr possible or is this a different animal vs post above re the positive that balanced out for olympus headphone amp?

A balanced output more or less entails doubling up the output stage. So that’s not particularly difficult to do. Volume control is a different matter and there are multiple ways to implement that which we already started exploring.

The mayor effort went into designing XDMI as a whole, designing more output options are a relatively minor exercise. Though naturally we could turn it into a mayor design effort if interest is high enough. Which so far seems to be the case, but we’d like to await feedback from the field on how XDMI is going to be used most in practice to determine the main direction of our efforts.

These efforts can be anything, from designing a better source to dac interconnect system, to adopting DAC manufacturers specific interface implementations, to designing even better analogue output options, including a diversity of analogue equipment to drive.

 

[cmarin]

A balanced output more or less entails doubling up the output stage. So that’s not particularly difficult to do. Volume control is a different matter and there are multiple ways to implement that which we already started exploring.

The mayor effort went into designing XDMI as a whole, designing more output options are a relatively minor exercise. Though naturally we could turn it into a mayor design effort if interest is high enough. Which so far seems to be the case, but we’d like to await feedback from the field on how XDMI is going to be used most in practice to determine the main direction of our efforts.

These efforts can be anything, from designing a better source to dac interconnect system, to adopting DAC manufacturers specific interface implementations, to designing even better analogue output options, including a diversity of analogue equipment to drive.

Hi Emile,

If we’re starting to count votes, my own preference is to use: (1) balanced analog output to a preamp (I don’t envision not using my current CH Precision L10 anytime in the near future - but never say never); and (2) AES digital output to DAC - specifically to a totaldac sublime.

 

[Taiko Audio]

Hi Emile,

If we’re starting to count votes, my own preference is to use: (1) balanced analog output to a preamp (I don’t envision not using my current CH Precision L10 anytime in the near future - but never say never); and (2) AES digital output to DAC - specifically to a totaldac sublime.

(2) we already have

 

[M27HF]

Hi Emile,

If we’re starting to count votes, my own preference is to use: (1) balanced analog output to a preamp (I don’t envision not using my current CH Precision L10 anytime in the near future - but never say never); and (2) AES digital output to DAC - specifically to a totaldac sublime.

I'm a fan of #1 too

 

[LampiNA]

I think we can measure all or our percolation with a seismic meter! I just felt a rumble! I sailed with another guy many years ago from Miami to St. Thomas 55'Cat, NO LORAN, (before GPS) Dead Reckoning. I feel like I'm back in that catamaran! We are kind of in unchartered waters. We all want to make informed decisions. Many of us have become quite fond of the Horizon and would like to be able to combine HDMI in the fold. Hence the uncharted waters. Do we "stick" with USB until Lampizator creates an alternative as @Steve Williams referenced? We all have some EXTREMELY interesting decisions to make....

Rest assured, guys. LampizatOr is on the case and presently working on being able to fully leverage XDMI in order that the best DAC on the planet can leverage the capabilities of the best Taiko server. Stay tuned all, we'll have an announcement in the New Year!

Best,

Fred A.

 

[microstrip]

I’m sorry but I wrote so much about this already that I would kindly request you to use the “search this thread” for posts by member “Taiko Audio” function on this particular subject. If that leaves any lingering questions, which I strongly doubt with your level of understanding, I’m more then happy to answer those.

Unfortunately, as far as see it, none of the technical problems I am specifying was addressed. You wrote about your preferences, something that I respect and like to know, but a DAC recovering its intrinsic timing from the AES/EBU is a Pandora box, mainly because the techniques used by DAC designers to extract the clock are very different, as well as their jitter and noise rejection techniques.

Your asynchronous USB solutions sound really great, you are now going in an orthogonal way towards a synchronous transmission that classically needs an external clock line to sound optimum. I think that such subject deserves some comment. We had long technical posts from you in the past addressing the USB subject in the past - they managed to make an happy Taiko customer - I would be very happy to read on this different problem.

 

[Taiko Audio]

Unfortunately, as far as see it, none of the technical problems I am specifying was addressed. You wrote about your preferences, something that I respect and like to know, but a DAC recovering its intrinsic timing from the AES/EBU is a Pandora box, mainly because the techniques used by DAC designers to extract the clock are very different, as well as their jitter and noise rejection techniques.

Your asynchronous USB solutions sound really great, you are now going in an orthogonal way towards a synchronous transmission that classically needs an external clock line to sound optimum. I think that such subject deserves some comment. We had long technical posts from you in the past addressing the USB subject in the past - they managed to make an happy Taiko customer - I would be very happy to read on this different problem.

AES/EBU is just an optional output module.. It’s technical deficiencies are intrinsic to its design. This has little to nothing to do with XDMI beyond it being a highly compatible interface allowing early adopters to hear what it can do with their own DACs, and compare it to direct analogue out or USB.

I’ll quote some of my posts for context.

 

[Taiko Audio]

@microstrip I ended up sending them in PM due to the forum post length restriction.

 

[MarcelNL]

regardless of the forum restrictions I'm also interested on the thoughts behind timing between DAC (where it should reside IMHO) and source

 

[Rob181]

regardless of the forum restrictions I'm also interested on the thoughts behind timing between DAC (where it should reside IMHO) and source

And you can add me to this list - thanks

 

[jeremya]

@Taiko Audio this is beautiful, compelling work!

