[slim.a]

Quote:

Originally Posted by jumping /img/forum/go_quote.gif Just a brief digression back to the original topic. I really appreciate your honest review of the USB to SPDIF converters that started this thread. I have a huge collection of cd's ripped to flac files on computer hard drives and have been trying for years to find a way to play them on audiophile equipment. The best of those cd's sound great on just about anything, but most have a very definite edgy quality played thru even the best of DAC's. The Terralink-X did just what you said it would and allows me to enjoy a huge collection of cd's played thru a usb hub linked to a PS Audio Digital Link 3 with level 3 Cullen mods. The harshness in the upper registers that has plagued me for the last 3 years with several decent DAC's is completely gone. I am not entirely sure why, since the PS Audio DAC upsamples everything to 192, but the combination of the Terralink-X thru an RCA coax link to my DAC sounds terrific on even my old cd's. By the way, the RCA coax is audibly better than the optical link. Makes me want to try the new PS Audio Perfect Wave DAC in the native mode. (When I next have a few grand to drop on their system I know this is probably heresy, but I think that all the upsampling aggravates the harshness of the standard 44.1 cd recordings.

Hi jumping,

Thanks for your impressions! Glad to see you have finally achieved a pleasing sound from your computer.

As for the harshness you have noticed with the PS Audio Digital Link 3, it doesn't necessarily come from the upsampling but in my opinion, it rather comes from the technology used in the DAC. The PS Audio Digital Link 3 uses a sigma-delta dac chip which are known to be harsh sounding compared to the R2R/multibit DACs. To not digress again, I will suggest reading this interesting article : Mother of Tone - Conversion Techniques

By the way, while I used to think that the low jitter hiface could be a good solution for every application, I came to realize that in some cases, using a slightly more jittery interface (such as the Teralink-x) can provide better subjective results whith DACs that are not neutral.
When I was reviewing the Purepiper DAC A-1 (which uses a sigma delta CS4398), the most "pleasing" sound to the ears was not achieved through the Hiface but through the warmer sounding devices (Teralink-X and Purepiper own usb to spdif converter).
I think that by lowering the jitter too much, we start hearing other types of distortions (that were previously masked by jitter).
Of course, not all "jittery" interfaces are pleasing to the ears. For example, the EMU 0404 (when used as usb to spdif converter) is less detailed than other converters and is more edgy at the same time.

Overall, from my own experience (and what I read about different units), it appears safe to say that the Teralink-X is a relatively detailed converter but slightly on the warm side while the Hiface is more detailed device but rather on the neutral side.

If you are planning to upgrade your DAC and want to improve both details and naturalness of the sound, I suggest looking towards some R2R/multibit based DACs (that use the PCM1704 chip for example).

 

[regal]

Any comparisons to the old terralinx X1 version vs x2?

 

[leeperry]

Quote:

Originally Posted by slim.a /img/forum/go_quote.gif not everybody thinks that glass is the ultimate and only way to go.

yes, that's why the higher grade glass toslink is only made of 65 strands as I understand it.

the blue jeans cable guy says the same stuff and that his Mitsubishi wire is much better because it's only made of one strand....but on the link I sent you the guys are saying that the mitsu cable was worthless, so as usual YMMV

"most other manufacture including more expensive one terminate their cable by hot glass or knife."
that's exactly what I suspected!

anyway, I've received my NG COAX from ebay, I'll be playing around w/ my audioquest coax cable and the ploytec drivers in KS

is it possible to get KS on XP w/ any teralink? you said that the CMI108 drivers were horrid..and C-Media is known to cheat in their drivers to great extend(like the stock CMI8768 drivers that only output 14bit).

 

[slim.a]

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif anyway, I've received my NG COAX from ebay, I'll be playing around w/ my audioquest coax cable and the ploytec drivers in KS is it possible to get KS on XP w/ any teralink? you said that the CMI108 drivers were horrid..and C-Media is known to cheat in their drivers to great extend(like the stock CMI8768 drivers that only output 14bit).

Did you notice any improvement with the NG coax vs. the spdif out of your motherboard?

As for KS, I haven't been able to make it work with the Teralink-X(1).
The weird thing is that when setting up ASIO4ALL in Foobar, it recognizes the KS ... but when I choose to output directly in KS output (ie no asio4all), it doesn't work and I have no sound.

