DIY amp/DAC with better measurements than O2/ODAC?

[MHOE]

Question is quite simple... I am trying to get more information about available amplifiers and DACs and as a very happy desktop O2/ODAC combo owner I am curious if there are any other great-measuring amplifiers and DACs available on the DIY scene. They must be neutral and technically excellent (more than O2 and ODAC), of course!

Benchmark DAC2 HGC looks exciting but the price is high, especially in Europe...

Thanks for any information!

 

[opc]

Hi MHOE,

As a shameless self plug, you could take a look at the headphone amps I offered under the "The Wire" moniker.

They are similar to the O2 in that they were designed from a very technical/objective perspective, but unlike the O2, they are a cost-no-object based design. That's not to say that they are very expensive, just that performance always came first in the design.

They do measure significantly better than the O2 headphone amp, and measurements are included in the thread. They also deliver significantly more output power and better drive into low impedance loads.

I've got an ES9012 based DAC in the works that will also measure quite a bit better than the ODAC, but it's probably another few months off.

Regards,
Owen

 

[MHOE]

Hi MHOE,

As a shameless self plug, you could take a look at the headphone amps I offered under the "The Wire" moniker.

They are similar to the O2 in that they were designed from a very technical/objective perspective, but unlike the O2, they are a cost-no-object based design. That's not to say that they are very expensive, just that performance always came first in the design.

They do measure significantly better than the O2 headphone amp, and measurements are included in the thread. They also deliver significantly more output power and better drive into low impedance loads.

I've got an ES9012 based DAC in the works that will also measure quite a bit better than the ODAC, but it's probably another few months off.

Regards,
Owen

Thank you very much for your recommendation, Owen! Are you planning to release your work as a single amp/DAC combo unit? And do you have a rough estimate about its price?

 

[jcx]

the O2 design explicitly adopts a "good enough fo audio" at a low price philosophy - not intended as a exercise in pushing specs to technology limits

likewise Vout DAC chips are "good enough" by his criteriea but most flagship monolithic audio DAC chips are Iout - and I/V conversion keeps analog circuit designers and tweakers happy with more scope for external circuit design arguments

 

[BuildMeSomething]

Do like the Benchmark units, but at £1000+ they are a touch expensive... More than happy with the sound of the ODAC, especially feed into one of opc's The Wire se-se. The O2 sounded not to far off The Wire, then I figured out I'd messed up the input side of The Wire. That small difference just opened up to Gand Canyon size proportions once sorted.

@ Owen. Is the DAC your working on to be a commercial/pre-built offering, or possible another addition to "The Wire" stable?

 

[opc]

Thank you very much for your recommendation, Owen! Are you planning to release your work as a single amp/DAC combo unit? And do you have a rough estimate about its price?

It will most likely be two separate units... one headphone amp and one DAC.

There's just too much variability on the headphone side to keep everyone happy with an integrated unit. Some want fully balanced output and input, some want SE, some need more power for planars, and some prefer lower noise and much less power for IEMs. Some want to use the head amp for vinyl or other analog sources.

The cost is also still a little up in the air at the moment. The original version of "The Wire" could be built for a little less than $80 and I will probably keep that option open (selling bare PCB's) but a fully assembled, tested, cased and "tweaked" unit in a case will likely be a few hundred.

the O2 design explicitly adopts a "good enough fo audio" at a low price philosophy - not intended as a exercise in pushing specs to technology limits

likewise Vout DAC chips are "good enough" by his criteriea but most flagship monolithic audio DAC chips are Iout - and I/V conversion keeps analog circuit designers and tweakers happy with more scope for external circuit design arguments

All very true, and every type of design has its place. It's hard to fault the ODAC or O2 in any way as they are very well documented, well designed, and very low cost for the performance you get. The only place I tend to disagree with the design is the overwhelming influence of price, and the fact that performance is indeed given up to save what amounts to the price of a coffee in the morning. It's quite a lot of trouble to build your own gear, so I feel like if you're going to put in the effort, you might as well spend a tiny bit more and know that you've got the best possible performance that can possibly be achieved. That's just my take on it though, and the "good enough" argument is also pretty solid.

