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

For those who just started reading up on Olympus, Olympus I/O, and XDMI, please note that all information in this thread has been summarized in a single PDF document that can be downloaded from the Taiko Website.

https://taikoaudio.com/taiko-2020/taiko-audio-downloads

The document is frequently updated.

Scroll down to the 'XDMI, Olympus Music Server, Olympus I/O' section and click 'XDMI, Olympus, Olympus I/O Product Introduction & FAQ' to download the latest version.

Good morning WBF!​

We are introducing the culmination of close to 4 years of research and development. As a bona fide IT/tech nerd with a passion for music, I have always been intrigued by the potential of leveraging the most modern of technologies in order to create a better music playback experience. This, amongst others, led to the creation of our popular, perhaps even revolutionary, Extreme music server 5 years ago, which we have been steadily improving and updating with new technologies throughout its life cycle. Today I feel we can safely claim it's holding its ground against the onslaught of new server releases from other companies, and we are committed to keep improving it for years to come.

We are introducing a new server model called the Olympus. Hierarchically, it positions itself above the Extreme. It does provide quite a different music experience than the Extreme, or any other server I've heard, for that matter. Conventional audiophile descriptions such as sound staging, dynamics, color palette, etc, fall short to describe this difference. It does not sound digital or analog, I would be inclined to describe it as coming closer to the intended (or unintended) performance of the recording engineer.

Committed to keeping the Extreme as current as possible, we are introducing a second product called the Olympus I/O. This is an external upgrade to the Extreme containing a significant part of the Olympus technology, allowing it to come near, though not entirely at, Olympus performance levels. The Olympus I/O can even be added to the Olympus itself to elevate its performance even further, though not as dramatic an uplift as adding it to the Extreme. Consider it the proverbial "cherry on top".

 

Those should appear sometime soon.

Thank you

 

Dear Taiko team,
When do you expect to publish the addicional cost for each size of internal hardrive (above 4Tb) for the Olympus?

 

Haha, nahh, they had a good stress test in Munich, good to go!

 

Will post updates on Wednesday! I’m currently out of office till then due to my kids May school holidays.

I can share that we more or less solved the storage SQ differentials to within about a ~1% margin. Will share the details on that on Wednesday as well.

@Taiko Audio: Before you left for Munich, you said that you had "more or less solved the storage SQ differentials to within about a ~1% margin." Unless I missed it, you haven't gone into any more detail about this. Since "storage SQ differentials" might be relevant to how much internal storage people will want to order with their new Olympus, could you provide further details on this?

 

@Taiko Audio: Before you left for Munich, you said that you had "more or less solved the storage SQ differentials to within about a ~1% margin." Unless I missed it, you haven't gone into any more detail about this. Since "storage SQ differentials" might be relevant to how much internal storage people will want to order with their new Olympus, could you provide further details on this?

What we solved is that adding an internal music drive no longer degrades SQ. What we didn’t “solve” is that network based playback is still slightly ahead, by about 1%, IF you own the Router/Switch/DCD with a good quality power supply. If you don’t local playback is superior, easily 10-15% better.

In practical terms this makes a NAS a cost effective solution for large libraries IF you own the Router/Switch/DCD, I’d say absolutely interesting to consider if you expect to exceed a 16Gb library size.

 

Dear Taiko team,
When do you expect to publish the addicional cost for each size of internal hardrive (above 4Tb) for the Olympus?

@tsaett is currently digging through all the quotes, he’ll post an update soon.

 

What we solved is that adding an internal music drive no longer degrades SQ. What we didn’t “solve” is that network based playback is still slightly ahead, by about 1%, IF you own the Router/Switch/DCD with a good quality power supply. If you don’t local playback is superior, easily 10-15% better.

In practical terms this makes a NAS a cost effective solution for large libraries IF you own the Router/Switch/DCD, I’d say absolutely interesting to consider if you expect to exceed a 16Gb library size.

Thanks for synthesizing Emile. Did you test impact of different NAS options on SQ? I have the Router/Switch/DCD + Olympus.

 

Thanks for synthesizing Emile. Did you test impact of different NAS options on SQ? I have the Router/Switch/DCD + Olympus.

No I only tested a 2 bay Synology. As the SQ is identical to streaming (same files downloaded from Qobuz) I wouldn’t expect much of a difference there, if any..

 

I will but it’s quite a story to share so it requires some time to prepare/write. And we’re at the Munich show right now!

<bump>

 

What we solved is that adding an internal music drive no longer degrades SQ. What we didn’t “solve” is that network based playback is still slightly ahead, by about 1%, IF you own the Router/Switch/DCD with a good quality power supply. If you don’t local playback is superior, easily 10-15% better.

