XXHighEnd

So we have learnt that USB sucks. Because it has a "speed limit" of 16/48. That is USB in its current version 2. Maybe the next version 3 will improve on this? Time will show. But then it will be another plug format, and probably not compatible with present USB equipment. A friend told me that Apogee DAC's DO accept 24/96 through USB, due to custom made in-house drivers. If there are any software pro's reading this topic, maybe they can shed some more light on the situation.

In the year 2008, if I wanna play hi-rez music, I have to go back the SPDIF interface using RCA plug format (made by the Radio Corporation of America in the early 1940s). Qiuite ironic, isn't it?

My PC (located in the basement) has Firewire 400 output. I will run a 5meter Firewire cable up to my stereo in the living room. Then I will connect it to a Fireface 800 box (purchase price about €1,200.-). The Fireface 800 box will be positioned in my equipment rack next to the Eximus DAC. Then I can run a 50cm short 75ohm coax RCA cable into the DAC. With this setup I can get my 24/96 through all the way to my DAC.

SOME NOTES:
I will experiment with high quality glassfiber optical Toslink between Fireface and Eximus. A galvanic separation can imporve the sound itself, due to reduced HF noice, mains ground loop, etc. If the Toslink interface works well, then the RCA connector will be donnated to the local museum.

€1,200.- for the Fireface 800 is a lot of money. The box has many more features which I never will use. (Build in A/D, D/A converters of medium quality, headphone outlet, digital volume, instrument inputs, etc). BUT: It also has a superb re-clocking circuit, which provides a State Of The Art low-jitter output of the SPDIF signal in both Toslink and 75ohm RCA format. Also it's not bad to have same configuration as Peter himself, when playing through XX.

There are cheaper "boxes" out there who can convert your PC signal into SPDIF for less money. The question is, if they have as low jitter performace as Fireface 800?

LynxAES16 (estimated price €900.-) is another option. It is also said to have high quality low jitter circuit. It delivers a SPDIF signal, but in the AES 110ohm XLR plug format. Unfortunately my Eximus doesn't accept this plugformat. Of course you can use a simple mechanical plugg adapter, but I'm afraid it would not be optimal.

Any comments, Gentlemen?

a. The Fireface obviously comes in the 400 version for 800 euro or so.
b. Reclocking of the Fireface is at some 800ps. It is not *that* good (20ps would be :)).
c. I am not sure what the true merits of b. are, because there's many parts in the chain, and e.g. a 100ps USB DAC could be destroyed by USB itself, or the fact that it won't connect with 96/24 for that matter. Just examples.
d. Now you have an XLR connection, and again there are compaints ? hehe, anyway keep in mind that the "mechanical" converter plug as you mention it, is not only a plug. A "working" one would have a resistor (or two, I'm not sure anymore) as well.
e. My current configuration won't last long anymore. Don't let it be an argument please.
f. If you are going to invest in a "reclocker", I think there are other options, which by pure coincidence (:nea:) connect by means of SPDIF. I didn't read this myself : http://www.teac.com/esoteric/G25UFinal6moonscomReview.pdf

Yes, I think the 400 is a better candidate. Less (unnecessary instrument) features and better price. It also includes BNC input of external word clock. (see next point)

I have spend some hours scanning the world market for digital audio interface candidates. There is a very big choice from inexpensive soundcards costing peanuts, to pro equipment with lots of unnecessary (for my, that is) features costing thousands of euros. None of the very cheap ones mention any active jitter reduction circuits. Fireface 400 is actually one of the few who really discuss it's jitter reduction system in detail and who reprints jitter measurements. you say 800ps is not "that" good. But what is then?

To answer my own question, I found the Apogee Big Ben costing 1,500USD. It's actually a quite universal product, like a Swiss army knife, regarding my needs. Apogee does not mention any specific measurements, but they "claim" it's very good (low). But is it below 800ps?

-------------

Some people think the digital interface is solved with inexpensive soundcards. I think they are wrong. Let's say you install a soundcard in your PC, which outputs a AES/EBU 110ohm digital signal (squarewave signals up to 5volts). That is a VERY tough task for the powersupply. -Which happens to be the shared "Mickey Mouse" powersupply of your PC. -Which is not only very weak, but also very poluted with RF noise. I believed that such a digital signal will be very "jittery" when it arrives to the DAC. (Although I have not seen measurements of such a case, I have discussed the topic with a High End DAC designer and manufacturer).

