|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
184.147.232.184
In Reply to: RE: cPlay - the open source high-end audio player using ASIO posted by cics on May 05, 2008 at 12:31:58
Bucking conventional wisdom, I gone with USB rather than the Julie etc sound card route. Recently switched to USB2 Asyncronous with really good results. Asyncronous protocol puts makes the PC a slave to deliver the bits to the converter. For sound, this is the way it always should have been. I chose WaveIO, XMOS based card. Beautiful design and execution. http://www.diyaudio.com/forums/digital-source/188902-xmos-based-asynchronous-usb-i2s-interface.html It also brings up to 24/192 resolution.
To get the drivers loaded, I needed a clean install. This card completely removed the USB conversion as the weak link from the system. It made the recent sound quality improvements outlined by our inmates via XP slimming even more apparent.
Starting with a base XP Pro Nlite install and minlogon + CMP optimisations the improvement from the WaveIO card was apparent.
I did the Steppe 1- 15 + Mihaylov tweeks. At this stage, sound was very clean and crystal clear with great sound stage etc. But it was not something I wanted to listen to. Compared to my well modded Rotel CDP I far preferred the Rotel. Simply more involving and natural.
I thought the WaveIO card might be the issue. However, for completeness I carried on first doing the Resource Hacker slimming of the remaining DLLs in Sytem32. Unbelieveable!! Now this rig exploded with natural texture and toe tapping involvement. Then I added all the registry slimming. Again a huge jump into that 'I am there' and involved with the performers who are real breathing humans. Instruments are made of wood, brass and other natural elements.. not some digital reproduction.
Now this thing is truely high end. Now no comparison with the Rotel CDP. It is useless by comparison.
Summary. USB2 is a viable alternative. If you have not ventured into the ultra level of resource hacking and registry hacking, you have no idea what is possible in PC based sound. I am no where at the level the core team has achieved. I think I am hovering around 23mb in Windows with a 600k registry. Even without the extremes the CMP based system can beat anything out there.
Thanks again to the trail blazers that have made this possible. Now its time for me to get back to building my OTL 6c33c monoblocks and see where CMP2 can take me.
Follow Ups:
There is a comparison of USB2 boards taking place. WaveIO wins first round,
http://www.diyaudio.com/forums/digital-source/208834-usb-i2s-m2tech-vs-xmos-vs-open-source-qnktc.html
Too bad they don't have an ultra slim CMP2 rig!
Hi wlowes,
Very interested in the XMOS card, Can you give more detail in your implementation? e.g.
-- do you need a power supply for the card? or
-- do you just plug it in the USB port at one end and the dac at the other?
-- does the ASIO driver bundles with the card? or just the demo version (with the beep)?
Thanks
KC
KC
You don't need an external PS. There is a jumper that sets power from USB or from external. On board handling is very good. In my system external was an improvement, and for the ultra in transperancy, you want that supply to be clean. With USB power it was still a big step up in SQ over a very good pcm2706 implementation it replaced.
You plug in USB at one end. I2S at the other. 2 choices for I2S. mBNC cables for non isolated signal, or a multi pin header for an isolated signal. USB also has choices. On board USB socket, or a header to allow you to run USB to a socket on your case away from the bd. Allows you to place the board close to the DAC. Your main challenge is to ensure you can connect I2S on your DAC. I just solder to the pins on the tda1541a.
The builder has customized the Theyscon driver. It is fully licensed without beeps. I paid as much just for the AQVOX asio driver on my last rig.
This thing is a steal. Allow 6 weeks for delivery as this is another DIYer with a day job, and he is a perfectionist.
Recently i stepped over from Juli@ to USB-async using a modded M2tech Hiface. This unit gets 5v from USB-bus and 3,3v from Lifepo4 batteries.
CMP2 setup: nLited small install of XP-pro, minlogon.
Hiface beats Juli@ in this setup, and equals SQ of my Marantz CD63 cdplayer.
