Taiko Audio SGM Extreme : the Crème de la Crème

[cat6man]

@Taiko Audio
is anyone including an extreme in their digital source chain at CanJam2023 in NYC next saturday?

@all
anyone else here going on saturday?

DCS and Chord will both be there. I wonder if PGGB files will be playable in NYC?

 

[rando]

? Extreme Restart button
? Extreme Shut Down button

Would I be going too far asking if a facet of ongoing software development is a growing need to restart or fully reboot arising?

The Extreme is growing more physically complex. Which no doubt is being accounted for here inside proprietary software managing or communicating with these additions. My suspicion here is this might be leaning more towards BIOS and MS Windows than Taiko products if any perceptible change has been noted. So not directly focused on Alpha player build .current

 

[kennyb123]

Note that Totaldac doesn't use at all the clock transported by the AES-EBU link, it uses exclusively the data. The clock is built completely in the Totaldac DAC or reclocker.

Sorry for the stupid question, but I thought AES-EBU was a synchronous interface with the clock in the source governing what happens downstream. From your wording it almost sounds like you have it operating more like USB. Is that right?

 

[totaldac]

Sorry for the stupid question, but I thought AES-EBU was a synchronous interface with the clock in the source governing what happens downstrea. From your wording it almost sounds like you have it operating more like USB. Is that right?

Yes, it is made to be synchronous but I don't use it that way.
It don't use it like bi-directional asynchronous USB either.

 

[cat6man]

Hello,
Some customers asked me news about the Taiko TACDA/TACDD project with Totaldac.

I talked with Emile, we think that because AES-EBU has no limit in the Totaldac design there is no point integrating a Taiko board in the Totaldac DACs or reclocker. It is just equivalent when this board is mounted in the Taiko server.
So for the best sound from a Taiko + Totaldac system, a Taiko server should be equipped with the future Taiko AES-EBU output (based one of the new TACDA/TACDD board).
The link from the Taiko server to a Totaldac DAC or Totaldac reclocker would be AES-EBU.

The best result should be found when using a d1-digital-sublime reclocker, which is the reclocker developed for the new d1-sublime DAC.
You can find some talk about this DAC and reclocker here: https://www.whatsbestforum.com/threads/totaldac-new-dacs.36082/page-4

this is wonderful news for totaldac owners as we won't need to add h/w to our DACs to take advantage of this.
so TACDD with a aes/ebu looks like the answer for us........can't wait to hear how this plays out and sounds.

since this does not require a DAC change, maybe the aes/ebu daughter card will be the first one released?

 

[gereric]

Emile, does that mean TACDD for TotalDAC owners, since we we wouldn’t need the digital to analog conversion on the PCIe card?

 

[Zeotrope]

Copying a file is different from streaming or sending over USB, let alone over AES/EBU or other synchronous methods. But please note that "Bit Perfect" is not the whole story.

Christiaan, I don’t think we ever got to the bottom of this:
Correct me if I’m wrong but copying (ripping) a CD IS bit perfect if you use error correction and recovery (eg “Secure” on dB Poweramp). Given this, does the CD player hardware matter? I don’t think so.

The reason this is very different from playback is that during playback (whether streamed or local) you don’t have the luxury of error correction and recovery (and the buffer required).

Is that correct?

 

[Zeotrope]

Good ripping hardware :

Ediscreation Ultra Silent Ripper : a CAS best-buy made by Edison Wong in Hong Kong

Following his Silent Switch OCXO, the Ediscreation Ultra Silent Ripper is another CAS best-buy made by Edison Wong in Hong Kong : https://nandeaudio.com/product/ultra_silent_ripper/ A friend of mine bought one last month. We together "opened-box" and tested ripping with different add-on...

www.whatsbestforum.com

I did see this again, thanks!

 

[oldmustang]

this is wonderful news for totaldac owners as we won't need to add h/w to our DACs to take advantage of this.
so TACDD with a aes/ebu looks like the answer for us........can't wait to hear how this plays out and sounds.

since this does not require a DAC change, maybe the aes/ebu daughter card will be the first one released?

Or maybe dual AES/EBU? This would serve both TotalDAC and dCS Vivaldi customers.