So, probably, it's better to take the signal from the PC by Firewire into a separate box like the Fireface 400 with its own powersupply/onboard local power regulators. Then feed the Fireface 400 with good clean power (including ferrits on mains and other usual High-End tricks) and let it generate the spdif signal, going to the DAC. Eventually the user can choose between spdif RCA, spdif Toslink or AES/XLR depending on his DAC input.

I believe (hope!) this will be a "very good" solution at a not "so high" cost.

Eventually, an external clock (like Apogee, Esoteric or others) can be added to the Fireface 400's BNC input for external word clock. (I think the future will bring more choices of separate word clock to be purchased separately). Also future stand alone DAC's will have separate in/out BNC clock options.

To sum up: "Good" SQ can be made with inexpensive soundcards/interface, but I think that further margins can be gained when doing as discussed above.

I do not have any particular technical insight into this technology. But I like to get to "the bottom" of all aspects of my hobby. -So, any views on the topic are appreciated. Hopefully we can reach some absolute conclutions. At least I will report back later on after purchase and testing.

 :)


PS: Some say a 75ohm RCA coax will perform better if it's LONGER than a meter, because the additional length will attenuate the reflections inside the cable.

I don't think the wordclock connection should be used to reclock better. Especially not when it is not the DAC. The wordclock connection is to synchonize (pro) stuff, behind that is again a DAC with good or bad jitter behaviour.

It's all apples and oranges, and yes, I came to such a configuration because I thought it was the best like that.

It would be my conclusion that the pro world stuff certainly is NOT the best (but it has its purposes), that the cheap stuff in 100% of cases also isn't right, and that if you buy from each part of that chain the best for its purpose, you end up with something optimal. So that's how you snobbishly end up with a 1300 euro Fireface800 which only passes through SPDIF. But indeed, it's outside of the PC, and it has its ferrite core filter. About the reclocking ... I actually don't know. Keep in mind the "strange" situation that the life of the soundcard starts behind that Firewire cable, which doesn't seem like a normal input to me. Also, when I had it behind a 10m Firewire cable it sounded worse. Smeared in the highs. "It" is the DAC behind it, SPDIF connected.
What I myself really should do, is find the best Firewire PCI card, because *that* makes all the difference of the world.

... but I actually stopped tweaking for a long time, because that happens in software in this modern world. Hahaha. Nonsense of course, but I really stopped, because otherwise how to judge the software.

Hello again,

1.
FF400 up and running!

2.
I will compare carefully the SQ differences (if any!) between USB and FF/SPDIF (both coax and TosLink).
Eventually I will also try to get a dedicated high quality FireWire PCI card. (At the moment I use the standard FW port on the motherboard).

3.
In the meantime here are the DAC TEST results, when using FF and TosLink. A strange thing is that my DAC is quoted to be 24/192. But the DAC TEST says it supports 384kHz. How can that be?

-----------------

Start Audio Device Analysis Exclusive Mode support
046CFD24

    Supported : 2|16|44100|176400|4
Not supported : 2|24|44100|264600|6
    Supported : 2|32|44100|352800|8
    Supported : 2|16|48000|192000|4
Not supported : 2|24|48000|288000|6
    Supported : 2|32|48000|384000|8
    Supported : 2|16|88200|352800|4
Not supported : 2|24|88200|529200|6
    Supported : 2|32|88200|705600|8
    Supported : 2|16|96000|384000|4
Not supported : 2|24|96000|576000|6
    Supported : 2|32|96000|768000|8
    Supported : 2|16|176400|705600|4
Not supported : 2|24|176400|1058400|6
    Supported : 2|32|176400|1411200|8
    Supported : 2|16|192000|768000|4
Not supported : 2|24|192000|1152000|6
    Supported : 2|32|192000|1536000|8
    Supported : 2|16|352800|1411200|4
Not supported : 2|24|352800|2116800|6
    Supported : 2|32|352800|2822400|8
    Supported : 2|16|384000|1536000|4
Not supported : 2|24|384000|2304000|6
    Supported : 2|32|384000|3072000|8

He he, that's because the FF is not protected against more than it's capable of. If you try to play such a high sample rate file it goes bananas.