In practical terms this makes a NAS a cost effective solution for large libraries IF you own the Router/Switch/DCD, I’d say absolutely interesting to consider if you expect to exceed a 16Gb library size.

My library is 16TB and I went with a NAS I was going to buy 32 TB of internal memory but went with an NAS 4 Bay with 4x 16TB WD Red Pro in RAID1

 

My library is 16TB and I went with a NAS I was going to buy 32 TB of internal memory but went with an NAS 4 Bay with 4x 16TB WD Red Pro in RAID1

Hi Steve,

That's a heck of a redundancy you've got there!

If you place 4 x 16TB disks in RAID1, you'll have less than 16TB of capacity.

It would probably be more interesting to put them in RAID6 (or SHR2, if you're using a Synology NAS).

It's probably worth considering

Cheers,

Thomas

EDIT :

And, if I may , a word of advice: buy your discs from different sources to ensure they don't come from the same batch, so that when they fail, and they will fail, they won't all fail at the same time.

 

I am running RAID0 on my 2 drive Synology NAS. My library is always backed up so I don't need any redundancy.

 

Hi Steve,

That's a heck of a redundancy you've got there!

If you place 4 x 16TB disks in RAID1, you'll have less than 16TB of capacity.

It would probably be more interesting to put them in RAID6 (or SHR2, if you're using a Synology NAS).

It's probably worth considering

Cheers,

Thomas

EDIT :

And, if I may , a word of advice: buy your discs from different sources to ensure they don't come from the same batch, so that when they fail, and they will fail, they won't all fail at the same time.

It’s all good. I got this advice from Taiko network experts and I’m quite comfortable pursuing this route. I do have my entire library backed up on several HDD. For what it cost me I’m fine with it all. I do appreciate your concern. FWIW I’m using Qnap and the drives were ordered on separate days in 2 groups

 

What we solved is that adding an internal music drive no longer degrades SQ. What we didn’t “solve” is that network based playback is still slightly ahead, by about 1%, IF you own the Router/Switch/DCD with a good quality power supply. If you don’t local playback is superior, easily 10-15% better.

In practical terms this makes a NAS a cost effective solution for large libraries IF you own the Router/Switch/DCD, I’d say absolutely interesting to consider if you expect to exceed a 16Gb library size.

@Taiko Audio Emile, I recall discussion about the startling improvement from playing WAV files, uncompressed from FLAC (or APE etc.), experienced by Ed, uncompressed with such software as dbpoweramp, in contrast to playing FLAC files locally. I play music exclusively from Extreme local HD and only in uncompressed format. The compression ratio used by FLAC or APE (highest level of compression) could reduce original WAV size by almost 2/3. I have known for a long time (thru comparison) that there is a clearly audible improvement of playing WAV files over compressed FLAC or APE files. De-compression in real time (especially when the compression level is high) when such compressed files are being processed in Extreme invariably introduces jitter and noise. Did you compare playing from locally saved WAV files with playing FLAC files streamed from Qobus or Tidal (which steaming I don't do myself; I tried Qobus for a month; I recall its Hi-Re files were in FLAC only)? If so, you conclusion above still holds?

 

@Taiko Audio Emile, I recall discussion about the startling improvement from playing WAV files, uncompressed from FLAC (or APE etc.), experienced by Ed, uncompressed with such software as dbpoweramp, in contrast to playing FLAC files locally. I play music exclusively from Extreme local HD and only in uncompressed format. The compression ratio used by FLAC or APE (highest level of compression) could reduce original WAV size by almost 2/3. I have known for a long time (thru comparison) that there is a clearly audible improvement of playing WAV files over compressed FLAC or APE files. De-compression in real time (especially when the compression level is high) when such compressed files are being processed in Extreme invariably introduces jitter and noise. Did you compare playing from locally saved WAV files with playing FLAC files streamed from Qobus or Tidal (which steaming I don't do myself; I tried Qobus for a month; I recall its Hi-Re files were in FLAC only)? If so, you conclusion above still holds?

On the Olympus I have a very small preference for flac.

 

On the Olympus I have a very small preference for flac.

Over uncompressed WAV? Incredible. If so, I'll find out for myself. Anything is possible. If it is so to my ears, I'll be compressing WAV files in the reverse direction. We have a saying in English: have your cake and eat it too.

 

Over uncompressed WAV? Incredible. If so, I'll find out for myself. Anything is possible. If it is so to my ears, I'll be compressing WAV files in the reverse direction.

Yes but I only compared flac and wav versions of the same files downloaded directly from Qobuz. I'm not spending too much time on these things as I have a long list of mayor infrastructural things to test with potentially much larger rewards, like driver code changes

 

Thanks for synthesizing Emile. Did you test impact of different NAS options on SQ? I have the Router/Switch/DCD + Olympus.