Don't have time (yet) to slim xp, but your results make me eager.
I'm very curious about your results if you would try on a Unix-based SO. As i read Voyage-MPD and MPDpup produce great SQ.
http://linux.voyage.hk/voyage-mpd
http://208.109.22.214/puppy/viewtopic.php?t=70052&sid=a211e2cab717aebc13cd5c51416918e9
If SQ on MPDpup is equal/better than CMP2 it could be a great experiment to run it on a very small ARM-based unit like Raspberry Pi (700MHz ARM), or (not so small but very tweakable) Alix-boards (500MHz AMD-Geode). And for me it could be interesting to change to XMOS (Hiface doesn't come with linux-driver)
Please keep us informed! tia!
Douwe
Douwe
It was in fact the Voyage Store that caught my eye. It strikes me that a small ALIX board acting as a headless network appliance serving up the bits is an interesting proposition. Also very simple to build a single 7V linear power supply and it is completely isolated from HDD. Will be a while before I take this step but will post results when/if I get to it.
Walter
There is a new HiFace Two with ASIO and Linux support that will be out soon.
You can find info at their site...
Hi wlowes,Congratulations to your achievement.
Is the WaveI/O card using ASIO ?
- I have the feeling you´re tricking the system to run async via WASAPI...(This will compromize the dedicated ASIO/buffer design of cPlay)
And to an extent, explain that your interface protocol gains quality from the software mods on Windows/cMP side.To run cMP2 with async ASIO protocol, it will always require an ASIO driver software to use the full benefits of bulk mode data transfer in an external DAC.
It is my experience that true external (async or iscochronous)clocking via ASIO does (almost) completly eliminate efforts on the Windows OS.
-so if you experience improvements, my guess is that some sort of jitter is allowed to influence the DAC via the protocol on the WaveI/O.
In other words : the protocol you´re running is the "weak" point, and could possibly gain even more quality if it could run ASIO.kind regards
cMP2 Computer w/digital crossovers & FIR filters > Lynx Aurora 8 FireWire /192kHz all channels. 2x AcousticReality Ref. 202 & 2x AcousticReality Ref. 601´s ICEpower. Magnepan MG3.3R beechwood frames & custom stands. Miller chokes
Edits: 03/12/12
I was kind of expecting async to be immune to our tuning efforts. Not so in my case. WaveIO uses ASIO 2.1 drivers from Theyscon. per the author's site.. "Since this board is entirely based on XMOS / Thesycon work (SW speaking) you can see details directly on XMOS web site: "USB Audio Class 2.0 is natively supported by Apple OS X version 10.6.3 and above. Support for Windows is provided via XMOS partners, Thesycon and Centrance. The drivers are configurable and provide support for WDM/Direct X and ASIO 2.1."
Cplay sees it as an ASIO driver. There is no trace of Direct X left in my slimmed XP.
I am toying with the notion of trying running it from a unix appliance as an experiment, for even slimmer hdw/softw than we achieve with CMP2.
WaveI/O card uses ASIO...
As it output I2S I would be interested to know which DAC you connected it to.
Bibo01,
My Dac is a DIY TDA1541a based DAC with a 6n2p tube output stage. I started with the audio board from an Arcam Alpha CDP. There is not much left from Arcam but the box. Power supply is all BG and Oscons with Burson regs. Tubes have choke supply on B+. DEM reclocking per Grundig design. Separate Plitron transformers for DAC and tubes. I use mBNC connectors for the I2S signal. Not perfect, but seems to pair well with WaveIO.
Dear Bibo01
hmmmm....see, ASIO is a software protocol, and I2S is not a protocol at all, ey ?
One can start wondering why cics implemented tiny, small, medium & large communication buffers, and drew so much attention to buffersizes, as in theory, when it comes to pure playback, long buffers should be as good as small ones.....
I´m not really convinced that just because USB is async and derives its clocking from the DAC, that everything is as straightforward simple as that....
