What's an example of a "good DAC"?

[pibroch]

111

Ok, then I'll refine the statement.

All DACs designed to reproduce a signal accurately and without deliberate or unintentional modification of the signal sound the same. That should be most of them. And those that sound different should be very easy to characterize with measurements, and IMO any DAC with an unusual or signature should be avoided unless the intent of that modification is fully recognized and applied to it's intended purpose.

Geez, I have to write like a lawyer here.

Also, DAC quality doesn't track cost at all. The DACs in iPhones and iPads, for example (the ones that still have a headphone jack anyway), are some of the best, and technically free with the device.

Top recording studios all over the world are dumping their stand alone dacs and mixing and mastering their top recordings on iPhones.

 

[pinnahertz]

Its not the loudness which makes the difference, but how 3D the sound stage is, how analog and smooth and detail well extended. Plus how well the interface is coded and integrated for the DAC.

So in your tests you've actually level-matched all DACs being tested?

 

[castleofargh]

Its not the loudness which makes the difference, but how 3D the sound stage is, how analog and smooth and detail well extended. Plus how well the interface is coded and integrated for the DAC.

variations in volume levels are known to often be confused by the listener for variations in perceived quality. it's the very reason why volume matching within 0.1dB has been introduced in listening test of gears. you not knowing that it's necessary to a proper listening test is different from it not being necessary.

 

[spruce music]

Its not the loudness which makes the difference, but how 3D the sound stage is, how analog and smooth and detail well extended. Plus how well the interface is coded and integrated for the DAC.

Job #1 in any useful comparison is precise level matching. Small level differences not obviously heard as different volumes create the impression of sound quality differences. Level matching even in sighted comparisons often reduces perceived differences by a large amount. Blinding usually makes them all go away. And yes, 3D soundstage, sense of space, air, smoothness and detail, can all be due to different levels of playback.

So match it up or really it isn't worth wasting your time. Unmatched comparisons are meaningless unless the device is so non-linear as to be a broken design.

 

[ShaggyFi]

Job #1
So match it up or really it isn't worth wasting your time. Unmatched comparisons are meaningless unless the device is so non-linear as to be a broken design.

variations in volume levels are known to often be confused by the listener for variations in perceived quality. it's the very reason why volume matching within 0.1dB has been introduced in listening test of gears. you not knowing that it's necessary to a proper listening test is different from it not being necessary.

So in your tests you've actually level-matched all DACs being tested?

Here is how I get my impressions and why I think it makes sense:

I have 3 DAC(HUGO, M8, MODI) at once for about 6 months, avg listening for 2 hours per day. There was one week in may when I switched between the 3 DAC, tested them and then listed one for sell, one gave away as gift, one for myself. I tested then with the same source(gaming notebook), input usb cables with jitter bug usb cleaner, output RCA cable, wall power, headphone AMP(jotunheim high gain single end) and headphone(HE500) and the same software setting(JRiver exclusive mode) and music files Beethoven Symphonies 5 & 7 - Honeck, Pittsburgh SO (Reference FR-718).

The 3 different DAC sound like 3 different DAC. The modi has the most to-your-face soundstage while all instruments are tied together and the rear instruments(basses, horns and trombones etc...) don't sound like on the back, but only sound like being covered by the front due to lack of details and their harmonic overtone got overwhelmed by high frequencies noise from pc. Modi also makes everything sounds much brighter than live perform. M8 has a more realistic brightness. The rear instruments don't get overwhelmed by the front too much while airy details from the front are still there. Its overall a more 3D, not overly bright sounding DAC. Therefore if the instruments are brighter than they should but instruments are still not well separated, then I must be listening to a MODI not a M8. Because their harmonic overtone sound like being overwhelmed by the brightness of all the fundamental sound of the instruments. When the high frequencies details and harmonic got messed up, the soundstage got messed up too. The Hugo is a larger jump again. Each instrument sounds more like what the sound in real life. The fundamental sound of the instruments are more detailed and weighted, without disturbing restless noise from PC which makes everything sounds too bright and out of place. Therefore the harmonic overtone is all there, its warm and airy and all instruments own a body on the stage. Maybe thanks to the well written Xmos interface and clean battery power or something in the HUGO, the shiiit PC noise do not get in my music as much as the 2 other dac while having good detailed sounding chip and stuff. Hugo simply tastes better.

I have to say all 3 DAC I have are not for symphonies because they all have the to-your-face soundstage compares to full size pro DAC. However they are still different enough for me to tell which is which. Back to the volume matching issue, if the violins sound too bright, no matter how you turn the volume up and down. You wont make it natural wood sounding again. If the orchestra instruments are well separated, there won be noise to jam the airy details no matter how you turn the volume up and down.

