[Willakan]

I've been trying to work out what DAC is best at a budget (up to £300) pricepoint, but have been getting increasingly confused as claims are regularly made that appear to be scientifically impossible. For example, the Benchmark DAC1 (out of my price range, but an excellent example here) is regularly referred to as bright and treble-centric - nothing in its totally flat frequency response or otherwise excellent specs suggests this.
 
In fact, looking purely at released specifications, it is unclear exactly where all these DAC sound signatures are coming from. For example, at the other end of the price spectrum, the Essence ST is regularly touted as bright and harsh by those with more expensive DACs, but I see nothing in the frequency response that would suggest this, and distortion and other specs are uniformly excellent.
 
In fact, looking at specifications, it seems increasingly difficult to distinguish all but deliberately coloured DACs. I understand that frequency response is by no means the supreme specification by which one should differentiate DACs, but apart from THD and noise, how else can I decide what's best?
 
What scientifically separates an Asus Essence ST and, say, a Yulong D100 as different classes of DAC, other than "I think one is bright and has no detail" ? The only obvious thing I could find is how more expensive products go to greater lengths to deal with jitter - is this the core explanation for the huge differences in presentation and sound signature regularly assigned to products which seem to go to great length to not have a signature of their own, yet are regularly described as bright/warm/harsh/smooth/detailed ect by the audiophile community?
 
 

 

[TStewart422]

This is easy. It's bullcrap. Placebo. People hearing the money they spent and not what they're really hearing.

 

 

[nick_charles]

Quote:

I've been trying to work out what DAC is best at a budget (up to £300) pricepoint, but have been getting increasingly confused as claims are regularly made that appear to be scientifically impossible. For example, the Benchmark DAC1 (out of my price range, but an excellent example here) is regularly referred to as bright and treble-centric - nothing in its totally flat frequency response or otherwise excellent specs suggests this.
 
The DAC1 has been measured many times and so verified as a technically superb DAC. It is as near perfect as it needs to be, it has noise/distortion levels and deviations from flat so good as to be beyond human detection. When Peter Aczel reviewed it he said ....
 
 
Quote:

[size=medium]It should be obvious from the above discussion, at least to those familiar with The Audio Critic, that the Benchmark DAC1 has no sound of its own, transparently passing on to its output the quality of its input. Whatever sonic peculiarities may perchance be audible are due to the input signal, not the DAC1 circuit. Even if the circuit were a lot less perfect, that would still be the case. Absolute sonic transparency is a concept innocent audiophiles are uncomfortable with, believing that all audio components—CD players, preamplifiers, amplifiers, tuners, all of them—exhibit varying degrees of soundstaging, front-to-back depth, grain, air, etc. That it isn’t so, except in the case of loudspeakers, is a fact calmly accepted by professional engineers but not by the high-end pundits and high-end manufacturers, who would be out of business if the truth were to sink in universally.[/size]
[size=medium]I am by now a little tired of harping on this subject but was still amused by John Atkinson’s comments on the Benchmark DAC1 in the May 2004 issue of Stereophile. John made sure to tell his readers that the Mark Levinson No. 30.6, which cost $17,500 before it was discontinued, still sounds better than the Benchmark, despite the latter’s perfect measurements. In a December 2004 followup (“2004 Editor’s Choice,” namely the Benchmark!) he adds the Theta Generation VIII ($10,000) and the Wadia 27ix ($9959) to the of-course-sounds-better list. I wonder what quality the Mark Levinson, Theta, and Wadia engineers dial into their products—above and beyond flat frequency response, low distortion, low noise, and the other usual suspects—that mysteriously makes them sound better. Maybe I should stop wondering after 28 years as an audio journalist and 57 years as an audiophile.[/size]  

 
 
But the DAC1 reaches heights that may not be necessary, see....
 
DAC Benchmark DAC1 vs Behringer DEQ2496
 
In fact, looking purely at released specifications, it is unclear exactly where all these DAC sound signatures are coming from.
 
Most of these "evaluations" come from flawed listening tests which are non level matched, prone to several types of cognitive bias and rely heavily on memory of events long passed for comparisons
 
For example, at the other end of the price spectrum, the Essence ST is regularly touted as bright and harsh by those with more expensive DACs, but I see nothing in the frequency response that would suggest this, and distortion and other specs are uniformly excellent.
 