Cheers,
Owen

 

[Ken Newton]

It's quite a lot of trouble to build your own gear, so I feel like if you're going to put in the effort, you might as well spend a tiny bit more and know that you've got the best possible performance that can possibly be achieved. That's just my take on it though, and the "good enough" argument is also pretty solid.

Cheers,
Owen

I second your stated philosphy here. Commercial vendors often have to count dimes and quarters, but any DIY project worth the effort should not be subject to such coin pinching constraints.

 

[xnor]

The thing is many diy attempts at building an excellent amp have resulted in worse performance and higher cost than the O2.

The "highest performance and cost is no object" philosophy only works if you can objectively quantify performance improvement due to whatever change.

 

[Ken Newton]

Cost can be an object

I don't think he's so much saying that cost isn't an object, but more that the lowest possible cost should not become an objective itself.

 

[xnor]

Either way, what I wrote still stands.

Nwavguy posted: "Unless you have really rare headphones that need more than 7 Vrms, or more than 200 mA peak current, the LME49600 has no audible advantages over the O2 design."

 

[opc]

I don't think he's so much saying that cost isn't an object, but more that the lowest possible cost should not become an objective itself.

Couldn't have put it better myself!

xnor:

You are correct that there are an alarming number of designs out there which just jam a bunch of expensive bits together and call it "cost no object" but that's really missing the point.

As Ken said, the driving force in a design should really be performance, and when you start using an AP or other proper measurement methods, you quickly realize that spending several hundred dollars on boutique capacitors and fancy wiring is not really the path to a better end product. The point of cost no object is spending your money where it provides the largest tangible gains, and if given the option between a good solution, and a superb solution (objectively) you pick the superb solution even if it costs a little more.

I know everyone has differing opinions on what constitutes "good value for money" but you are correct in stating that people often do spend a lot and get very little when it comes to DIY tinkering, and especially untested/unmeasured "mods" to gear that actually has been properly designed and tested

You'll find that all my designs use good tightly matched parts (like 0.1% thin film SMD resistors and good X7R or C0G ceramic caps) but nothing exotic, and pretty much never anything with "audio" written on it. I also measure all changes carefully, and make sure that if I'm doing something, there's a sound technical reason, and measurable performance advantage.

Cheers,
Owen

 

[opc]

Either way, what I wrote still stands.

Nwavguy posted: "Unless you have really rare headphones that need more than 7 Vrms, or more than 200 mA peak current, the LME49600 has no audible advantages over the O2 design."

Statements like that are a little scary, because they're decidedly not objective.

I always get a little chill when someone says "well, sure you can measure it, but you can't hear it, so don't bother"

I intentionally avoid commenting on exactly what people can and can't hear because I have no tangible proof of what the human ear is capable of. I only know what my ears are capable of. I have a general idea of what is audible and what is not, but having no proof, I keep my opinion to myself.

I generally believe that if you're passionate about getting good sound, you should strive for the most technically superior solution to the given problem. If you head down the path of "good enough" then you might as well just head over to BestBuy and pick up an ipod with a decent set of headphones and call it a day.

Cheers,
Owen

 

[counter culture]

Heading down the path of 'good enough' might save you a lot of money.

IIRC nwavguy has done some A/B testing with a number of subjects who were unable to distinguish between his amp and the Benchmark DAC1's amp. A/B testing qualifies as objective. If the O2 is transparent then the LME49600 has no audible advantage over the O2 in the use cases quoted and he is entirely justified in making the statement he did.

Tangible proof of what the human ear is capable of hearing is obtainable, you just haven't taken the trouble to obtain it.

When it comes right down to it you're just trying to sell your amp by using the same technique of instilling 'fear, uncertainty and doubt' that the worst of the subjectivists employ. You're probably unaware of doing it, I can understand how easy it is to slide into the example set so commonly, and it's difficult not to be partisan in these circumstances.