Concerning the cmedia drivers, as I mentioned earlier they are pretty poor sounding and not bit perfect (the loss of resolution also shows at analog stage of my DAC). So even if they worked with KS, the data would already be messed with by the drivers. So there is no real sonic benefit to expect from those cmedia drivers.

By the way, leeperry, if you figure out a method to output KS without those cmedia drivers let us know.

 

[leeperry]

yes, ASIO4ALL can get KS from most and every audio interface...there's an internet board where audio drivers engineers exchange tips, and they're all blown away by how well ASIO4ALL works

so sometimes you won't get a properly working DirectKS miniport, but ASIO4ALL will still work fine...tbh, I'd trust ASIO4ALL more than a poorly coded DirectKS implementation.

I didn't try the NG COAX yet, I'm getting lazy these days....a groundloop-free connection to the computer, a clean linear discrete PSU, some high-end opamps...I'm not too hard to please after all ^^

well, KS is mandatory for me as I'm a major Reclock enthusiast...and it happens to only work in KS or WASAPI, and I'm quite fund of XP tbh.

KS in the ploytec drivers should work just fine on XP w/ generic TI chips

 

[leeperry]

BTW, on XP uLilith sounds MILES better than foobar(that seems heavily flawed) to my ears..I use it w/ ASIO4ALL: Index of /download/uLilith



uL/Reclock/foobar all sound completely different to my ears(and to other ppl's too, mind you

)

the nice thing about uL is that the whole audio pipeline works in 64bit float...lossy audio decoding, DSP/VST plugins, volume attenuation etc etc

tbh, even in 100% bit-perfect it kills foobar on XP....jitter at the software level would appear to be very real:

cMP² | CPlay / SoftwareInducedJitter

Measuring XXHighEnd ...

 

[punk_guy182]

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif tbh, even in 100% bit-perfect it kills foobar on XP....jitter at the software level would appear to be very real:

Not much difference in Vista 64 IMO and I used Direct Sound and WASAPI for comparaison with Foobar2000.
Now jitter at the software level, that's something I don't understand.
I remember reading somewhere that Amarra is a less jittery* software compared to others.

 

[slim.a]

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif BTW, on XP uLilith sounds MILES better than foobar(that seems heavily flawed) to my ears..I use it w/ ASIO4ALL: Index of /download/uLilith uL/Reclock/foobar all sound completely different to my ears(and to other ppl's too, mind you ) the nice thing about uL is that the whole audio pipeline works in 64bit float...lossy audio decoding, DSP/VST plugins, volume attenuation etc etc tbh, even in 100% bit-perfect it kills foobar on XP....jitter at the software level would appear to be very real: cMP² | CPlay / SoftwareInducedJitter Measuring XXHighEnd ...

I have also noticed differences between media players.

By the way, which version of Foobar are you using? Because they all sound different. I have tested (shortly) the uLilith you linked and I didn't find that it was as good as Foobar v1.0 as it sounded harsher to me (using the hiface).

From the previous (more lengthy) comparisons I have done before between different players, here is my personal order of preference (using the hiface as a transport):
1) a) cPlay
1) b) Foobar v1.0 with KS output
2) Foobar v0.8.3 with ASIO4ALL and otachan 0.51.7
3) Foobar v0.9.x
4) MediaMonkey
5) Windows Media Player

cPlay was perhaps slightly better than Foobar v1.0+KS but I rarely use it because it not very user friendly.
I have also noticed that with different usb converters, the results might differ a little bit. So I won't be surprise if someone finds MediaMonkey to perform better with a specific converter.

Whether it is jitter or something else entirely, those different playback methods give different sonic results.
Personally, I would like to try the XXhighend player but it doesn't seem to work correctly in my laptop (not enough memory I think).

Edit -- that http://cplay.sourceforge.net/pmwiki....eInducedJitter article is very intersting to read. Thanks for the link!

 

[leeperry]

Quote:

Originally Posted by punk_guy182 /img/forum/go_quote.gif Not much difference in Vista 64 IMO and I used Direct Sound and WASAPI for comparaison with Foobar2000. Now jitter at the software level, that's something I don't understand. I remember reading somewhere that Amarra is a less jittery* software compared to others.

yes, I mentioned XP because the difference would hardly be audible in the newer OS...for the good or the bad remains to be seen, as many ppl have said that audio was very different between XP and Vista/W7.