 

[ShaggyFi]

Top recording studios all over the world are dumping their stand alone dacs and mixing and mastering their top recordings on iPhones.

indeed

 

[spruce music]

Here is how I get my impressions and why I think it makes sense:

I have 3 DAC(HUGO, M8, MODI) at once for about 6 months, avg listening for 2 hours per day. There was one week in may when I switched between the 3 DAC, tested them and then listed one for sell, one gave away as gift, one for myself. I tested then with the same source(gaming notebook), input usb cables with jitter bug usb cleaner, output RCA cable, wall power, headphone AMP(jotunheim high gain single end) and headphone(HE500) and the same software setting(JRiver exclusive mode) and music files Beethoven Symphonies 5 & 7 - Honeck, Pittsburgh SO (Reference FR-718).

The 3 different DAC sound like 3 different DAC. The modi has the most to-your-face soundstage while all instruments are tied together and the rear instruments(basses, horns and trombones etc...) don't sound like on the back, but only sound like being covered by the front due to lack of details and their harmonic overtone got overwhelmed by high frequencies noise from pc. Modi also makes everything sounds much brighter than live perform. M8 has a more realistic brightness. The rear instruments don't get overwhelmed by the front too much while airy details from the front are still there. Its overall a more 3D, not overly bright sounding DAC. Therefore if the instruments are brighter than they should but instruments are still not well separated, then I must be listening to a MODI not a M8. Because their harmonic overtone sound like being overwhelmed by the brightness of all the fundamental sound of the instruments. When the high frequencies details and harmonic got messed up, the soundstage got messed up too. The Hugo is a larger jump again. Each instrument sounds more like what the sound in real life. The fundamental sound of the instruments are more detailed and weighted, without disturbing restless noise from PC which makes everything sounds too bright and out of place. Therefore the harmonic overtone is all there, its warm and airy and all instruments own a body on the stage. Maybe thanks to the well written Xmos interface and clean battery power or something in the HUGO, the shiiit PC noise do not get in my music as much as the 2 other dac while having good detailed sounding chip and stuff. Hugo simply tastes better.

I have to say all 3 DAC I have are not for symphonies because they all have the to-your-face soundstage compares to full size pro DAC. However they are still different enough for me to tell which is which. Back to the volume matching issue, if the violins sound too bright, no matter how you turn the volume up and down. You wont make it natural wood sounding again. If the orchestra instruments are well separated, there won be noise to jam the airy details no matter how you turn the volume up and down.

I understand this methodology, and how it seems so right.

I repeat: Job #1 is matching levels. Not by ear either. By measuring with a test tone and getting outputs within 1.5% of each other (or .1 db). Otherwise you will lead yourself astray without knowing it. All these flowery descriptions of depth and in your face sound and 3D quality can often disappear with level matching. Best unsighted, but even sighted matching levels is simply essential. Let me re-emphasize that. Matching level is

ESSENTIAL.

In comparative listening it actually isn't even worth doing if you don't match levels. Yes, I know, have been there and done that and it seems like your descriptions are solid, and real and no need to bother. Bother to do it. Level matching is essential and step number one. Matching wood sound to natural sound blah, blah, blah (I really am not picking on you personally just trying to get something important across). Match levels when you compare things.

 

[old tech]

Transparent DACs are ubiquitous these days, as others have noted, even the humble Iphone.

While the high end DAC is the poster child of some digital audiophiles, they belong in the same category as what tube amps are to some audiophiles. If these DACs do sound different from their generic cousins it is only because they are distorting the sound. Subjectively some listeners do prefer the distorted sound, whether it be a top end roll off or unevenness in frequency response to make it sound more like analog, that is fine if that floats your boat but it is departure from true high fidelity. Then there is the issue that there may be timing errors or jitter with external DACs which is non-existent with most internal DACs.

 

[ShaggyFi]

Matching level is

ESSENTIAL.

Sure thing, Its very important. Human ears like to hear things loud and clear. harsh thin bright sound does becomes more enjoyable when lowing the volume, and loosy muddy sound does becomes more enjoyable when increasing the volume. However when I decreased the modi volume, it was still way thinner brighter than the other 2 DAC. When I in creased the volume, modi was again less instrumental separated than the other 2. Same goes for decreasing Hugo's volume, the energy and smoothness of the music remain, but less enjoyable than normal volume tho.

 

[Strangelove424]

A test without falsifiability is an exercise in confirmation bias. There has to be a way of being wrong. If you have two DACs in front of you, and can actually see the source, there is no falsifiability, no matter whether you think you hear a difference or you don't think you hear a difference. Either way, you're "right" because there's no way to prove yourself wrong. Unless you use a (level matched) ABX. I don't personally like ABX tests, I think they're a pain in the ass, and there's more objective ways than ABXing to measure phenomenon directly. But as a measurement of what is perceivable, it's the falsifiability component - the ability to confirm that you can or can't hear something, rather than just believe yourself - that makes ABX so critical.

The power of the mind is underestimated. You can focus on one part of the song, lets say the kick drum, and your brain will zero in on that part of the freq band. You'll perceive that as more articulation in the bass. Now you go to headphone/DAC #2, and start focusing in on the same kick drum as the diligent ABXer you are, but this time your mind focuses on the initial high frequency element of the drum strike instead. Confirmation bias will start kicking in, and you'll think that headphone/dac #2 has higher frequency emphasis. If you think I am crazy, then you haven't done an ABX test yourself. It's only when the listening is done blindly, and when results can be falsifiable that you'll find the error in your perception. Most people don't like to think that there minds can be easily led astray, but that's the ego talking. ABX neutralizes the ego. That first moment when you realize that the thing you heard was just an emphasis in your head, and then the second moment, and third, and on and on... until you feel statistically like you're at the crap tables in Vegas. It's like cold water to the face. It's essential for honest appraisal. We are best at lying to ourselves.

You don't hear the term "DAC" and "ABX" thrown around together very much. Hell, you don't even see detailed measurements often either. For such an essentially technical element in the chain, that's strange to me.

 

[spruce music]

Sure thing, Its very important. Human ears like to hear things loud and clear. harsh thin bright sound does becomes more enjoyable when lowing the volume, and loosy muddy sound does becomes more enjoyable when increasing the volume. However when I decreased the modi volume, it was still way thinner brighter than the other 2 DAC. When I in creased the volume, modi was again less instrumental separated than the other 2. Same goes for decreasing Hugo's volume, the energy and smoothness of the music remain, but less enjoyable than normal volume tho.

I don't know if such a thing happened. However, I have experienced it and seen it in others over and over and over and over.

They listen to something sighted. Without any intent or realization their mind has pegged the relative sound when doing comparisons. It may have been due to loudness differences making harshness more apparent on one vs the other or just the appearance of the box or something they had been told. Hidden in their awareness is the description of the differences. After that changing volume, changing headphones or other gear that biased idea follows the gear. Match levels and cover the gear up and it all disappears into thin air. If it doesn't disappear it was real.

I'm giving you good advice I was hard headed about for years with all the same excuses and some you haven't brought up.

Do yourself a favor and consider matching levels a necessity. Because it is if your comparisons are to have good validity. Not with me. Who cares what I think. With matching genuine physical reality. You can't hoodwink that.

 

[spruce music]

A test without falsifiability is an exercise in confirmation bias. There has to be a way of being wrong. If you have two DACs in front of you, and can actually see the source, there is no falsifiability, no matter whether you think you hear a difference or you don't think you hear a difference. Either way, you're "right" because there's no way to prove yourself wrong. Unless you use a (level matched) ABX. I don't personally like ABX tests, I think they're a pain in the ass, and there's more objective ways than ABXing to measure phenomenon directly. But as a measurement of what is perceivable, it's the falsifiability component - the ability to confirm that you can or can't hear something, rather than just believe yourself - that makes ABX so critical.

The power of the mind is underestimated. You can focus on one part of the song, lets say the kick drum, and your brain will zero in on that part of the freq band. You'll perceive that as more articulation in the bass. Now you go to headphone/DAC #2, and start focusing in on the same kick drum as the diligent ABXer you are, but this time your mind focuses on the initial high frequency element of the drum strike instead. Confirmation bias will start kicking in, and you'll think that headphone/dac #2 has higher frequency emphasis. If you think I am crazy, then you haven't done an ABX test yourself. It's only when the listening is done blindly, and when results can be falsifiable that you'll find the error in your perception. Most people don't like to think that there minds can be easily led astray, but that's the ego talking. ABX neutralizes the ego. That first moment when you realize that the thing you heard was just an emphasis in your head, and then the second moment, and third, and on and on... until you feel statistically like you're at the crap tables in Vegas. It's like cold water to the face. It's essential for honest appraisal. We are best at lying to ourselves.