Sell your amp on the basis that it's technically superior, not that it might sound better. It might not, and what evidence there is suggests it won't, it will just sound the same.

I genuinely believe that if you are only passionate about getting good sound you should pick the lowest cost solution that has proved transparent in A/B testing, or for that matter, exceeds the likely known limits of the perceptible. If you want the intellectual satisfaction of knowing you have a technically superior solution, by all means pay for it, but be aware that the scientific evidence, and all common sense, suggests that there is a limit to perception and it makes no audible difference whether you exceed it by a little, or by a lot. Certainly there are those who cannot reconcile themselves to this idea, but to deliberately evoke doubts is just salesmanship, flies in the face of what is known about hearing and attempts to take advantage of a weakness of character.

 

[TheShaman]

Nwavguy did a good job on the O-* family of products, but he kinda looses...objective marks when he suggests he has reached some kind of performance sweet-spot where anything worse is a "compromise" and anything better is "needlessly good".
If such a sweet-spot does exist, I'm sure all of us would be interested to hear (pun intented) more about it. But, since we're all "happy objective campers", as it seems, proof would have to go well beyond vague estimations of performance parameters and anecdotal claims of A/B tests against a Benchmark unit.... I'm sure it'll be a hell of a seminal paper and I'll be happy to read it once Nwavguy decides to follow proper scientific procedures, document his work and get it published. I know at least two other guys, i.e. Floyd Toole and Sean Olive, who will be happy to read it as well.

In any case, the decision of where to set the "good enough" threshold is one the designer has to make. It might be affected by many objective (at best; objective & subjective in most cases) parameters, but the "weight" of each parameter is set subjectively so, in the end, the decision is subjective.

 

[xnor]

@opc: So do you hear a difference between the wire and the o2 in a blind, level-matched comparison, if so what do you hear, what do you think it is in the measurements and what headphones did you use?

@counter culture: Excellent post.

 

[jcx]

how does name dropping Loudpeaker/Room/Sound researchers add to discussion of headphone amplifier electronic design thresholds/requirements? - very different fields - I don't recall either saying much about amplifier electronics influence on audiblity

in fact I scraped this quote from AVS Forum:

Originally Posted by Sean Olive

I have to admit that I've done very few amplifier listening tests over the years. At the National Research Council, controlled tests on amplifiers usually found that well designed amps sounded much the same until you started driving them into distortion.

 

[Ken Newton]

XNOR:

Okay, you appear too be an absolute objective-ist, then? If so, there's little to be gained from further discourse about the subjective benefits of capacitors costing more than 50 cents each, I think.

 

[opc]

Nwavguy did a good job on the O-* family of products, but he kinda looses...objective marks when he suggests he has reached some kind of performance sweet-spot where anything worse is a "compromise" and anything better is "needlessly good".

This is precisely the crux of the matter. If you go down the road of "good enough" then there's absolutely no need to stop at the O2. Pretty much anything will do the trick. Just use your headphone output on your laptop or MP3 player and stop wasting your time on here

Tangible proof of what the human ear is capable of hearing is obtainable, you just haven't taken the trouble to obtain it.

No... this is absolutely not true. I don't think you properly understand how these tests are run. I've participated in them, and been involved in the crunching of data, and I can assure you that the results are not scientific proof. They are at best a vague estimation of what the general public is capable of hearing and that's not a legitimate test as far as I'm concerned.

Let's consider an anecdote for fun:

I task you with telling me how fast a human can possibly run 100m. As a response, you gather 50 random people you find on the street, get them to run 100m, and report back that the fastest a human can run 100m is 13.8 seconds. Do you think that's true? Let's say I tell you to try harder, and instead of 50 random people, you collect 50 pretty solid athletes. Now your result is 11.1 seconds. Is that true? let's say you're even more clever, and you tell me that the fastest man on earth is Usain Bolt and he can do it in 9.58 seconds. Can you say with 100% certainty that he is the fastest man on the face of the planet? Maybe he's just the fastest man we know about to date.