See above but add expectation bias caused by low price and nature of device
 
In fact, looking at specifications, it seems increasingly difficult to distinguish all but deliberately coloured DACs. I understand that frequency response is by no means the supreme specification by which one should differentiate DACs, but apart from THD and noise, how else can I decide what's best?
 
FR is far and away the most important parameter, it is the basis of the fundamental sound, other things can take away from quality such as THD, IMD, Noise, and Crosstalk but all things being equal you will notice a 3db drop in midrange FR far more easily than an extra 0.005% distortion.
 
.....The only obvious thing I could find is how more expensive products go to greater lengths to deal with jitter - is this the core explanation for the huge differences in presentation
 
 
NO ! or highly unlikely anyway, I'd bet on winning the lottery first, there is no evidence that jitter in the levels found in moderately competent digital audio kit is in any way audible in musical playback, anti-jitter device manufacturers have not provided any controlled listening evidence for the value of jitter reduction (beyond absurd levels) and the few peer reviewed papers on this (Ashihara et al 2005, Benjamin and Gannon, 1998 ) have not supported audibility of jitter in the sub ns magnitude regardless of type or spectra.
 
Anecdotes abound but real evidence = zero, jitter is just not something to worry about
 
 

 

 

[Willakan]

So after frequency response and THD+N there really is nothing else significant to it? To use my earlier example products, what separates the line out of an Essence ST from something like the Benchmark? They both have a very flat frequency response in the audible range (around -0.2db falloff at 20khz in both cases), both have excellent THD+N results (in the region of 0.0003-0.0005%) and both have excellent SNR results, the Essence actually having the edge in this regard. (-120db vs -114db)
 
Seeing that Benchmark produce professional audio equipment, I'm assuming the Benchmark unit is not so very well regarded because it looks pretty; likewise, the Essence is not the best bargain in audio by around 600 dollars. What's the factor here that results in audible differences?

 

[EthanWiner]

Wow, great thread. Yes, claims of "bright" sound when a frequency response graph is flat proves how silly this stuff can get.
 
Quote:

So after frequency response and THD+N there really is nothing else significant to it?

Only four parameter categories are needed to define everything that affects audio fidelity:
 
Frequency response
Distortion
Noise
Time-based errors
 
These are described in detail here:
 
Audiophoolery
 
--Ethan

 

[Willakan]

Thanks for the article link - as it happens, I've already read it - should be required reading for every new Head-Fi member before they start buying cable burn in CDs and platinum plated USB cables.
It is refreshing to find a section on Head-Fi full of sane people - ironic that the sound science forum is the least popular major section - people are apparently far more interested in learning which utterly useless cable to buy over another, entirely different, yet equally useless cable. Science seems to have fallen out of fashion. And don't get me started on the anti-placebo comments (not sure if placebo is exactly the right term, but it's clear to what I refer):
 
Person A: What you supposedly heard from changing a usb cable/wearing purple socks/praying to your cat is impossible, so it must be the placebo effect.
Person B: BUT I HEARD IT!!! THEREFORE IT MUST BE REAL!
 
Righto, before I derail my own thread, back on topic. I found Asus' complete set of measurements for the Essence ST: 
http://files.computeraudiophile.com/2010/1217/e4253_Xonar_Essence_STX_Audio_Test_Report.pdf
 
I am now even more confused than before. Here are the Benchmark specifications:
http://www.benchmarkmedia.com/sites/default/files/documents/dac1-usb_spec_sheet_0.pdf
 
To the uninitiated, they both seem to perform equally well from the line out - could someone with more knowledge than me on the matter chime in to see how close these numbers actually are and whether they are directly comparable?
 

 

[nick_charles]

Quote:

Wow, great thread. Yes, claims of "bright" sound when a frequency response graph is flat proves how silly this stuff can get.
 

Only four parameter categories are needed to define everything that affects audio fidelity:
 
Frequency response
Distortion
Noise
Time-based errors
 
What about crosstalk, channel imbalance and phase distortion ?

 

 

[jcx]

I assume Nick knows most of this but for general education purposes:
 
crosstalk is a easily measured # that apparenty has little audible significance at levels easily achieved by DACs: world class phono cartridges often just have "better than 30 dB" spec for crosstalk, even closed back circumaural cans have leakage at -50-60 dB levels, electronics can easily reach/exceed 100 dB channel isolation/crosstalk - TRS common gnd contact ~3-10 mOhm series R gives bigger crosstalk with low Z cans
 
channel imbalance in DAC chips can be ~1% internally, few audio equipment use better than 1% tolerance resistors anywhere in the signal chain - so yes this number could be trimmed but it will normally be as good as any other audio equipement - touted HD650 driver matching is +/-1 dB ~ 10%
 
phase error is a "time domain" error - analog anti-imaging filters are more variable than the DAC's internal digital filtering and the analog filter design/implementaion is easier with oversampling DACs - there are claims of "linear phase" FIR digital filter pre-ringing being a cause of audible differences between higher sample rate systems and RedBook - but convincing, replicated DBT evidence is lacking
 
in loudspeaker crossovers Linkwitz-Riley LR4 with driver phase inversion are widely accepted and have ms group delay nonlinearity - as long as phase "errors" are small, matched in L,R channels it is another fairly low audibility isssue - in DACs the R/L differential phase errors can be sub us

 

[EthanWiner]

Quote:

What about crosstalk, channel imbalance and phase distortion ?

If you read the article you'll see those are addressed. Crosstalk falls under noise, and phase shift is part of time-based errors. Channel imbalance isn't really a parameter. Good point though! I'll have to think about that further.
 
--Ethan

 

[Prog Rock Man]

Welcome to the back of the bus Willakan

. Asking detailed and challenging questions which are evidenced based is sadly, not that popular on this and indeed most other audio/hifi forums.
 
The lack of consistency between 'measurements; and sound differences/quality is something that has been raised a number of times with cables. It is great that you have taken the time to do the same with DACs.
 
I suspect that whilst on the digital domain, unless a signal is corrupted in anyway and so breaks up, all 'sound' the same. Finding no consistency between 'sound' and DAC is evidence to suggest that that is the case.

 

[Willakan]

In the interests of comparing some popular DACs (some generally, others around here) I have put together this spreadsheet comparing the main attributes of various popular products. It probably has a mistake in it somewhere and by no means is a comprehensive evaluation of all the DACs.

What I find interesting is firstly how well the Essence holds up - apparently not needing to bother with an enclosure and huge mass production allowed ASUS to really go all out in designing this thing for decent audio. The Yulong also doesn't embarrass itself alongside the DAC1, from which it copies most of the feature set and a large chunk of the jittter reduction techniques, at a far smaller price. 
The thing that caught my eye is how the DACMagic measures. I was considering getting instead of the 'lowly' Essence, but it appears you pay for the nice metal case and the plethora of inputs. The only comparison between the two I found on these forums concluded the DACmagic was a major upgrade from the Essence though!  
So I suppose the question this time around, aside from confirming this table is A) Vaguely useful and B) Mostly right, is what other strange components of audio could mean that people perceive improvements between the DACMagic and the Essence (aside from cognitive bias ect) or does DAC pricing rely to a large extent on casing and I/O options, rather than absolute sound quality?
 
 

 

[Shike]

In the interests of comparing some popular DACs (some generally, others around here) I have put together this spreadsheet comparing the main attributes of various popular products. It probably has a mistake in it somewhere and by no means is a comprehensive evaluation of all the DACs.

What I find interesting is firstly how well the Essence holds up - apparently not needing to bother with an enclosure and huge mass production allowed ASUS to really go all out in designing this thing for decent audio. The Yulong also doesn't embarrass itself alongside the DAC1, from which it copies most of the feature set and a large chunk of the jittter reduction techniques, at a far smaller price. 
The thing that caught my eye is how the DACMagic measures. I was considering getting instead of the 'lowly' Essence, but it appears you pay for the nice metal case and the plethora of inputs. The only comparison between the two I found on these forums concluded the DACmagic was a major upgrade from the Essence though!  
So I suppose the question this time around, aside from confirming this table is A) Vaguely useful and B) Mostly right, is what other strange components of audio could mean that people perceive improvements between the DACMagic and the Essence (aside from cognitive bias ect) or does DAC pricing rely to a large extent on casing and I/O options, rather than absolute sound quality?
 
 

You have to consider that many people are driving directly out of the Essence into headphones, and the Xonar cards like the STX are known for high output impedance which can impact frequency response. Furthermore, not every installation will have the same specs with the Asus in reality (computers are noisy environments and not identical).

