Friday, 18 October 2013

MEASUREMENTS: Separate vs. AV Receivers (Emotiva XSP-1 vs. Denon AVR-3802 vs. Onkyo TX-NR1009) as Analogue Preamp.


Okay, so maybe I'm being a bit too dramatic using that epic battle between Godzilla and Rodan above :-).

With the recent acquisition of the Emotiva XSP-1, I wanted to see just how well a separate pre-amp with audiophile design in mind stacks up against something more ubiquitous like the AV receivers out there. Remember that a preamp at its core has very basic functionality - it allows switching of the source and volume control by adjusting voltage gain on the output. The essential difference between a good preamp versus poor one (beyond features, ergonomics, remote control quality, etc...) is how well it maintains a high signal-to-noise ratio (SNR). If you supply a high resolution line level DAC output, you want to see that signal come out of the preamp with as much resolution as possible; this demands that the preamp introduce as little noise as possible.

Is there good evidence that spending money on a good analogue preamp will result in more accurate music reproduction? Let's find out in this installment...

First, let me introduce the contenders today:

Emotiva XSP-1:

Currently the highest quality Emotiva preamp out. The claim to fame is the differential design for balanced operation. It provides 2 balanced inputs along with a host of unbalanced RCAs. Volume control is through a digitally controlled, analogue resistor network. For this test, I will not be using the analogue bass management or tone controls that could affect signal presentation / quality. SNR for this device is rated at >110dB across the board for both RCA and XLR operation.

Denon AVR-3802:
A 7.1 channel "classic" from the decade when SACD and DVD-A were starting life and home theater lovers started seriously investing in discreet surround receivers with Dolby Digital and DTS (as opposed to the matrix surround of previous Dolby Pro Logic receivers). Analogue 7.1 channels input available. I bought this unit back in early 2002. Though not the highest end back then, it wasn't cheap (I think I paid almost $2000CAD). I don't remember the results of actual 3rd party testing but the rated amplifier power is 105W into 2 channels at 8ohms with 0.05%THD.

Analogue input SNR is rated at (only) ~86dB based on the table in the manual. For the sake of measuring the best possible audio output, all measurements will be performed through the CD input in "DIRECT" mode which bypasses all processing including tone controls and bass management. Measurement will be off the front (stereo) channels of the 7.1 "pre-out" analogue outputs.

ONKYO TX-NR1009:
My new receiver mentioned in the previous post. Capable of 9.2 channels processing with 7.1 analogue external inputs. Again, not the most expensive in the Onkyo line but certainly in the upper end of the previous generation released in mid-2011. Amplifier is capable of more power than the Denon with a rating of 135W into 2 channels at 8ohms with 0.08%THD. Sound & Vision measured it as 145W into 2 channels at 8ohms with 0.1% distortion. Line level SNR rated at 110dB.

I'll measure it through the analogue CD input in "Pure Audio" mode where all extraneous audio processing is turned off. Likewise, it's supposed to quiet the video circuitry and even the front LED indicators and display are turned off. I will measure from the front (stereo) channels of the 9.1 "pre-out" analogue output.

Firmware was updated to the latest version as of September 2013 - 1131-1399-0211-4108.

Setup:

 

First, I just want to discuss the general setup. The main thing I wanted to know was just how much resolution was maintained with the signal going through the preamp and look/listen to the results through a variety of levels standard across each device.

For the TEAC UD-501 DAC at 24/96 (which to me is the sweet spot) using the "SHARP" digital filter measured the following way:

HTPC (AMD A10 "Trinity") --> shielded USB (Belkin gold) --> TEAC UD-501 --> shielded RCA --> E-MU 0404USB --> shielded USB -->  Win8 laptop

we get these results:

The hope of course is that when we pass the above signal through the pre-amp, there will be minimal loss in resolution (noise level remains low around -113dB, and no change to frequency response to suggest "coloration" of the sound). In my previous TEAC measurements, I noted that the XLR output was too "hot" for the E-MU 0404USB without volume attenuation which drops the resolution. My guess would be that the noise level drops to less than -116dB. I will measure the XLR output from the Emotiva XSP-1 and see (it's the only device out of the 3 capable of balanced operation).
 
Using the digital oscilloscope, I found the following correlation between peak voltage output from each preamp device (accurate to <0.05V) and the volume control setting (using the TEAC RCA or XLR input of course):

Nice to see the volume control accuracy of each device - every halving of output peak voltage corresponds with a 6dB decline of the volume "knob". Notice that the Emotiva and Denon are using a relative system of volume control (dB below 0dBFS) and the ONKYO is set to an "absolute" measure between 0 to 100.

The setup incorporating the pre-amp in-line therefore looks like this:
HTPC (AMD A10 "Trinity") --> shielded USB (Belkin gold) --> TEAC UD-501 --> shielded RCA/XLR --> Pre-Amp device --> shielded RCA/XLR --> E-MU 0404USB --> shielded USB -->  Win8 laptop

Results:


I. Emotiva XSP-1 RCA:

Without further ado, here is what the Emotiva looks like with the unbalanced RCA setup:


Frequency Response

Dynamic Range
This looks really good. At 2V peak output, the dynamic range at 111dB is very close to the "ideal" (113dB directly from the TEAC). Notice a very small amount of roll-off in the high end when using the Emotiva. Also as expected, when the volume is reduced (2V --> 1V --> 500mV --> 250mV), the signal-to-noise ratio goes down and we see a concomitant reduction in the dynamic range and rise in noise level.

II. Emotiva XSP-1 XLR:

Let's have a look at the XSP-1 operating in a balanced configuration:

Frequency Response
Dynamic Range
Well folks, proof that if you want absolute fidelity, you really need to squeeze out those last few dB's down below 110dB with balanced XLR cables! Irrespective of whether you can hear it or not :-)!

Seriously, these are some fantastic measurements. As I said previously, unfortunately, I don't have direct measurements for the TEAC's XLR output. When passing the XLR output from the TEAC to the Emotiva pre-amp, the results at 2V peak volume coming out of the Emotiva slightly exceeds the direct RCA output from the TEAC DAC across the board from noise level to distortion levels to even lower stereo crosstalk.

The high frequency roll-off is less than with RCA. Notice just how clean the dynamic range graph looks as well through XLR cables. Fantastic.

III. Denon AVR-3802 RCA

Now we get into the AV receivers:

Frequency Response
Dynamic Range
Remember, I am measuring the Denon in "DIRECT" mode with all audio processing including bass management turned off (front stereo speakers set to "Large" for the sake of completeness). Not unexpectedly, these results are clearly a step down from the Emotive XSP-1. With a dynamic range of ~96dB at 2V, the Denon is capable of passing through 16-bit CD resolution but nothing more.

Roll-off above 20kHz is similar to the Emotiva's unbalanced mode but worse than the Emotiva below 100Hz with a -1dB bass roll-off at 20Hz.

IV. ONKYO TX-NR1009

Finally, let us have a look at what approximately a decade of advancement (between this and the Denon) in AV receiver technology can do:


Frequency Response
Dynamic Range
With the ONKYO in "Pure Audio" mode, no video processing at all, HDMI and video inputs all disconnected... Wow! That's very impressive IMO for a machine that's a "jack of all trades". In fact, these results are almost the same as the Emotiva XSP-1 functioning in unbalanced mode!

As excited as I am about those results above, a modern AV receiver is meant to process HDMI and be connected to a TV. This receiver has a HDMI "passthrough" which is essentially always in operation and for most people, it would not be left in "Pure Audio" mode with all the video gear disconnected. As such, look what happens when I connect my LG 55LW5600 TV (55" passive 3D, LED TV from 2011) to the ONKYO and repeated the measurements:

Frequency Response
Dynamic Range
Ugly, my friends... Clearly having the TV HDMI connected has injected very significant amount of noise in the system! Dynamic range has dropped to ~80dB across the board (equivalent to 13-bits). Notice a very strong 60Hz mains hum which is even showing up in the frequency response graph... What is happening here is that I'm seeing the effect of ground loops. There are ways to overcome this of course. For example, using a 3-to-2 prong adapter to disconnect the TEAC DAC from ground resulted in the following:

Frequency Response

Dynamic Range
About 10dB improvement just by doing this. For now, I'm not going to bother isolating the problem further (I'll be moving house in about a month!) but suffice it to say that in a receiver setup with complex component interconnections, be very careful of noise polluting the analogue output as demonstrated above. Ground loops are very common especially with TV systems where ethernet and coaxial cables are often connected to the TV/receiver creating a number of potential ground points beyond the individual device plug-ins to the wall.

Summary:

 

So there you have it. The Emotiva XSP-1 measures as a very capable pre-amp unit with excellent resolution especially when used in a balanced configuration. There was barely any loss through RCA and the XLR performance is beyond the E-MU 0404USB's measurement capabilities. Note that in all these tests, I'm just using generic "Radio Shack" type RCA connectors and the XLR cables are inexpensive Monoprice brand. No reason for spending money on expensive cables when these kinds of results can be obtained with standard decent interconnects.

In "Pure Audio" mode without the HDMI system connected, the ONKYO performed excellently. It bested the 12-year old Denon AVR-3802 handily and is essentially neck-and-neck with the Emotiva in an unbalanced configuration. However, beware of the potential noise pollution and ground loops once you plug in all sorts of things into these receivers (like your fancy big screen TV)!

A little while back, I spoke about how music sounded better through the Emotiva XSP-1 compared to using the Denon as pre-amp. These results are supportive of my subjective impressions (I'm showing >10dB dynamic range difference and bass roll-off differences between the two). As for the ONKYO, it does sound much like the Emotiva as a stand alone audio device. Since I will be listening primarily with the ONKYO as a HDMI DAC (either from the computer or through Blu-ray player), for me the digital audio performance is much more important than the results I show here from the analogue input.

Music this evening:
My kids enjoyed Les Misérables (the movie) and really love to listen to it in the car on the way to school... My favorite recording of this is the recent 25th anniversary UK Tour cast from 2010's performance called "Les Misérables Live!" - certainly much better singing than Hugh Jackman and Russell Crowe!

I've got another trip coming up at the beginning of November and then the house move by the end of November. I hope to put up some results of the ONKYO as HDMI DAC before I go. Until next time... Enjoy the tunes...


5 comments:

  1. Really looking forward to your tests of the Onkyo as an HDMI-connected DAC. I have the NR-809 (very similar to your 1009) and have been using it this way for some time. I think there it may sound better with an external DAC (right now, using the DragonFly) through analog in, but I also like the idea of the direct HDMI connection with support up to 24/192. Will be interesting to see how it fares.

    ReplyDelete
  2. This comment has been removed by the author.

    ReplyDelete
    Replies
    1. Thanks for another interesting and informative post. Regarding the Onkyo as a HDMI DAC, I think that there must be something I don't understand; it seems to me that the post-DAC audio signals will be passing through the exact same analogue stages in the pre-amp as the analogue inputs i.e. the results will be just as hideous. I'd be grateful if you could tell me what I'm missing (it might affect a planned purchase!)

      Delete
    2. Spoiler: It's actually not as hideous :-)

      I'll get it up soon... Just thinking of what other probing I can do before packing up the receiver for the move!

      Delete