I know it might sound silly but conventional NAS options wouldn't even hold a candle if we're actually comparing them to either Extreme or Olympus.

In other words, the "real" test would be enabling SMB shares on a separate unit of Extreme (with Extreme Network Card) and then use that as our super-duper NAS option.

At some point we'll probably see refurbished Extreme to appear, most likely we'll figure out that we're gonna do about that?

 

About the Olympus design change:

As mentioned earlier in this thread we ended up having 2 Olympus designs.

The original design is already about a year old and was based on an AMD Epyc 32-core CPU with an Asrockrack or Supermicro motherboard (server motherboards are generally based on the same reference design with mainly differences in I/O options offered, networking, storage functionality etc).

We completed and introduced this design to the world in December.

Warning, NERD talk ahead:

As part of my general interest in computer hardware I tend to order every major new CPU / motherboard release to test it's effects on sound quality. Asus announced one of, if not the, most monstrous motherboard ever created in October which I ordered together with its accompanying AMD Threadripper CPU, which is more or less the same CPU as the Epyc, but for the HEDT (High-End Desktop) market. This set new world records of performance with insane power consumptions, like between 1 and 2 Kilowatts (!) of peak power. I more or less forgot all about this but it arrived halfway of January. So we hooked it up, oh my, a true monster indeed, renamed it to FET RIPPER due to it's insane transient current draws on initial boot, we're talking fast repetitive 30-40A spikes, enough to turn most FETs into charcoal. To power this beast with a linear power supply is absolutely challenging, BUT, we have battery supplies which can supply up to 2000A current peaks! Now due to the enormous amount of power consumption it would deplete a battery supply very quickly, within minutes, so I configured a BPS to charge continuously, the BPS doesn't blink at those transients, so we could boot and have a listen.

I basically copied the Olympus design and just exchanged the Motherboard/CPU and it's accompanying power supply. Well of course it didn't sound as good as the already completed Olympus design, BUT, it had the upper bass/midrange density of the Extreme's motherboard/CPUs. I wrote about this before how the Extreme's motherboard/CPUs have this unique density quality no other motherboard/CPU combination has, and rest assured I've tested virtually everything. So this appears to have something to do with Asus's HEDT/Workstation motherboard designs. (It's not present in the Asus motherboards I tested with AMD Ryzen and Intel consumer level chips, so this appears to be something exclusive to their High-end motherboard designs).

Intrigued by this midrange density feature me and Jeroen started tweaking it a bit in our spare time, just for fun, to see if we could achieve the overall performance of the original Olympus design which was on another level in transparency, sound staging, holography, micro detail retrieval and realism. It was actually very far behind so we did not expect to actually get there, but it was a fun spare time project to "tame the beast"! Besides that it would never fit in any reasonably sized chassis, the motherboard was physically larger, required a 4th BPS and a more elaborate cooling solution..

While studying the motherboard architecture we noticed a distinct difference with other motherboards, it's designed to be powered by multiple power supplies to be able to provide the unwieldy power it can consume. It has 2 separate power planes to feed the FET ripper CPU, and another for the PCIe slots(!). With our original Olympus design we deliberately had 4 "grounds" to isolate devices from each other (motherboard/CPU, Network card, Drives and XDMI. 1 ground in the linear power supply, and 3 battery grounds). We then changed the whole system setup to use a hybrid parallel / star ground arrangement using 2 (battery) grounds, minimizing ground current loops resulting in overall better performance (!) while halving the number of battery supplies used. Now it really started to sing AND it would actually fit it into the Olympus chassis! By now we were halfway into February and facing a dilemma, we have 2 Olympus servers, and one of them actually sounds better with by now not only an increase in midrange density and colour palette, but also a significant increase in both micro as macro dynamics. So we pulled the trigger and ordered everything we need to switch over to this design, we did waste some resources on this move, but.. it's just even more "awesome", what can you do? We need to launch the best we can do, even if just for my peace of mind.

In a nutshell this leads to the following changes in the Olympus design:

1) AMD Epyc 32-core -> AMD Threadripper 24-core (86.4 GHz to 100.8 GHz total base clock)
2) 128Gb memory -> 64Gb memory
3) 3*BPS -> 2*BPS
4) 1*LPS + 1* charger PS -> 1* monster charger PS
5) 4* Regulator -> 8* Regulator
6) Larger / more sophisticated heatsink design
7) Even better sound

This is the Olympus as displayed at the Munich show. Functionally there's no difference, it only changes the specs list!

 

Warning, NERD talk ahead:

These last-minute discoveries and changes must have been quite an adventure!

It's really interesting to see the extent to which data transport affects sound quality.

I can imagine that the future holds many more discoveries and possible improvements.

I suppose we'll soon be seeing a new version of the Taiko Audio website, won't we?

Cheers,

Thomas