I always come to think of how much difference a new Lynx WDM driver can improve sound quality and how much time & effort cics put into cPlay´s ASIO/DSP optimization.....
I seriously believe that driver software is the most underestimated part of any digital enginering and that it is too easy just claim that async is the blessing of the holy grail.
kind regards
cMP2 Computer w/digital crossovers & FIR filters > Lynx Aurora 8 FireWire /192kHz all channels. 2x AcousticReality Ref. 202 & 2x AcousticReality Ref. 601´s ICEpower. Magnepan MG3.3R beechwood frames & custom stands. Miller chokes
ASIO is a software protocol and I2S is not a protocol at all, ey ?Correct. ASIO is a soundcard driver protocol; I2S is an interface standard, not a protocol.
The Thesycon drivers used by the Wave I/O device and other USB-based ASIO drivers (e.g. the Ploytek/AQVOX driver) bypass the Windows USB Audio drivers to transfer data straight to a device (using the PCM270x, that XMOS something-or-other or the TAS 1020) which in turn pass the data over the I2S bus to the DAC. They do so asynchronously (Thesycon) or (if I have the term right) adaptively (Ploytek).
This ground has been covered before - see link and the replies, esp, as ever, the one from John Swenson.
I seriously believe that driver software is the most underestimated part of any digital enginering
I'm sure you're right but reports are that the Thesycon drivers used for Wave I-O device are particularly robust. IIRC, Thesyon supplies drivers for some of the much-respected Ayre products. Its people do know what they're doing.
it is too easy just claim that async is the blessing of the holy grail.
Also correct only I'm not sure that anyone does claim it's a Holy Grail. Again, reports have it that the protocols work well in many implementations - including cMP^2/cPlay. If you take a few moments to visit the thread wlowes linked to, you'll see that a number of cMP2 users are already successfully using the Wave I/O device.
. . . the protocol you´re running is the "weak" point, and could possibly gain even more quality if it could run ASIO.
It doesn't "run" ASIO, it conforms to the ASIO standard.
Edits: 03/12/12
Hi Dave,
Pardon me, but I did read the two links you´re refering to, but I´m still not convinced that it´s irrelevant how cPlay connects to the DAC.
It´s however not a question of that an async USB DAC can work with cPlay, but rather how.....
Apart that I´m very sceptical to how well all these newish async DACs actually "eliminate" the (jitter-) influence from the computer.
The theoretical thesis is that when cPlay (i.e. the PC) is slaved to the DACs masterclock, jitter becomes irrelevant, right ?
So how can wlowes experience a difference then ?
-is it only ground-noise that makes the difference ?
According to cics (and many others) it is jitter that alters soundquality (bit-perfect transfer granted), so it does not make sense that there should be any perceptible SQ difference.
When I first switched from internal Lynx Two-B (fully optimized cMP2)to external Lynx Aurora via isochronous FireWire, I was somewhat surprised that switching from cMP2 to regular XP, did hardly make any difference at all....I did expect at least a little.
Second, I´ve witnessed several software upgrades on the Lynx drivers and a lot of us have marvelled about cics work on cPlay´s DSP and ASIO software.....
My point is that software is an essential part of a devices performance and most manufacturers have a very "casual" attitude to this.....
Surely, it´s cheaper, easier, faster and less complicated to licence al sorts of "asynchronous USB" circuits and advertize it as "immune to jitter".....
Infact I have the impression that manufacurers just pick´n mix design features from a catalogue. And where this is perfectly acceptable in order to implement well-enginered technology at a price-point, I feel it´s problematic to claim that only because it´s "async" it eliminates all jitter issues.
-it seems that there are ways of "async" that do not capture the benefits entirely....
kind regards
cMP2 Computer w/digital crossovers & FIR filters > Lynx Aurora 8 FireWire /192kHz all channels. 2x AcousticReality Ref. 202 & 2x AcousticReality Ref. 601´s ICEpower. Magnepan MG3.3R beechwood frames & custom stands. Miller chokes
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: