Audio Asylum Thread Printer
Get a view of an entire thread on one page
|For Sale Ads|
In Reply to: RE: ? posted by Jim Pearce on June 02, 2017 at 07:17:11
Oppo claims the outputs on its player sound identical, yet a reviewer found "major sonic differences...noticeable immediately". So what's up with that?
In my experience, most (even those who should know better) don't control all the variables properly when conducting comparisons such as this. I've seen experienced reviewers writing for print magazines attempt to reach conclusions while using different brands and models of cables for each connection. And even if there is a difference, is it due to the source or the receiver?
It's a difficult experiment to control, and may not even be possible with some equipment. An example of this was Sony's original SACD player, the SCD-1. It was pretty much universally agreed that the single-ended outputs sounded better than the balanced outputs. When you look at the circuit, this makes perfect sense. At some point after the DAC chip, the signal was purely single ended. The RCA jacks connected directly to this point, but the positive (+) and negative pin (-) of the XLR went through two op-amp buffers with 100% negative feedback (to create unity gain), one wired in non-inverting mode and the other in inverting mode (to create the out-of-phase signal). It's no surprise that when connecting that output signal to a true differential balanced input on a preamplifier that the result would be a reduction in sound quality, as signal was passing through an additional amplification stage of lower performance than the discrete unbalanced circuitry used for the single-ended output.
The reason for audiophiles to use balanced has almost nothing to do with the reason balanced connections are mandatory in professional recording studios. In the pro environment they are worried about picking up external magnetic hum fields in long (100') cables carrying low-level (microphone) signals. That is not a problem at all for home playback - I've even used 20' unbalanced cables between a low-level MC phono cartridge and the preamp without hum problems.
The real sonic advantage of balanced only comes into play if the internal circuitry is truly differential from input to output. Since every circuit in the world (analog or digital, tube or solid-state, audio or video) is nothing more than a modulated power supply, it only makes sense that the quality of the power supply is important. And as there is no such thing as a "perfect" power supply, gains can be made by reducing the audio circuit's sensitivity to imperfections in the power supply. A differential balanced stage will typically exhibit an increase in power-supply rejection ratio (PSRR) between 40dB (100x) and 80dB (10,000x) over identical single-ended circuitry. Therefore using balanced differential circuitry is like making the power supply between 100x and 10,000x better.
The easiest way to hear what difference this makes is to listen to a Pono Player using a headphone with detachable cables. When comparing otherwise identical single-ended and balanced harnesses (same cable manufacturer and model), the signal path from the DAC chip to the headphone driver is fully balanced differential all the way through for a balanced harness, while a single-ended harness is fully balanced differential *except* for the final output buffer stage. The only variable in this test is whether a single stage in the chain is operated in single-ended or differential balanced mode. There are many reports on the internet about the improvement in sound quality available from using a balanced connection, but I've never seen the opposite conclusion.
The bottom line is that properly executed differential balanced circuitry will always sound better - but only if all else is held equal. The penalty is pretty much a doubling of the circuitry and parts count - hence price, size, and power consumption. There are other ways to both improve power supplies and also reduce a circuit's sensitivity to imperfections in the power supply, but those can also be applied to differential balanced circuits for further improved performance. In other words, it is also possible to make a single-ended circuit perform better than a balanced circuit - but only by improving the single-ended circuit and not the balanced one. That is why the caveat "if all else is equal" is so critical.
As usual, strictly my personal opinion and not necessarily that of my employer or hairdresser.
Oppo may claim the balanced and single-ended outputs of the 205 should sound identical, but using a balanced connection between my 205 and Ayre K-5xeMP resulted in more detail and a greater sense of transparency. The improvement over single-ended was very noticeable. I played CDs, SACDs, and Blu-rays, jazz trios, baroque, and opera. In each case, the sound was superb.
Thank you, Charles Hansen, for a post that convinced me to make the change.
> > using a balanced connection between my 205 and Ayre K-5xeMP resulted in more detail and a greater sense of transparency. The improvement over single-ended was very noticeable. < <
Thanks for sharing your experiences. Enjoy your new Blu-ray player!
Oppo actually claims that the xlr output sounds better, no matter what customer service says:
"The stereo output section offers both XLR balanced and RCA single-ended connectors. By transmitting a pair of differential signals, the balanced output provides better common-mode noise rejection and improves signal quality."
Actually their claims across their audiophile players - BDP-95, BDP-105 and now UDP-205 - remain consistent:
1. The dedicated L,R RCA outputs sound better than the L,R front 5.1
2. The xlr outputs sound better than the dedicated L,R stereo outputs.
Personally I can't attest to having an opinion even with the BDP-95 because I use the xlr outs to defeat bass management in the player without having to dive into the menu. The one thing I'm sure of is that any given stereo or mono disc will sound better in 2.1 (or 1.1) with BM or through the xlr outputs.
> > By transmitting a pair of differential signals, the balanced output provides better common-mode noise rejection and improves signal quality." < <
Yes that is true for specific situations, that are mostly found in recording studios and not home playback systems. My experience is that the common-mode noise rejection of the signal input is not nearly as important as the fact that *differential* balanced circuitry will also provide the same type of rejection to imperfections in the power supplies. The importance of power supply quality becomes apparent when one realizes that the audio circuitry (indeed *all* electronic circuits) are various types of ways to modulate the power supply. Just as with computers, garbage in = garbage out (GIGO). Hence the importance of power supply quality. Differential balanced circuits add an entire extra level of rejecting imperfections in the power supplies.