Steve Z

 

[nonesup]

I have the impression that if you use the new Taiko card with AES/EBU output, you will get an improvement over the current AES/EBU output, but you will be losing for a part of the "invention". If it wasn't like that and the sound quality was the same, Emile would have been wasting his time and money developing the new interface, and I get the impression that he doesn't usually do this.

 

[Christiaan Punter]

Emile, does that mean TACDD for TotalDAC owners, since we we wouldn’t need the digital to analog conversion on the PCIe card?

The early information surrounding the TACDA and TACDD could be interpreted to mean one is analog and the other is digital but this is not the case. As time went by, new insights were gained and new decisions were made. Rather than making several versions, the board will now be modular and will support output modules that can be analog or digital and can cater to a range of formats. Note that the final name is still pending, “TACDA” is a temporary unofficial name that we use internally.

- TACDA is the card hardware (that can be either analog or digital)
- TACDD is what we call the interface format

Please see the attached TACDA / TACDD documentation for more information.

 

[Christiaan Punter]

Would I be going too far asking if a facet of ongoing software development is a growing need to restart or fully reboot arising?

The Extreme is growing more physically complex. Which no doubt is being accounted for here inside proprietary software managing or communicating with these additions. My suspicion here is this might be leaning more towards BIOS and MS Windows than Taiko products if any perceptible change has been noted. So not directly focused on Alpha player build .current

That's a fair question but there is no need to worry. When used normally with XDMS, the Extreme is in no need to be restarted any more than one would normally do with Roon. Also, I can confirm that the XDMS is actually only getting more stable with every new build. So, the addition of the restart button is not a sign of decreasing stability. However, XDMS is in an Alpha state which implies that it is not yet as stable or complete as full-production software. As part of the Alpha testing, bugs do arise that will indeed sometimes require a restart. Also, as part of the testing, people are more inclined to switch between Roon and XDMS and if that is not done in a particular sequence, some manual settings need to be made a little bit deeper in the system which is not something all testers know how to do or want to dive into. The addition of these buttons in the software just makes it easier for everyone involved.

 

[Christiaan Punter]

Christiaan, I don’t think we ever got to the bottom of this:
Correct me if I’m wrong but copying (ripping) a CD IS bit perfect if you use error correction and recovery (eg “Secure” on dB Poweramp). Given this, does the CD player hardware matter? I don’t think so.

The reason this is very different from playback is that during playback (whether streamed or local) you don’t have the luxury of error correction and recovery (and the buffer required).

Is that correct?

That would sound logical indeed. But I have not made any comparisons with different hardware using error correction, so I cannot confirm from personal experience. All I can say is that in the tests that I did when not using error correction, comparing different ripper/computer hardware and different ripping software would result in a different-sounding rip. For instance, a rip made in iTunes on a Macbook sounded noticeably leaner than one made on a large Windows tower PC. But I also noticed that ripping speed (where this could be adjusted) had a large effect on the sound. Faster rips would also sound leaner. At that time, I eventually settled on 8X. At that time, I preferred not using error correction as I felt it made the music sound overly controlled and "digital", compared to the same CD played on what I think would have been a Levinson 390S. In hindsight, I was compensating for the below-par server hardware that I used at that time. Those were just very different times. But as my system developed and my preferences changed, many years ago, I started noticing that other people's rips sounded better than mine and I decided to switch to a secure ripping method which from that moment on yielded very consistent results. These rips sound tight and articulate (which I would not have liked back in the day), but with incomparably better server/streamer hardware, the sound is no longer "digital" or over-controlled.

 

[Taiko Audio]

Emile, does that mean TACDD for TotalDAC owners, since we we wouldn’t need the digital to analog conversion on the PCIe card?

Yes, we are now 100% sure we are going to take the modular route. That makes it a TACD mainboard with either a D/A or D/D daughterboard.

This is also the most flexible, the TACD (interface card) will be the priciest part but it can be outfitted with a variety of cheaper daughterboards containing either a DAC with analogue outputs or digital interfaces like (dual) AES/EBU or MSB Pro ISL. This allows you to retain the value of your investment in the interface card if you change DACs by simply purchasing the optimal daughterboard to interface with your new DAC.

Additionally to this we are creating a new transport/interconnect system which will need a daughterboard to be mounted inside the DAC to work.