This just kinda leaps off the page and then hangs in the air doing a little hula dance

Cajun, momentarily. :)

I changed from USB to RME Fireface 400 interface between PC and DAC. The sound improved. Not a huge difference, but it was worth it. A slightly more transparent and "focused" sound. USB sounds (slightly) veiled in comparison. My gues is that the RME re-clocking circuit provides a spdif signal with less jitter, than the input USB receiver of my DAC is capable of.

Read more about RME claims here. (http://www.rme-audio.com/english/techinfo/steadyclock.htm) (it's sales literature of course, but shows some serious measurements).

Can the FF400 be further improved with an external super precise clock? -I don't know. If I can get my hands on such a clock I will try it. Maybe a dCS Verona Master Clock? (http://www.stereophile.com/digitalprocessors/305dcs/)

Where does it end? An Alpermann Rubidium GPS 10MhZ clock? (http://alpermann-velte.com/prod_e/erubidium/rub_gps10.html) (Does anybody know what this things costs BTW?)

I too can choose between USB and SPDIF; besides I agree with your remarks pedal, I could add that SPDIF is more snappy, USB seems to present the bass better, but the most important remark I have with USB is : the musicians don't want to present themselves. I know, a strange description, and I guess it is related to everything being less profound (and volume doesn't help hehe).

Of course this information is useless because of the USB connections being different (and in my case USB is converted to SPDIF internally), but still ...
The reclocking the Fireface does presents an overwhelming 800ps jitter which would be the world record in badness these days. But if I compare my same DAC in a (commercially) presented 0ps jitter (btw USB) version (which XXHighEnd CANNOT influence) ... that doesn't cut it. There's just much much much more going on and let's give it another year to find out. I think we are about to have the tools for it ... :whistle:

:good: Tell me more, tell me more, tell me MOOOORRRREEEE


In the meantime I can tell you that I have purchased an expensive Firewire PCI card + an upgraded external powersupply unit for my media PC. I will compare if these new components improves the SQ or not, compared with standard motherboard and standard cheap wall-wart powersupply.

Also I am curently comparing high quality coax, cheap Toslink and expensive Toslink ($1000.- AQ glass fibre) cables between FF400 and DAC. There ARE SQ differences, but I need to listen more closely before drawing a conclusion.

One more end note: The RME FF400 allows for bitperfect digital loops, making it possible to use plug-in programs like SOUNDLENSE (http://www.juicehifi.com/) in the pure digital domaine. This PC program can correct just about anything in your system: Frequency response, Phase, Impulse response, Room reflections, etc. It can also act as a crossover with unsurpased functionality (whatever number of drivers and speakers you have - Soundlense can do the XO). Initial testings are very promising.

Hahaha, but that is great ! I always thought that such a "tweak" would be possible somehow, but didn't pay the real attention.
Now think in ADAT channels instead of SPDIF (just choose them in the FF) and start connecting outputs (DACs, speakers, speaker drivers ...).

Although we could say it is too stupid not to see this option in a glance, I say you are great pedal !. So allow me to say it :
GREAT !!

:clapping:

Keep in mind though that the Fireface (at least the 800) has two SPDIF outputs only, one coax and one Toslink (or two Toslinks ?). Thus, connecting more DACs (with amps behind them) is a bit of a strange operation.
A less strange operation occurs when you use the analogue outs from the FF directly to the amps. An again more awkward situation emerges when you combine this with using one SPDIF, and use this for one of the XOver channels (though stereo).
Fact is that the FF just can do this, as long as you use ADAT (which contains 8 channels).

In either case a digital volume would be the most convenient, but assuming 16 bit output towards the DAC this should be avoided, unless the digital volume of XX is being used (and then preferably not more than -18dB).
The real problem would be the individual volume of the XOver output, which should not occur in the digital domain with 16 bit output (24 bit is ok). Don't get tricked by the bits used for internal processing; it is about the output (or IOW the input to the DACs).