 

[noinimod]

keep those teralink x2 impressions coming in! I'm just about to pull the trigger on these hehe

 

[slim.a]

Quote:

Originally Posted by noinimod /img/forum/go_quote.gif keep those teralink x2 impressions coming in! I'm just about to pull the trigger on these hehe

So you decided to go with the Teralink-X2? Are you sure you are not tempted by the hiface?

 

[leeperry]

Quote:

Originally Posted by slim.a /img/forum/go_quote.gif I have also noticed differences between media players. By the way, which version of Foobar are you using? Because they all sound different. I have tested (shortly) the uLilith you linked and I didn't find that it was as good as Foobar v1.0 as it sounded harsher to me (using the hiface).

on what windows version did you run those tests?

well, a simple test(conducted many times!) on XP SP3 w/ different soundcards(Envy24HT/CMI8768 DSP's mostly):

-foobar 1.0 final in bit-perfect KS/ASIO/ASIO4ALL: poor bass response, harsh trebles, agressive and edgy = digitis here I come! the problem seems to lie within the foobar audio pipeline...it sounds like erroneous integer<>float conversions or sumthing like that.

-Reclock in bit-perfect KS: much better bass response, very clear(but a tad agressive for music) trebles, excellent stereo imaging...fantastic for movies dialogs clarity all the way to the top! Reclock relies on a 0.17ppm clocking system based on the motherboard PMTimer...this is serious business! I've also set the windows timer granularity to the lowest figure possible(9766 ms) and changed the win32 priority system to give a 64X higher priority to "high priority" processes...this is as tight as XP can be! of course HPET in Vista/W7 would kill it

I'm a dedicated Reclock beta-tester, and at some point I annoyed the guys to look at improving the float<>integer conversions...and the improvements were pretty amazing, as they were ever so slightly off.

-uLilith w/ ASIO4ALL : analog/vynil sounding(a perfect match to the AD797B opamp

), it's like Reclock w/o the harsh trebles...digitis is gone, stereo imaging is perfect, bass is very tight and clear, mids are as creamy as can be...I'm in love

now, I sound like a complete idiophile, be it!

 

[slim.a]

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif on what windows version did you run those tests? well, a simple test(conducted many times!) on XP SP3 w/ different soundcards(Envy24HT/CMI8768 DSP's mostly): -foobar 1.0 final in bit-perfect KS/ASIO/ASIO4ALL: poor bass response, harsh trebles, agressive and edgy = digitis here I come! the problem seems to lie within the foobar audio pipeline...it sounds like erroneous integer<>float conversions or sumthing like that. -Reclock in bit-perfect KS: much better bass response, very clear(but a tad agressive for music) trebles, excellent stereo imaging...fantastic for movies dialogs clarity all the way to the top! Reclock relies on a 0.17ppm clocking system based on the motherboard PMTimer...this is serious business! I've also set the windows timer granularity to the lowest figure possible(9766 ms) and changed the win32 priority system to give a 64X higher priority to "high priority" processes...this is as tight as XP can be! of course HPET in Vista/W7 would kill it I'm a dedicated Reclock beta-tester, and at some point I've annoyed the guys to look at improving the float<>integer conversions...and the improvements were pretty amazing, as they were ever so slightly off. -uLilith w/ ASIO4ALL : analog/vynil sounding(a perfect match to the AD797B opamp ), it's like Reclock w/o the harsh trebles...digitis is gone, stereo imaging is perfect, bass is very tight and clear, mids are creamy as can be...I'm in love now, I sound like a complete idiophile, be it!

I did all my testing on XP media Center SP3.

And for the listening tests, I tried to make sure that the hardware side was as transparent as possible.
For my tests I used the m2tech Hiface which is the most transparent and neutral usb to spdif converter I have. If had done the test with the Teralink or the EMU 0404 usb, I would have probably ended up with another order of preference to compensate for their flaws.
Also, I used the dac19mk3 for the tests: again, If I had used the Audio-gd FUN or the Purepiper DAC A-1, I would have found other results because both are colored. The dac19mk3 is a well designed R2R/multibit dac with a zero feedback (current gain) output stage. So if I lower jitter either by the usb converter/the digital cable/... I don't risk having other kind of annoying distortions become more audible. I have listened to many CD players and DACs (cheap and very expensive) based on delta sigma dacs and none of them sounded natural.

So to cut it short, for my different tests I used high quality tracks (either in 16/44 or 24/96) many of them contain unamplified instruments: piano, violin, cello, ... And at no time, I found that Foobar v1.0 was harsh or unnatural sounding. Playing Keith Jarrett - Paris / London – Testament or Glenn Gould - The Goldberg Variations you just feel like there are real pianos in the room. No digitis whatsoever.