> > The dedicated L,R RCA outputs sound better than the L,R front 5.1 < <
For the Oppo players this is certainly true of the measured performance. They normally use an 8-channel DAC chip for the 7.1 analog output, whereas the dedicated stereo outputs typically parallel 4 DAC chip channels for each stereo channel. This improves the S/N ratio by 6dB, assuming all else is held constant. Oppo may also use better parts and or power supplies on the dedicated stereo channels (they may be too expensive to use for all 8 channels, just as most customers use higher quality main speakers than surround speakers).
> > I use the xlr outs to defeat bass management in the player without having to dive into the menu. < <
That's fine for all CDs, but it turns out that there are some DVDs that were authored improperly with regards to bass-management. The normal 2-channel mixdown will *not* recover the bass on these discs if you have an HT-2.0 or HT-2.1 system. Ayre added a feature on the DX-5 to correct for this, as described in the owner's manual. To the best of my knowledge, this is the only Blu-ray/DVD player in the world with that feature. Hope this helps.
I play all MC discs in MC mode through the 5.1 analog outputs. I thought it was generally accepted that most common-mode noise comes from the power supply.
If you are not mixing down to stereo, there is no problem created by improperly authored video discs.
> > I thought it was generally accepted that most common-mode noise comes from the power supply. < <
There is at least one other source of common-mode noise found in all digital products - switching noise from the DAC chip itself. With each new sample the output switches to a new state. Real world devices are imperfect and have parasitic capacitances, such that some of the electrical energy used to turn the individual DAC switches on and off is coupled into the audio output signal. Using either two single-ended DACs or one balanced output DAC per audio channel means that the switching noise will be the same in both phases. If the first analog stage is a true differential balanced design, it will reject the common-mode switching noise from the DAC chip. (This is akin to what happens in the recording studio where a low-level microphone signal picks up line frequency hum from the AC wiring in the studio walls equally in both conductors of the balanced mic cable.)
As always, only my personal opinion and not necessarily that of my employer or dog-catcher.
Will the ESS minimum phase slow filter on the Oppo 205 likely approximate what you have done with "MP"?
> > Will the ESS minimum phase slow filter on the Oppo 205 likely approximate what you have done with "MP"? < <
Good question, but hard to say with any certainty. The only real way to know is with listening tests, but there is enough information that we can say "probably not". Although both are described as "slow rolloff, minimum-phase digital filters", there are many differences we can see and likely more that we can't see. About the only visibility I have into the ESS filter comes from the datasheet, which is not publicly available. It shows graphs of both its frequency response and its impulse response. Based on that alone, there are significant differences between the filters. Ayre's single-rate slow rolloff MP filter has about 1-1/2 cycles of post-ringing, while the ESS has about 4-5 cycles of post-ringing.
We have a few data points to help us predict the audibility of this difference. Ayre DACs with a "Listen"/"Measure" filter selection allow the choice of two minimum-phase filters, one with about 20 cycles of post-ringing, the other with 1-1/2 and the difference is fairly easily noticed. In contrast when JA recently reviewed the Meridian Ultra DAC (link below), he found it virtually impossible to hear any differences between minimum-phase filters, the shortest with about 6-7 cycles of post-ringing and the longest with ~35 cycles of post-ringing. Based on these two sets of data points, I suspect that the filter length alone would create an audible difference between the filters. However I also suspect that the slow rolloff MP filter in the ESS DAC chip would be my preference of the 7 choices offered.
That is just the tip of the iceberg, however. While the digital filter is one part of the design that affects the sound, there are many other aspects to digital filter design not mentioned above, including the window shape, the dithering applied, the interpolation rate (ESS uses 8x, while Ayre uses 16x). Further there are many more factors that affect the sound of a D/A converter In my experience, digital filters are one of about a half-dozen major design aspects that will significantly affect the sound of a digital product. Also important are the analog circuitry, the power supplies, the clock implementation, the DAC chip itself, and the presence or absence of any DSP algorithms (such as Asynchronous Sample Rate Conversion - ASRC - and many others). I would be loathe to rank the importance of them, but would agree that the digital filter (or lack thereof, in the case of "non-oversampling") plays an important factor in the overall sound quality.
As always, strictly my own opinions and not necessarily those of my employer or trash collector.
> when JA recently reviewed the Meridian Ultra DAC (link below), he found it
> virtually impossible to hear any differences between minimum-phase filters,
> the shortest with about 6-7 cycles of post-ringing and the longest with ~35
> cycles of post-ringing.
My comment was made auditioning a 192kHz file, whereas my impulse response
measurements were made with 44.1kHz data. After the review was published,
I was told that the Meridian's filters behave differently at 2Fs and 4Fs rates from
how they do with baseband data. I had assumed they were the same at all rates.
You're ability to explain things in a simple and concise manner, so clear that EVEN I can begin to understand, is much appreciated!
Thanks for the kind words, and glad I could be of assistance. However many would disagree with your assessment of my posts as "concise"... :-)
Enjoy your weekend!
Thanks for the detailed explanation. I think I now understand why I should use a balanced connection between the Oppo 205 and Ayre preamp.
Post a Message!
This post is made possible by the generous support of people like you and our sponsors: