[Jon Sonne]

I have been searching the web (and head-fi) for an ABX test of balanced vs. single ended setup, but with no luck. Has this ever been done? (with volume levels matched within ±0.1 dB).

 

[jcx]

there are measureable differences in almost every setup - and Psychoacoustics gives a few guides on human hearing limits, for some "clearly audible" thresholds
 
but most with the knowledge to construct such a test, with good quality amplifiers expect a null result
 
check the flow chart: http://www.delta.dk/imported/senselab/AES125_Tutorial_T4_Perceptual_Audio_Evaluation_Tutorial.pdf
 
expecting null results reduces the motivation a lot - while we may make inferences from many accumulated null results there will always be hold outs that null results aren't "proof" that the next ripe apple shook off the tree won't go orbital
 
so the "believers" really have to be the ones to do the tests since they expect positive results - but few of them will use the controls or will argue after the fact that the controls interfered with the difference they "know" they can hear sighted but the blind statistics say not
 
 
and it is fair for "believers" to criticize many amateur DBT - some are poorly designed and executed by "debunking enthusiasts"

 

[sk3383]

Jon Sonne

The difference between Balanced and Singled Ended has many factors.

I think you are asking the wrong question.

I concede, I am not an electricial engineer by trade, however I have a wealth of related experience in the subject.

Firstly, we need to set some variables, guidelines if you will, in order to narrow down the differences.

Focus on one device at a time as to not complicate things (ie. A single chasis integrated DAC & Amplifier). Yes, level matching is extremely important as the outputs will produce a vastly different ouput volume.

The single most important factor would be to direct your attention to the design, implementation and voicing of a given product.

An engineer can meticulously develop a circuit design that exhibits very similar measured response regardless of output selection.

In that same breath, an engineer can create a circuit design that has a distinctly different sound signature depending on which output is selected if so desired.

It is true, that, 'generally' the balanced topology typically exhibits a lower noise floor and distortion characteristics.

However, this does not necessarily mean that this would be the preferred selection.

Earlier, I noted the words 'circuit design'. But there is much more to it. Each and every part has an effect. From the actual board material, the type of trace material, solder, grade of resistors, quality of capacitors, etc., etc. Each of these things contribute to the overall Sound.

There have been many instances where I preferred how the engineer voiced the single ended ouput as compared to the balanced on the same product and vise versa.

There really is no right answer. There is no answer that applies to every piece of gear. Each is designed, implemented and voiced according to the engineers tastes.

You must listen with your own ears and determine which you prefer.

If the piece of equipment was well engineered, you would need the most highly resolving headphones/speakers. Forget audiophile-grade or audiophile approved recordings. You're gonna need master tapes or master recordings to decipher the microscopic differences.

If your at the point where you need to ABX test to tell a difference then you may be missing the point of enjoying your music altogether.

Don't drink the kool-aid my friend.

 

[higbvuyb]

It is true, that, 'generally' the balanced topology typically exhibits a lower noise floor and distortion characteristics.
 

 
I am led to believe that the opposite is in fact true:
 
Quote:

  Quote:

Originally Posted by KarateKid /img/forum/go_quote.gif Yeah well I just thought I could see what the fuss regarding XLR is all about. I mean isn't XLR directly off the DAC1 USB better than the HP-2 amp (the national opamp in the XLR vs the TI in the hp2)? Shouldn't it be better or good enough to use until I can get a balanced amp?

The XLR outputs on the DAC1 USB are not designed to drive headphones. The output impedance is too high (anything over 0.5 ohms is too high for driving headphones). The HPA2 headphone amp has <0.01 ohm output impedance. This is one reason it achieves ultra-low-distortion headphone amplification.

The 4562's are great line amps for the line outputs (XLR and RCA), but they are not meant to be headphone amplifiers. They will do better then the 5532's at that job, but still not ideal. The HPA2 headphone output driver is not a 5532...it is a BUF634, which is extremely well suited for driving the awkward load that headphones present.

There's a lot of discussion about balanced headphones earlier in this thread, but I'll give a brief description of the problems with balanced headphones...

ALL headphones have non-linear mechanical impedances (that is, the mass and shape of a speaker will resonate more at certain frequencies and much less at other frequencies). This means the physical build of the headphones (as well as other physical impedances, like your head and ears!) will try to override the electrical system (amplifier and speaker coil).

To create low-distortion headphone response, one must consider 'damping factor'. A high damping factor will control the response of the speaker, thus preventing the physical impedances from dictating frequency response. Damping factor is the ratio of speaker (load) impedance to amplifier (source) impedance. In other words, the best damping factor will result from a low source impedance. Again, the source impedance from the HPA2 is less then 0.01 ohms...as low as gets!!

Balanced headphone amps will double the source impedance of an unbalanced headphone amp. No matter how low the impedance of a balanced headphone amp, it could be half that much if it was unbalanced. This is one reason balanced headphone amps are not a good idea. (It should also be noted that the balanced output of the DAC1 / USB / PRE is 60 ohms or greater, depending on the attenuator settings).

Not only will the source impedance double with balanced headphone amplifiers, but the total distortion and noise of the amplifier will double as well!! Every output device (opamp, transistor, tube) creates some distortion and some noise. If there are two opamps or transistors or tubes driving each headphone speaker, twice as much distortion and noise will be added!!

The result of balanced headphones is less damping factor, more distortion, and more noise. Also, balanced headphones configurations offer no real benefits, to boot.

Feel free to use the XLR outputs of the DAC1 / USB / PRE for balanced headphone outputs (as mentioned above, the DAC1 USB and DAC1 PRE will do better then the DAC1 at this task, because of the 4562's). It won't damage anything to operate in this configuration. But, for the reasons above, I don't recommend it.

Thanks,
Elias

 
 
 
 

However, this does not necessarily mean that this would be the preferred selection.

Earlier, I noted the words 'circuit design'. But there is much more to it. Each and every part has an effect. From the actual board material, the type of trace material, solder, grade of resistors, quality of capacitors, etc., etc. Each of these things contribute to the overall Sound.

There have been many instances where I preferred how the engineer voiced the single ended ouput as compared to the balanced on the same product and vise versa.

There really is no right answer. There is no answer that applies to every piece of gear. Each is designed, implemented and voiced according to the engineers tastes.

You must listen with your own ears and determine which you prefer.

If the piece of equipment was well engineered, you would need the most highly resolving headphones/speakers. Forget audiophile-grade or audiophile approved recordings. You're gonna need master tapes or master recordings to decipher the microscopic differences.

If your at the point where you need to ABX test to tell a difference then you may be missing the point of enjoying your music altogether.

Don't drink the kool-aid my friend.

 
This could be true in theory, but it is not very useful as advice. Yes, you could treat every single product as a "black box" and rely only on what it sounds like. Unfortunately nobody has the time or patience to do this, and so human beings have been making assumptions based on the correlation between two different attributes ever since the beginning of intelligent thought.
 
Further, the beauty of a difference test is that by keeping all other factors constant, you only need to pay attention to any difference in the one factor you're changing. You don't need the most highly resolving headphones or recordings (p.s. pure sine waves are as resolving as you can get).
Neither is the point of an ABX test to resolve small differences. The point is to exclude bias to ensure the difference you're 'hearing' really is a difference in the sound.