So if you noticed any harshness/agressiveness/poor bass response on Foobar v1.0 I am almost certain it comes from elswhere from your chain: Envy24HT, digital cable, ... or maybe from Foobar settings : Did you disable replay gain and other DSPs?

BTW, where did you get this info : "Reclock relies on a 0.17ppm clocking system based on the motherboard PMTimer...this is serious business!"
First there is a difference between ppm (clock stability over time) and ps(jitter). And even there was a super low jitter clock somewhere inside a computer, mohterboards are so noisy that any clock would be messed up before it gets to the spdif outputs. If you look at high end DACs and transports you will see that a lot of care is taken to keep PS noise minimal, circuits very short, impedance, shielding, ... and I see none of that in most motherboards.

 

[dszabi]

Re the delta-sigma vs R2R/multibit DACs:
Does anyone know which type of DAC is the Musical Fidelity X-DACV3 (DSD1792/SRC4192/3 chipset)? Thanks!

 

[leeperry]

I was running all the players(on XP Pro SP3) in 100% bit-perfect...no DSP whatsoever.

a few links: PERFGURU : Explanation for the USEPMTIMER switch in the boot.ini
Quote:

configuring the /USEPMTIMER switch forces the machine to select the ACPI clock for the performance counter timer instead of using the time stamp counter (TSC).

Guidelines For Providing Multimedia Timer Support
Quote:

PM Clock The ACPI timer, also known as the PM clock, was added to the system architecture to provide reliable timestamps independently of the processors speed. Because this was the single goal of this timer, it was designed to provide a time stamp in a single clock cycle

open CPU-Z, look at your CPU clock speed...this is not a good clocking device AT ALL.

you can set the windows timer granularity using /TIMERES: Boot INI Options Reference
Quote:

/TIMERES= Sets the resolution of the system timer on the standard x86 multiprocessor HAL (Halmps.dll). The argument is a number interpreted in hundreds of nanoseconds, but the rate is set to the closest resolution the HAL supports that isn't larger than the one requested.

so, you can just add this to your XP boot.ini:
Code:

[left]/NOEXECUTE=ALWAYSOFF /FASTDETECT /USEPMTIMER /NODEBUG /TIMERES=9766[/left]

actually, you can read more about all this in this PDF from the cPlay's author: The art of building Computer Transports v0.3.pdf - Google Search

so now, Windows will base its timestamps on the PM clock, w/ the lowest granularity available...that's the only way for me to reach 0.17ppm in Reclock:



XP wasn't meant for realtime(HPET being the cure

), but this is as good as XP will get tightness-wise.

to get back on the jitter issue...how is that that a lower jitter would always be desirable?

when I compared the Asus STX to the ST...the stereo imaging was far more accurate on the ST, but it was also *very* fatiguing...my brain clearly couldn't stand it

it'd be naive to think that the only difference between these two cards is the clock conditioner to get a more stable 44.1kHz multiple clock out of the the 24.576MHz PLL...as the ST runs a bridge to allow the CMI8788 PCI native DSP to work over PCI-E, and that the PCI/PCI-E layouts are night and day...still, I'm not too sure that my brain craves for an uber-low jitter.

Burson explains here what to expect from a low jitter clock:
Quote:

By reducing the jitter error, you will hear clearer positioning, also details are further refined vocally and instrumentally. Sound stage and positioning will improve noticeably and that includes deeper sound stage and darker background.

and it matches w/ the diff. I heard between the STX/ST...it made the ST over-fatiguing and the STX having an unbearable mushy stereo image.

OTOH, the RMAA measurements for the ST were quite funky(I sent it back for RMA, but the replacement card was the exact same)....I'm still not convinced that an uber-low jitter is useful/desirable and something to crave for.

besides you're only as weak as your weakest link(again) so as you said, CS8414/DIR9001 will react differently..

and I wouldn't get too hung up on the DAC...as the opamps you're using as LPF or I/V on the DAC output will color the sound far more than switching DAC's ever would from what many ppl believe(it's also been my personal experience

): http://forum.rightmark.org/topic.cgi?id=4:504-3
Quote:

Everybody -in different locations, at different times, without knowing from each other- told the same story, that they found the differences between opamps more important than the differences in dac chips.

damn, does "John Barry - Midnight Cowboy.flac" sound good