Will there be a way for DIY Extreme builders to add an Olympus I/O to their setup? (A PCIe to XDMI interface board perhaps?)

 

[Taiko Audio]

@Taiko Audio this is beautiful, compelling work!

Will there be a way for DIY Extreme builders to add an Olympus I/O to their setup? (A PCIe to XDMI interface board perhaps?)

Hi @jeremya ,

In theory yes, you should be able to add an Olympus I/O to your DIY server assuming it runs windows 10 or 11 and uses either Intel Xeon datacenter (not the xeon versions of consumer level chips) or AMD Epyc gen3 or gen4 CPUs. However we just cannot guarantee performance as we cannot put considerable effort in supporting this very limited use case. As an example we did indeed attempt to supply some components for the DIY market in the past, but we ended up spending a lot of time and selling very few components. For example we ran a batch of 100 pieces of a GaN fet based DC-DC ATX power-supplies and we sold ..35.. While the accompanying ULPS design ended up being copied 1 on 1 in a competitor’s product.. We’re absolutely sympathetic to the DIY community but in the end we do have a business to run and our main focus should be on our customers buying our servers.

 

[Taiko Audio]

regardless of the forum restrictions I'm also interested on the thoughts behind timing between DAC (where it should reside IMHO) and source

In a digital interface system the source and destination device clocks should be synchronised, aka there needs to be a master clock.

Ideally the DAC’s clock should be the system's clock master as jitter is most significant where digital signals are converted to analogue. Timing is most critical here as samples should be reconstructed at the right moment in time. This is something proprietary interfaces like MSB’s Pro ISL do, the DAC clock is the master and provides the clock for the source.

AES/EBU and SPDIF are self clocking, the source is the master clock, the clock signal is embedded in the datastream. This is not ideal. So a lot of high-end dacs use a buffer (reclocker) to isolate the embedded signal clock from the DAC clock.

An alternative to using a buffer/reclocker is to use an external master clock to provide a clock to both source and signal. This will need to be supported by both source and DAC, they will both need to have an external clock input.

 

[jeremya]

Hi @jeremya ,

In theory yes, you should be able to add an Olympus I/O to your DIY server assuming it runs windows 10 or 11 and uses either Intel Xeon datacenter (not the xeon versions of consumer level chips) or AMD Epyc gen3 or gen4 CPUs. However we just cannot guarantee performance as we cannot put considerable effort in supporting this very limited use case. As an example we did indeed attempt to supply some components for the DIY market in the past, but we ended up spending a lot of time and selling very few components. For example we ran a batch of 100 pieces of GaN fed based DC-DC ATX power-supplies and we sold ..35.. While the accompanying ULPS design ended up being copied 1 on 1 in a competitor’s product.. We’re absolutely sympathetic to the DIY community but in the end we do have a business to run and our main focus should be on our customer base buying our servers.

That's completely understandable, Emile. Bummer to hear someone copied your design (ugh).

As you may recall, my DIY Extreme is based on the same motherboard (ASUS SAGE), CPUs (Xeon Silvers), and underlying OS (Win10 LTS) as the Extreme, so it should be "plug and play." I still have my DC-DC ATX board waiting patiently for me to assemble Nenon's v3 unregulated linear PSU. I'm hoping to knock that out in the next week or so... once the family holiday visitation calendar calms down a bit. =^)

 

[MarcelNL]

Hi @jeremya ,

In theory yes, you should be able to add an Olympus I/O to your DIY server assuming it runs windows 10 or 11 and uses either Intel Xeon datacenter (not the xeon versions of consumer level chips) or AMD Epyc gen3 or gen4 CPUs. However we just cannot guarantee performance as we cannot put considerable effort in supporting this very limited use case. As an example we did indeed attempt to supply some components for the DIY market in the past, but we ended up spending a lot of time and selling very few components. For example we ran a batch of 100 pieces of a GaN fet based DC-DC ATX power-supplies and we sold ..35.. While the accompanying ULPS design ended up being copied 1 on 1 in a competitor’s product.. We’re absolutely sympathetic to the DIY community but in the end we do have a business to run and our main focus should be on our customers buying our servers.

WOW that few? You should have signed all of them, would have made my 2 ATX boards unique collectors items ;-)

We almost had more custom chokes made when it came to the point of replacing the Hammond chokes in the ULPS for something more substantial.

 

[TWJeff]

In a digital interface system the source and destination device clocks should be synchronised, aka there needs to be a master clock.

Ideally the DAC’s clock should be the system's clock master as jitter is most significant where digital signals are converted to analogue. Timing is most critical here as samples should be reconstructed at the right moment in time. This is something proprietary interfaces like MSB’s Pro ISL do, the DAC clock is the master and provides the clock for the source.

AES/EBU and SPDIF are self clocking, the source is the master clock, the clock signal is embedded in the datastream. This is not ideal. So a lot of high-end dacs use a buffer (reclocker) to isolate the embedded signal clock from the DAC clock.

An alternative to using a buffer/reclocker is to use an external master clock to provide a clock to both source and signal. This will need to be supported by both source and DAC, they will both need to have an external clock input.

As an Extreme and Vivaldi System user, it would be good if Vivaldi Clock can also output clock signal to Taiko Server! Which would be an option for tuning the sound and technically locking the digital components all together by an external master clock.