XMOS-based Asynchronous USB to I2S interface

well its pretty bandwidth limited, so wont be possible to do DSD over USB as this needs 176.xx; clocking regime (Master/Slave) seems not very flexible and the guys in the widget thread arent really giving it the seal of approval for stability. so i would say there is at least a few things lacking

windows has had 3 major OS overhauls and 2 completely new releases since UAC2's conception and it still hasnt been added, this speaks of something more than a missing update
 
Lorien,
in a direct comparation with M2Tech and Open source your XMOS modules is the winner.

http://www.diyaudio.com/forums/digi...2s-m2tech-vs-xmos-vs-open-source-qnktc-2.html

I hope it's not a bit late guys... but I want to thank you all, especially audiodesign for having time and will to make that test. Frankly I'm a little bit surprised about WaveIO's performance since my goal was / is simply to improve the XMOS USB reference designs. Now, how far my efforts went - only you guys could tell... Anyway, I'm pleased to see positive feedbacks, it means that WaveIO card is doing a good job.

Lucian we've been begging for a while for xmos based multichannel usb-i2s, what can you do for us? :D

For now I'm focused on improving WaveIO's outputs. We will see what will happen next.

@ Carsten.Witt: Glad you have it. Please let us know about your impressions on SQ!

Best wishes to all,
L
 
I hope it's not a bit late guys... but I want to thank you all, especially audiodesign for having time and will to make that test. Frankly I'm a little bit surprised about WaveIO's performance since my goal was / is simply to improve the XMOS USB reference designs. Now, how far my efforts went - only you guys could tell... Anyway, I'm pleased to see positive feedbacks, it means that WaveIO card is doing a good job.



For now I'm focused on improving WaveIO's outputs. We will see what will happen next.

@ Carsten.Witt: Glad you have it. Please let us know about your impressions on SQ!

Best wishes to all,
L


Hi Lorien,

How can I buy one of your WaveIO boards?
 
well its pretty bandwidth limited, so wont be possible to do DSD over USB as this needs 176.xx; clocking regime (Master/Slave) seems not very flexible and the guys in the widget thread arent really giving it the seal of approval for stability. so i would say there is at least a few things lacking

Stereo PCM @176.4 requires a very small fraction of available USB2 bandwidth -- what in DSD requires so much more? Stability is a function of DPC latency and is potentially a function of every other driver in the system. Wolfsin searched for evidence of thesycon instability given that his system works flawlessly.

Can you provide pointers? Admittedly this is academic in my case given my modest needs but I fear others will take your remarks seriously. Are you running your WaveIO from Mac or Linux.
 
it doesnt have much to do with USB bandwidth, maybe i was reading an old website last night, but it intimated it (theyscon) was limited to 96khz, so obviously that presents a problem with 176. i found it by a link last night, but just now i googled it and hit a totally different page. It still doesnt seem to be as easy as downloading a driver though, hardware has to be sold with it or license the driver, so i cant just use the driver with an existing UAC2 device or at least they dont frame it that way?? doesnt seem very UAC2, which is 'driverless' on Mac. it does seem like it would work fine if its reliable, but the only stuff prominently placed were the development kits and demos

i'd be wrapped if it works well, it would mean i can run Win on my Mac mini and run allocator natively with my USB hardware, which currently is only UAC2 and mac only and allocator is PC only so i'm stuck, i can run win on my mac but then the hardware is disabled; on the other hand i can run the hardware on mac but cant run allocator; other comparable digi XO software is either VERY expensive, or not as good.

the only reports ive seen that have tested it extensively have been in the widget thread and it didnt end well, they ended by saying it wasnt ready for primetime

anyway we're OT, i'll shoot you a PM tomorrow
 
Last edited:
no smileyz so not sure whether you are joking. There iz USB 1.0, 2.0, and 3.0 -- getting faster but with backward compat. Where izTheLine?

@rsdio I am as skeptical of instabilities in the thesycon driver as you appear to be of the glorious 'impressions' reported recently in this thread. I think WaveIO is a very fine product, skillfully implemented, reasonably priced, whose image might be tarnished by OS bigotry and innuendo. Just as 'difficult' understates the impossibility of improving on bit perfect, effusive praise or unsubstantiated rumor potentially undermine efforts of one-man shows. I am in Lorien's corner all the way.
 
no smileyz so not sure whether you are joking. There iz USB 1.0, 2.0, and 3.0 -- getting faster but with backward compat. Where izTheLine?
I assume this question was for me? I was not joking. All new USB devices are USB2. Thus, saying "USB2" really says nothing about speed. I just released a new USB2 device that is Full Speed, and would only be capable of medium bandwidth. USB2 is not faster than USB1 in all cases. However, High Speed is only available with USB2 or newer.

I mention this because a brand new USB audio device and firmware could easily be USB2 and UAC1 at the same time, or USB2 and UAC2 but only Full Speed. Considering all the variables, I was genuinely recommending that folks say "Full Speed" when they talk about the faster speeds that were added with USB2.

Note that the USB folks practically force all new USB devices to be USB2, whether they're High Speed, Full Speed, or even just Low Speed. I think they made a mistake by confusing the terminology - it's certainly not your fault - but I wanted to at least make a suggestion that would allow folks to be precise in their discussions.
 
thesycon? yeah its unfortunate thats the best windows has to offer, its soundly beaten for free in Linux or Mac. why it has taken 3 years to still not have a reliable UAC2 driver for win i have no idea.

A free & Open Source UAC2 ASIO driver have been recently developed for this project: "Open-source USB interface: Audio Widget".

I guess it may be easily adapted for other UAC2 devices, too.
 
Last edited:
no smileyz so not sure whether you are joking. There iz USB 1.0, 2.0, and 3.0 -- getting faster but with backward compat. Where izTheLine?

@rsdio I am as skeptical of instabilities in the thesycon driver as you appear to be of the glorious 'impressions' reported recently in this thread. I think WaveIO is a very fine product, skillfully implemented, reasonably priced, whose image might be tarnished by OS bigotry and innuendo. Just as 'difficult' understates the impossibility of improving on bit perfect, effusive praise or unsubstantiated rumor potentially undermine efforts of one-man shows. I am in Lorien's corner all the way.

again you seem to be confused about USB data bandwidth standards vs. the USB Audio Codec 2.0 standard aka UAC2 (there most certainly ISNT such a thing as UAC3 even though USB 3.0 exists) but as i said it seems i followed an old link to a theyscon old website last night that had recent (late 2011 info) and thus took it for up to date info. that doesnt impact on your confusion, look it up

there is thunderbolt/lightpeak too but you dont see audio using that much bandwidth... I wonder why you are skeptical about the stability of something that is only literally months old??.... i thought you had worked in the industry.

like i said i'm happy, but i'm not sold on it just yet enough to swap a fully functional and mature implementation of the UAC2 codec on Mac OS 10.7, for a substandard OS that is the devils work (lol) just so i can have a better XO, i have to try out all the demos and it wasnt immediately obvious where i can just DL the driver no strings attached to try out? do you have a link?

what i might do is buy a used i5 windoze machine to run a dedicated XO and slave it over the network from my mac

anyway again OT

back to the thread i've stretched a tenuous link to the thread too far; dont encourage me
 
Last edited:
A free & Open Source UAC2 ASIO driver have been recently developed for this project: "Open-source USB interface: Audio Widget".

I guess it may be easily adapted for other UAC2 devices, too.

yeah i know, but it's reports in that thread that havent been all that encouraging, with most people trying but going back to the linux or mac, also it seems to be tied to firmware rather than being a standard that is independent. my application is very high bandwidth xmos based (4 x 32bit/384kHz or 8 x 32bit/192kHz).

it does seem to be making progress but doesnt seem ready for primetime and doesnt seem to be a standard really ie. its 'UAC2' like, needing special FW rather than just working with UAC2 hardware that already works on mac and win.

@Lorien, which have you been finding most reliable as far as win goes for your Xmos solution?
 
it does seem to be making progress but doesnt seem ready for primetime and doesnt seem to be a standard really ie. its 'UAC2' like, needing special FW rather than just working with UAC2 hardware that already works on mac and win.
well, it's a young project: some problems/limitations are expectable at this very early stage of development. Moreover, I guess that the "audio widget" is the only UAC2 interface the developer have currently access to for testing.

But it's FLOSS, thus everyone is welcome to contribute to development. ;)
 
@ UnixMan: With all the respect I want to remember you that THIS thread is called: "XMOS-based Asynchronous USB to I2S interface". Is something in your last statement that has anything to do with the XMOS chip itself or the XMOS reference design... the base of WaveIO card? I do not recall on the other thread to mention even a single word about WaveIO. Most of us might say that your statement is a simple one but few could dig further and see besides that few attempts to promote other products. Please stay on topic... as I do and thank you for understanding!

@ qusp: I saw your last statement in my email and your choice to switch to PM! It's much appreciated! Regarding your XMOS - based application (4 x 32bit/384kHz or 8 x 32bit/192kHz) you make at least curious ... I'm interested in the 32bit option of your setup. In my experience, there's no such a "free meal" ... at least not with Thesycon :) And if we speak about Windodsdsdsd ... we are on the same side but if I have to choose from Windoze and Thesycon on one side and None on the other I would choose the first option... even if, sometimes I wish to be a third path :)

Kind regards,
L
 
Speaking about 32 bit depth. Do you know if there's an USB product that works with Thesycon drivers and offer 32 bit feature because I found none so far...
Well, it's not USB or Thesycon, but the Metric Halo Labs MobileIO line of audio interfaces offer 18 channels of 32-bit audio over FireWire, with additional bandwidth for parameters and metering. This is all supported transparently within standard CoreAudio applications under Mac OS X.

A similar USB interface could work without changing any application code, provided that the USB Device implemented UAC standards. Unfortunately, I still have not obtained any of the multichannel USB Devices that my clients use. I tend to write the application code using FireWire devices for testing, and it all runs just fine when the customer uses the same code with USB devices. The only thing missing is that I haven't put my hands on the USB devices with more than 6 channels of output.