Please note that a large part of the performance here is in the software/drivers/firmware. Hence it will most likely produce superior results when used with the optimal interface for the DAC of choice, which we now are pretty sure of being PRO ISL for MSB, AES/EBU for Totaldac, and dual AES/EBU for DCS. The optimal interface for Aries Cerat is yet to be determined. All the software enhancements we have developed to improve USB performance will now be available for these methods of connection including some XDMS technology improving general streaming/file playback.

There is still some work that needs to be done, we're most likely looking at a phased rollout between May and August.

 

[totaldac]

I have the impression that if you use the new Taiko card with AES/EBU output, you will get an improvement over the current AES/EBU output, but you will be losing for a part of the "invention". If it wasn't like that and the sound quality was the same, Emile would have been wasting his time and money developing the new interface, and I get the impression that he doesn't usually do this.

A good AES-EBU source has almost no limit with a Totaldac DAC, see this experiments I am doing with an analog master tape custom original (first) recording and a good ADC converter in the middle of this page https://www.whatsbestforum.com/threads/totaldac-new-dacs.36082/page-3

I am not aware of any server providing a similar of better AES-EBU signal, so if a new card + software would allow to provide a truly good AES-EBU signal it would a revolution.

 

[Zeotrope]

That would sound logical indeed. But I have not made any comparisons with different hardware using error correction, so I cannot confirm from personal experience. All I can say is that in the tests that I did when not using error correction, comparing different ripper/computer hardware and different ripping software would result in a different-sounding rip. For instance, a rip made in iTunes on a Macbook sounded noticeably leaner than one made on a large Windows tower PC. But I also noticed that ripping speed (where this could be adjusted) had a large effect on the sound. Faster rips would also sound leaner. At that time, I eventually settled on 8X. At that time, I preferred not using error correction as I felt it made the music sound overly controlled and "digital", compared to the same CD played on what I think would have been a Levinson 390S. In hindsight, I was compensating for the below-par server hardware that I used at that time. Those were just very different times. But as my system developed and my preferences changed, many years ago, I started noticing that other people's rips sounded better than mine and I decided to switch to a secure ripping method which from that moment on yielded very consistent results. These rips sound tight and articulate (which I would not have liked back in the day), but with incomparably better server/streamer hardware, the sound is no longer "digital" or over-controlled.

Hi Christiaan,
Thank you. Now, what compression level would you suggest for FLAC? Is there a benefit to ripping a CD using "Lossless Uncompressed" FLAC in dBpoweramp, or should the default Lossless level 5 be used?

I guess there will be less processing time for the Extreme is the file is stored uncompressed?

 

[SwissTom]

Thank you. Now, what compression level would you suggest for FLAC? Is there a benefit to ripping a CD using "Lossless Uncompressed" FLAC in dBpoweramp, or should the default Lossless level 5 be used?

I guess there will be less processing time for the Extreme is the file is stored uncompressed?

The following might help :

 

[Zeotrope]

Thanks, interesting!
So from this we can deduce that the CPU load for Decoding is roughly the same for FLAC compression levels >3.
And I/O load also is about the same for FLAC compression levels >3.

Now the question is will there be an audible sound quality benefit from using No FLAC compression (i.e., level 8)? I'm going to say "yes", or since CPU and I/O load is essentially the same, if you have the HD space, use level 8.
Thoughts?

 

[oldmustang]

Thanks, interesting!
So from this we can deduce that the CPU load for Decoding is roughly the same for FLAC compression levels >3.
And I/O load also is about the same for FLAC compression levels >3.

Now the question is will there be an audible sound quality benefit from using No FLAC compression (i.e., level 8)? I'm going to say "yes", or since CPU and I/O load is essentially the same, if you have the HD space, use level 8.
Thoughts?

No compression is level 0. Most compression = smallest file size is level 8.

Since one of the ways Taiko strives to reach its goal of best sound quality is to minimize CPU load I'd guess that not using compression would be in line with that aim.

I personally don't use FLAC where I can avoid it, in favor of WAV or AIF. Drive space is relatively cheap compared to some of the things we buy for our systems.

Steve Z

 

[Zeotrope]

Thanks, I'll try AIFF instead of FLAC, with I suppose 16bit depth for CD.

Does the Extreme play AIFF? I thought it may not since AIFF is an Apple standard...

At least with dBpoweramp, FLAC uncompressed is Level 9, or no level, to be precise