That said, I've been all over the chain in terms of DACs and can honestly say that the non-proven high-end is at best equal if not worse than some lower priced gear. I started with an Entech DAC which worked quite well honestly, but after being talked into buying a NOS DAC by various sources of conflicting interest I purchased a Shek D1. This . . . was a silly mistake. It was $300 for a sub-par circuit that anyone could throw together on a breadboard in an hour or so. I let it slide under the guise of being "musical", but in reality it was just a bad piece of ancient tech sold in a shiny case at a high price. Eventually I got a Cary Audio Xciter . . . and found out it too measured quite poorly when testing it via RMAA. I decided to be smart, cut my losses and sold it for $900 or so. Eventually I started digging into measurements, and I think the DacMagic was probably the best balance if you didn't use its USB being fairly comparable in a few fronts to the DAC1 (at least as far as jitter was concerned, and it measured passed audibility without question in all other fields).

Bit the bullet, bought one new for $375 shipped and haven't looked back since. The remaining money from the $900 went into equipment that actually makes a substantial difference.

 

[Uncle Erik]

Willikan, would you feel better about yourself if you were wearing a $50 jacket or a $1,000 jacket? Assume both offer equal warmth, protection from the elements and durability.

The difference between the two jackets is fashion. Audio is exactly the same.

People pay for looks and label, and that's really all you need to know.

The people selling this stuff will start the furious hand waving and uttering angry cries, but you really don't need to spend much to get quality audio today. I think most manufacturers accept this. Which is why they've turned this into a fashion show. In fashion, you can play games with what is "in" and "out," causing people to replace things that are still perfectly fine.

Kinda sad, though, because high-end audio is really inexpensive today.

 

[Currawong]

From my experimentation with various pieces of gear, the "bright" comments may be the result of noise or distortion of some sort in the treble.  Plugging various pieces of gear into a power filter with a measured attenuation of noise removed this.  It would be interesting to see how the measurements of each amp or DAC I tried this with are affected.  Related to that, an interesting comment from a DAC maker I read was that he said he could tell how a DAC would sound by its ability to reproduce a square wave.  A lot of gear can't reliably do this (depending very much, I imagine, on the input impedance of what it was connected to) so that may be another factor that doesn't show up in regular RMAA tests.  I think a lot of you are looking at manufacturer's specs or RMAA tests and assuming you've seen all there is to see.

 

[nick_charles]

 
Quote:

From my experimentation with various pieces of gear, the "bright" comments may be the result of noise or distortion of some sort in the treble.  Plugging various pieces of gear into a power filter with a measured attenuation of noise removed this.  
 
While one would assume that manufacturers would test their kit under optimal conditions I think electrical noise may frequently be over-rated as a problem, I inhabit a 1948 built house with the original ungrounded wiring, it is a disaster zone electrically and our light-bulbs blow with alarming frequency, sometimes when the air con switches off it is impossible to get the kitchen light to switch on, ( we really should get the house rewired)  many of our power strips are bodged with cheater plugs, despite all of this we do not have a notable noise problem on any of our many audio systems. Perhaps we are lucky.
 
I would also ask for better evidence of the effect of power filters on fundamental audio parameters, measured noise reduction is one thing but I would want to see measured differences in FR and so on, plus of course some DBTs showing that the electrical noise reduction was audible in music as opposed to listening to digital silence where it would be easy to notice...
 
It would be interesting to see how the measurements of each amp or DAC I tried this with are affected.  
 
Agreed !
 
Related to that, an interesting comment from a DAC maker I read was that he said he could tell how a DAC would sound by its ability to reproduce a square wave.  
 
Above 5k (I think) a square wave and a sine wave of the same frequency are perceptually indistinguishable, it is impossible for a DAC to render a perfect square wave due to the whole infinite harmonics thing, square waves do not occur in nature and seldom in music, they are useful things like testing rise time i.e getting an idea of transient response, but it is a bit of a parlor trick, it is the sort of trick vinylies use to prove that LP is better than CD after all

 
A lot of gear can't reliably do this (depending very much, I imagine, on the input impedance of what it was connected to)
 
so that may be another factor that doesn't show up in regular RMAA tests.  I think a lot of you are looking at manufacturer's specs or RMAA tests and assuming you've seen all there is to see.