Guys, I decided to put up a blog page with my XONAR E1 measurements with full commentary on these frequency anomalies...  IMO the frequency roll off at 24/96 and 24/192 are a bit unusual as well even apart from the upsampling issue.

 

Anyhow, I do hope ASUS has a look at this and come up with some fixes in the firmware!

 

http://archimago.blogspot.ca/2013/02/measurement-asus-xonar-essence-one.html

Great job arch.

Darn, double posted on accident.

I have been enjoying the XLR the best, and your finding proves I am right about my finding.

Makes you doubt many of what we are reading:

"it's interesting that most 'official' reviewers have not mentioned this rather obvious anomaly."

Thank you for sharing, it was one of the best readings about the E1.

Edited by turokrocks - 2/24/13 at 2:20am

Hi, It would be great if you can plot phase measurements for normal and upsampled signals. Maybe there is a secret.

Best regards

Hi DM. I see you posted some graphs of phase characteristics earlier in post 1625.

 

Unfortunately it looks like I'm hampered by the (free) version of RightMark I'm using and have not been able to recover the phase data measured. What's your observation on this?

 

Best regards...

Edited by Archimago - 2/24/13 at 9:07am

Yup, I'm quite weary these days about the purely subjective reviews since rarely have I seen these reviewers taking a stand on real problems like the high frequency roll off...  They'd need to prove it, and that requires hooking up the gear to have a look objectively. Even worse are some of these "reviews" found on line which essentially reads like company PR material ;-(

 

John Atkinson's objective measurements are also the main reason I still have my subscription to Stereophile.

I can loopback calibrate my Terratec EWX 24/96 soundcard and make pretty detail measurements in analog domain. Problem is when I output signal through E1 via USB and measure  analog output of E1 there is no calibrated reference, that is very important for phase measurement.

My posted graphs show relative difference between normal and upsampled processing. It seems to me that upsampled phase response is smoother (close to zero). At 44.1 and 48 kHz sampling it is of no use because of early roll-off, but at  higher sampling rates it is preferrable. I would like to read listening observations, especially with top of the line headphones. 

Edited by DMarasovic - 2/24/13 at 10:08am

Thank you guys for looking into this!

As far as I know, Asus is currently working on a fix.

 

I can't hear the diffrence between upsampling and non-upsampling on every track, however, I do notice a loss in the higher frequencies/sparkle on certain quality, high dynamic songs.

I've got my E1 set to 96khz/24bit via optical to my soundcard. I notice a significant increase in volume (1/3~ volume only over optical) if I use the usb output directly to my pc. However, I'll also start to hear that buzzing noise in the background - probably because of the custom opamps I use.

I hope ASUSXONAR has a response to this.

Thanks for the note DM. I see what you mean.  Will have to look into the phase response deeper next time I run it thru the tests...

Kenion, has ASUS ever responded to this in the past?

 

I agree Dynamics, I hope ASUSXONAR gets back to us (and headquarters!) on this...

Hi Tim

I am new here and been looking back through the forum to see what you have done to upgrade your E1.

I have the E1 Plus. In the op-amp manual it recommends to change slot 2 with Muses 02.

I am thinking of trying this in slot 1 & 2 and maybe in the RCA buffer output stage.

Has anyone used Muses 02's instead of 01's? I heard they are a better op-amp, Bi-polar not FET.

What upgrades have you tried? & is that a custom op-amp in the buffer RCA output stage?

I also seen this link for a custom op-amp.

http://www.hificollective.co.uk/components/newclassdopamps.html 

01 is superior for the Essence One.
I only have two MUSES02's.

Your link is down.

RCA & Headphone buffers are using two LME49710HA metal can op-amps on TO-99 to DIP8 adapters. Less bias/crosstalk/distortion than the stock LME49720. If you're too lazy doing all that, then just get a LME49720HA pre-soldered to a DIP socket. Should sound a little more "real" than the non metal version.

I've swapped most electrolytic capacitors from Nichicon KT 25V to Nichicon KA 50V.
Also swapped all film capacitors from Wima FKS 10% / FKP 5% to Wima FKP 2.5% Tolerance.
Swapped one Sanyo polymer capacitor for a Nichicon FP (which Schiit seem to also use and like), Don't believe me? Check their Bifrost and Gungnir pcb photos.
Swapped (2) RCA capacitors with Elna Silmic II as i find them to be the best performing audio capacitor (especially after 100+ hours) and close to close to film.

Changed power supply mosfet's thermal paste with Prolimatech PK-3, and Added a custom heat sink on the right side which oh boy, is very warm whenever i touch it, so its working like a charm. New paste dissapates heat very sell, i have replacement psu capacitors, exact brand/model/capacitance/voltage for whenever they give out.

All wires directly connected to or on the main pcb has been changed out, since im a hardcore believer in wires.
Last but not least, my led change from stock blues to less bright reds.

If you ever try the muses02's let us know.

Tim

 

I can create phase response and group delay graphs if you upload the impulse response(s) in WAV or FLAC format. Also anything that can be seen here if you post the loopback recording of the test signal.

 

By the way, the higher noise on the RCA outputs could be the result of a ground loop if both the E-Mu 0404 and the Essence One were connected to the same computer. Of course, with a grounded external amplifier, that is still a real problem, and not just a measurement artifact, but without a ground loop, the unbalanced output would likely be better.

Edited by stv014 - 2/25/13 at 6:28am