Quite another problem could be (or will be ?) the lack of drive, because of no "buffer" being present. This by itself urges for the pre-amp again, which now must be multi-channel (but obviously without processing). And once the pre-amp is there anyway, the problem of the digital volume is not there anymore (but it requires the pre-amp to attenuate in the analogue domain).

So it's not all *that* easy, but the base is there. :)

True. A friend of mine (http://memweb.newsguy.com/~stigerik/body_index.html) has purchased this  8 channel DAC from RME (http://www.rme-audio.com/english/adi/adi8.htm). It's their top model, about €3,000.- retail price. He says it really sounds up there with the expensive audiophile DAC's. The clue is that from FF800 you can export 6 or 8 chanels from ADAT to the RME DAC in 24/96 rez. And then you are much closer to true digital XO and volume controle, since you can attenuate the signal digitally within the DAC, and the DAC can drive the poweramps directly. (BTW: My friend says the FF800 has enough capasity to output 8 channels of 24/96, but the FF400 hasn't).

[Aaarrrrrgggg] but I wanted to build something this summer [/Aaaarrrrggg]

hmmm  looking to at least get spdif out to dac to try 24/96, and more importantly to me use xxhe volume.  It would be great to be able to get rid of the analog crossover and use something like SOUNDLENSE (or wait for XXHE...); but, even if I build a new computer, I wonder how running software in the background will influence SQ... Seems like everything does with my little laptop
One moment I'm thinking about picking up a FFx00, the next moment I'm building a whole new computer and thinking about soundcards, ... too many choices awwwww...... I just want it to sound the bestest... is that too much to ask?

anyway, build something around that juli@ card ... maybe???

Gentlemen,


http://www.audioasylum.com/forums/pcaudio/messages/9585.htm (http://www.audioasylum.com/forums/pcaudio/messages/9585.htm)

I just found this link to an interesting thread about the (initial) shortcomings of USB. I write "initial" because the thread is from 2006, 2 years old. However most so-called USB-DAC's for sale today, still have the limitiations as discussed in by Mr Swensson. Judging from Benchmark's last DAC with custom drivers enabling 24/96, there are still areas where USB can improve. Time will show. Remember though, that of-the-shelf USB DAC's necesary doesnt show all the potentional of PC-audio (and XX, ofcourse!).

(Thanks to LMC for hinting about the thread).

There were several reasons I finally broke away from usb with my Stello, not least of which when I found differences between the 5 usb ports I had available.  In my latest pc build I have all usb shut off at the bios end.  Can't say I've done back and forth comparisons on the new pc, but now I have 96/24 capability through spdif into the Stello, and I'm betting part of the great reason my front end is so superior to before is because of not using usb.  Reading your posts gave me some nudges too.

That all said, Wavelength audio and Empirical audio would say they're doing it like no one else (they can now do 96/24 through usb.... still);
I don't think even they can get away from the fact that different usb cables and different usb ports present better/worse.

Pedal, your link isn't working ...

Then again, on the Benchmark or PS Audio DACs, it's a real to-and-fro comparison between ports: a good coax cable sounds better than a stock USB cable, but to my ears a good USB cable sounds better than both: I've heard enough to be convinced that a cheap USB cable is a significant bottleneck and that many USB implementations are judged unfairly because listeners are not factoring in the cable.

I really can't understand why everyone has gone completely gaga after USB DACs when all modern mainboards have S/PDIF out; even an onboard S/PDIF will arsewoop most USB connections SQ-wise when hooked up to a DAC.

Moreover, a sound card with a high quality toslink or coax out is dirt cheap. There is no need for USB DACs at all, and it is starting to look like a hoax of some kind. I particulary enjoy the belief that toslink is OK for Macs, but if you have a PC, you need USB :wtf:

Because the quality of that spdif out is awful. It is polluted by all components that resides on the mainboard (including the cpu) and the drivers are usually poor as well.

And where does the USB come from? From Mars?
I still enjoy my spdif motherboard output very much even if some say it's cr*p...
But I'm willing to try a PCI-e solution if I can find it!

