I appreciate your reply. However, about 10 months ago the Taiko announcement for TACDA/DD said the following:
"Sure, we have developed a USB card with noise levels at or below those of an AES/SPDIF/I2S interface without the jitter/timing sensitivities of those interfaces. But still there were deviations and discrepancies from expected performance. It took designing our own USB driver, querying all DACs our customers support, and designing our driver to be adjustable to... different implementations of USB receivers in DACs."
Can you kindly expand on your reasoning as to why XDMI + AES/SPDIF is a significant performance improvement from the USB output you took painstaking efforts to be superior not too long ago? Are the jitter/timing issues you referred to for AES/SPDIF no longer relevant?
I'll happily expand on that, thank you for asking. I fully understand this release may come as a surprise to many, as how this turned out came as a surprise to ourselves as well.
We did not start out designing "TACDA/DD" as a replacement for USB. It was meant to be a relative cost effective analogue output option. The vast majority of people actively participating on this forum are what you could consider "hardcore" audiophiles. They don't shy away from purchasing multiple components and build a system taking up considerable shelf space and/or real estate. We did not start out designing "TACDA/DD" for that segment of our customer base.
There is another fast growing segment of people interested in high quality audio, these people are willing to purchase expensive equipment to enhance their music playback experience. However they are not willing to sacrifice a lot of space to this end and compromise their interior designs in the process. These people tend to use active loudspeakers, but a much larger part use headphones, high-end headphones is a fast growing market. This is also the segment of people of an average younger age, you could call them the next generation of audiophiles, as the average age of "traditional" large system owners is now near or at retirement age. They are interested in innovation and technologically advanced gadgets. Although some of them are interested in what they consider to be vintage playback media, vinyl, the vast majority use online streaming services. As our goal is to build a durable lasting business in high-end audio it is my opinion that we should not ignore but fully embrace this segment, as it will likely be your main source of revenue say 10+ years down the line. As innovating is my passion, my ultimate motivator, I intent to continue doing it as long as I'm physically able to, I'm hoping to have a good 20-30 years of this exciting and fulfilling life ahead of me, hence future proofing our business is one of the highest priorities I have.
I'm getting of track a bit here, back to "TACDA/DD":
So we started designing "TACDA/DD" with the following goals:
1) to be able to provide a one box streaming solution with analogue outputs, aka a build in DAC
2) to have a better performing AES/EBU + SPDIF solution as what we were offering back was based on off the shelf, ancient PCI to AES/EBU / SPDIF / I2S controllers. Our approach to USB was so much better then this that we ended up recommending the use of an external USB to AES/EBU or SPDIF converter rather then using direct AES/EBU or SPDIF out. It as always difficult to explain this to a customer, why would an external USB to AES/EBU - SPDIF converter be superior over a direct AES/EBU - SPDIF solution? Well ancient chips, ancient generic drivers, not specifically designed with audiophile performance in mind. Ok but why didn't you design something yourself? Because we're not a company of the scale of for example Texas instruments who can design and manufacture microprocessors, which they will in fact only do if they can sell millions of them as there's no ROI otherwise. You can see this is not a particularly inspiring conversation to have to spend your time on on a regular basis.
We had the first prototype of "TACDA/DD", a single PCB with both analogue and SPDIF out, with a rudimentary driver stack, up and running just over a year ago. Initially it was a bit of an underwhelming experience, it was not very transparent, veiled even, lacking in colour, a bit "grey", grainy and not very exciting in the dynamics department. I remember thinking oh dear, this is not going to be just a matter of burn-in, but let's keep it running and see what happens. Over the next few days that did not change in any dramatic way, some of the usual improvements but the overall signature remained. However I started to notice a different aspect, being one of an unfatiguing presentation. It rendered music in a very "continuous" way, with prolonged decays, and it did cast a huge soundstage although the images within that soundstage were somewhat flat and 2-dimensional. Switching back to USB and external DAC was infinitely more spectacular, really a different overall performance level, yet lacked this intriguing flow, continuity and immersive soundstage rendering. Over time I caught myself preferring to listen to this different presentation of the "TACDA", there was something inherently "musical" and engaging in it's presentation, making me more forgiving to it's obvious weaknesses. We had more people over listening to it and they unanimously shared my opinion after pointing out what to listen for. The verdict, you may have accidentally discovered something here, this has serious potential, but it needs work, a LOT of work.
At this point we decided to scale up the design effort, perform a full scale in depth analysis on all parts and segments to identify weaknesses, performed almost a complete redesign, a truly large scale effort headed by a project manager and a team of engineers, analysts and software designers, all externally sourced. Over the course of the next few months adjustments to the prototype circuitry and software increased it's performance to the point where I felt confident in claiming this could be considered an alternative to external DACs retailing at up to 15-20K. By May we had a complete redesign and new PCBs ordered, designed to run off battery power matching it's inherently low noise nature. These arrived in August. We have been adjusting and finetuning those again over the past 3 months, performance has once again increased dramatically and the 3rd iteration, which we consider to be the final product, is now scheduled to arrive in January in quantity.
In parallel we also launched a project to widen possible applications, we designed an additional interface enabling moving "TACDA/DD" to an external chassis without a performance (sonic) penalty which ended up providing an even larger uptick in performance and the possibility to be able to apply it to the Extreme while at the same time providing the Extreme with the option to utilise battery power without the need of returning it to base for a refit.
We have now renamed "TACDA/DD" to XDMI after incorporating some knowledge gained from the XDMS development path and it's now bearing some similarities in it's approach to music streaming.
As to specifically address your question: "Are the jitter/timing issues you referred to for AES/SPDIF no longer relevant?"
Yes they are relevant, I have referred to this before as "driving with the handbrake engaged", but it still takes advantage from the XDMI architecture preceding it.
We have more then met the design goals we originally initiated this project for:
1) Direct analogue out -> check
2) Good performing direct AES/EBU - SPDIF out -> check
What the new modular architecture allows beyond the original design goals is:
3) Integrate XDMI natively in an external device like a DAC, though it's now quite substantial in physical size
4) Allows designing multiple output options to directly interface to DAC manufacturer proprietary interfaces, like MSB Pro ISL for a well known example
5) Allows designing a direct headphone output (with volume control), and in extension of that perhaps even preamp like functionality
6) Even allows designing input options for a high quality 1 stop digital hub/source
7) for those who like the direct analogue output functionality we can add things like balanced outputs, or perhaps different DAC options, etc.
Therefor our main future hardware development efforts will be centered on expanding XDMI functionality.