The problem with people making claims about A/B testing is that it's just like the first part of my example above, except instead of 50 people you have more like 2 or maybe 6. That's not a big enough cross section to draw a scientifically significant conclusion, and it's certainly not enough data to make a wild claim like "nobody can hear the difference between an O2 and a Benchmark". It may have been true for the two or three people in the test, but that doesn't make it a universal truth, or a scientific fact.

@opc: So do you hear a difference between the wire and the o2 in a blind, level-matched comparison, if so what do you hear, what do you think it is in the measurements and what headphones did you use?

With my Denon AH-D2000 phones the answer is no... I cannot tell them apart reliably in an AB test.

With my HE-6 phones, I can reliably tell them apart at higher levels. The Wire is significantly better at higher output levels, especially in the lower registers.

Would I make the statement that nobody can tell them apart based on the above? No... that's very presumptuous.

Back to the main topic... the OP asked if there was an amplifier available with superior measured specifications to the O2, and indeed, there is. Perhaps he wants to listen to both and draw his own conclusions about what he feels is "good enough" for him. After all, it will be different for everyone.

I suggest everyone here try the same thing and decide for yourself. That's why we're on here to begin with isn't it?

Cheers,
Owen

 

[TheShaman]

...Loudpeaker/Room/Sound ...

...+ headphone + perceived sound quality vs measured performance, in general. What makes you think they wouldn't be interested?
I'm sure they would be very interested in reading some proper audio-perception-related work, if such is ever published on headphone amps, DACs and whatnot by our Objective guy.
If he's read any of their papers, he'll know what a proper blind test and proper research in general looks like; and the effort required.

Back to the main topic... the OP asked if there was an amplifier available with superior measured specifications to the O2, and indeed, there is.

And it's not like TheWire is an amp built on magic/audiophool grounds, with an enclosure milled out of solid aluminium blocks and sold for $$$$....
A 40-50 euros difference for the better amp is not prohibitive. Not to mention that, if you're a tech spec/measurement guy, you also "buy" the peace of mind and pride of having the top notch amp in this regards.
Well, OK, you'll have to take the risk of it being "too good" or "needlessly good", but I'm sure most will cope with it and sleep at night.

I guess some people need to preach, even if its the choir they're preaching to.

 

[xnor]

Okay, you appear too be an absolute objective-ist, then? If so, there's little to be gained from further discourse about the subjective benefits of capacitors costing more than 50 cents each, I think.

No, I'm just a non-BS-ist.

This is precisely the crux of the matter. If you go down the road of "good enough" then there's absolutely no need to stop at the O2. Pretty much anything will do the trick. Just use your headphone output on your laptop or MP3 player and stop wasting your time on here

Some portable players do have very respectable performance, but not very high voltage which is (I dare to say) the reason most people are using headphone amplifiers in the first place.
Other headphone outputs, like on many laptops, are a mix of line-out and headphone out. They have high output impedance, high distortion driving low impedance loads etc. I don't need to tell you that, so don't say it doesn't matter.
Compared to something like the O2 you can clearly hear differences. But compare the O2 to high-end headphone amp X ...

With my Denon AH-D2000 phones the answer is no... I cannot tell them apart reliably in an AB test.

With my HE-6 phones, I can reliably tell them apart at higher levels. The Wire is significantly better at higher output levels, especially in the lower registers.

Thanks for being honest. HE-6 is very inefficient.

Back to the main topic... the OP asked if there was an amplifier available with superior measured specifications to the O2, and indeed, there is. Perhaps he wants to listen to both and draw his own conclusions about what he feels is "good enough" for him. After all, it will be different for everyone.

I also encourage others to draw their own conclusions, not by looking at numbers though but rather by doing level-matched, ideally blind comparisons.