You don't hear the term "DAC" and "ABX" thrown around together very much. Hell, you don't even see detailed measurements often either. For such an essentially technical element in the chain, that's strange to me.

Plus 100 to everything he just wrote.

 

[ShaggyFi]

I don't know if such a thing happened. However, I have experienced it and seen it in others over and over and over and over.

One of the most important factor in my test which makes my 3 DAC sound so different is definitely not the volume, but my very poor source(gaming notebook). I can't believe none of you point it out yet. Its has tones of jamming noise passses to the DAC. Since my modi is USB powered, it only amplifies the weakness of the DAC and make the already not very revealing chip even worse. I do believe the DAC is not what makes the music really bad but the tech failed to handle the noisy power is. I actually used a high end 204 interface between the PC and DAC. That proved what I thought. PC + int204 + modi uber sounds much much better than PC + Hugo/M8/modi/modi uber. However the int204 costs more than the rest of the system.

 

[pinnahertz]

A test without falsifiability is an exercise in confirmation bias. There has to be a way of being wrong. If you have two DACs in front of you, and can actually see the source, there is no falsifiability, no matter whether you think you hear a difference or you don't think you hear a difference. Either way, you're "right" because there's no way to prove yourself wrong. Unless you use a (level matched) ABX. I don't personally like ABX tests, I think they're a pain in the ass, and there's more objective ways than ABXing to measure phenomenon directly. But as a measurement of what is perceivable, it's the falsifiability component - the ability to confirm that you can or can't hear something, rather than just believe yourself - that makes ABX so critical.

There are more objective ways to measure, because all measurement is objective, ABX is subjective. This seems like being pedantic, but it's very critical to understand the difference.

You don't hear the term "DAC" and "ABX" thrown around together very much. Hell, you don't even see detailed measurements often either. For such an essentially technical element in the chain, that's strange to me.

Not hard to understand, really. To ABX a pair of DACs takes a hardware ABX comparator, something most people don't have access to as it is a fairly expensive device. Measurement of a DAC requires a measurement system with better resolution than the DAC, also non-trivial, and quite expensive. That's why you don't see much of either.

 

[Strangelove424]

There are more objective ways to measure, because all measurement is objective, ABX is subjective. This seems like being pedantic, but it's very critical to understand the difference.

In this instance not completely pedantic, as it's a line between science and statistics. I regard ABXing as using objectively controlled environments (and to that extent some measurement might be involved) to set up statistical analysis of subjective perception. It is a systematic but not scientific way of measuring subjective perception, a statistically significant and far more valid way of saying "I heard". What exactly one might be hearing, which I think is a far more important question from an engineering and consumer purchasing standpoint, would require rock solid measurements. There's examples of people being "successful" at ABX tests because of a non-linear indicator like intermodulation distortion, which was mistaken for high frequency extension. So even if you hear it, that still doesn't necessarily mean you understand what you're hearing, or that you should buy it.

Not hard to understand, really. To ABX a pair of DACs takes a hardware ABX comparator, something most people don't have access to as it is a fairly expensive device. Measurement of a DAC requires a measurement system with better resolution than the DAC, also non-trivial, and quite expensive. That's why you don't see much of either.

When it comes to hobbyists, yes, ABXing is resource intensive, often impossible, and just generally a pain to do even if you can. But I don't think the audiophile press, who drive so much of the hype, have those excuses, especially with all the review gear being given away to them. Unless a stipulation of being given review gear is making no blind comparison. Which is probably the case. The high end audio market thrives in the grey nebulous of subjective impression. It is exactly the lack of falsifiability that gives them an open license to make whatever claims they want. Same for the manufacturers of the products. If Stereophile published this kind of an ABX-based review here (http://www.tomshardware.com/reviews/high-end-pc-audio,3733.html) it would put their advertisers out of business, and eventually themselves. I get the business model. But the strange part to me is why the paying market goes along with this, and doesn't demand more verification for their dollars, atleast measurements. Consumers always have the ultimate power, but "vote with your wallet" doesn't seem to be a thing in the audiophile market, sadly. This ultimately frustrates and alienates value-oriented shoppers who feel that their purchasing power is being diluted by the underperforming, ruthlessly inflated "high end". I am often hearing $2,000 amps/dacs nowadays refered to as "great budget values". It's gotten ridiculous.

For anyone who believes in DAC differences, or is just interested in a value-oriented perspective from some computer geeks, please read that Toms Hardware article. The $2k Benchmark DAC they used was not a loaner, it was the reviewer's own, so he had some skin in the game too!