Is this necessarily so? I vastly prefer toslink out from a modern mainboard than the average USB-connection for hooking up to a DAC. The newer Realtek codecs doesn't meddle with the signal and support all necessary sample rates.

And the toslink out of my Onkyo card is very very good, and the card is rather cheap. There is no reason whatsoever to use USB.

I´m gonna try this:
http://www.m2tech.biz/products.html
might strip it and take I2S out to the Buffalo input
have 2 on incoming :)
best
Leif

Hey Leifchristensen,

What is the price of that? 

grtz  :)

207 euro for 2 incl.freight to norway
these have rca connector
bnc is also available at a slightly higher cost
paid via paypal
best
Leif

 :thankyou:

No, the first USB receivers in DACs were also awfully bad :) And if the mainboard has a good power filtration...
You know, there is never an absolute in computer audio. On paper, we know which are the best solution. But then something may sound different, exceptionally good or bad respect what the theory says.

Toslink vs coax is a matter of isolation vs jitter. Toslink has higher jitter but it is intrinsecally isolated, while coax will require some isolation in the device (dac), which is normally MISSING.
etc etc. I think you get the point.

Interesting device and pretty cheap as well.  We'll look forward to hearing how well it works for you.  I've tried coax, toslink and AES/EBU into my Berkeley Alpha DAC and the AES/EBU interface was easily superior.  AES/EBU compared to S/PDIF uses a voltage range of 2-7 volts versus S/PDIF's range of 0.5-1 volts.  Clearly, other variables come into play such as what's on either end of the interface.  Good luck with your experiment.  Trying the I2S output is an interesting idea.

Leif,

Did you try it already with XXHE?

At their website I get the impression it only works with direct sound and kernel streaming.
Not with Wasapi yet. Is this true for you?

Email from m2tech says they are working on a wasapi-like interface for W7 64 bits....

I haven´t received it yet
will be back
best
Leif

Ordered the Hiface usb/spdif converter and it arrived today. Doesn't work with xxhighend, so I'll be sending it back unless...... I tried it in kernel streaming mode with media monkey and it sounded MUCH better than the M-Audio Audiophile usb that I currently use to convert usb to spdif. Now lets hope Engine #4 uses kernel streaming, cause media monkey just sucks compared to XX.

What happens ? Messages ?

Don't send it back yet, and for others : don't get discouraged - yet.
boleary, it may be your system, of which you know it is (somehow !) special. And FYI, my own Engine#3 not working anymore (as told in the other topic) is exactly and again related to that. Well, out of everyone, you know ... which in this case means I didn't change anything to the code part, but nevertheless it doesn't work anymore. There is just something going on in that WASAPI stuff which is still beyond me, although I expect it to be a compiler bug.

With XX it doesn't show up as an output device in the settings tab. I don't get any error messages. To get it to work with media monkey I had to download Steve Monks kernel streaming DLL as a plug in to media monkey, then select that plug in within media monkey. Whats weird is that it doesn't show up in "Sounds" tab of control panel at all, but it does show up in device manager under Sound Controllers.

But if that is so (besides it doesn't look good to ma at all), not any normal player would be able to use it.

I can indeed imagine Kernel Streaming is able to find the device, because KS depends merely on things like Direct Show filters, and once you know where they sit in your system, they can be controlled directly, without selecting any sound device first.

Notice I am not in the stage of being able to select the Sound Device with KS (currently it takes the Primary), but at the moment I have worked that out, it will be more clear to me.

Be careful though not to depend too heavily on KS as off yet, because you may not like it (for SQ) at all.
But I think you will ... :)

Well, since engine #4 relies on Kernel Streaming I guess I won't send it back (I have up to 30 days to do so.) The spdif audio  from the hiface is much superior to what I currently use. Can't wait to try it with XX!

Am at work now so I can't do any testing. Because the highface is supposed to work with Direct Sound it really should be showing up in "Sounds" in controll panel. I'm going to email the M2tech about this issue and will let you know what they say.

Just heard from M2tech. They will be releasing a wasapi driver for both Vista and Windows 7 sometime soon.

Very happy with my setup. See sig. Optical Spdiff is absolutely superb. Buffalo has a jitter correction anyway. Tried coax but could not hear an improvement, there was even less detail.