Review: NwAvGuy's O2 DIY Amplifier - Page 45

Quote:

Originally Posted by mikeaj 

Performance aside, what would you describe as the difference between a design that uses op amps and a discrete design?  You have RLC elements and individually nonlinear electronics components constructed from doped silicon with both, just in different configurations (circuit layouts, size).  Why would this be a distinguishing factor in sound quality?  (in amount of power or supply rails that can be handled in practice, sure)

That's a good question. Open up an amp with discrete components and you see all the parts and layout - pretty simple. I'm not sure that's the case with op-amps. I've noted some manufacturers hide the design / schematics for their top-tier op-amps (e.g. LM4562) . Even when the designs are indicated in the data sheet, it's my understanding it's a very simplified representation of what's actually implemented on the die. I would have to defer this question to those who design semi-conductors or are experts in op-amp design.
 

Quote:

Originally Posted by purrin 

First step is to understand what the numbers mean. As you mentioned: "fraction of audible distortion". This would indicate that we are measuring the wrong thing. Why are we measuring things that can't be heard? As I said earlier, who cares if a manufacturer says their headphone goes up to 54kHz.

 

So I repeat, where is the scientific data that shows the O2 sounds as transparent as the discrete designs?

 

 

 

 

Why are we measuring things that can't be heard? To ensure they can't be heard in order to ensure absolute transparency. :|  

The O2's THD is ~.002%, IMD is simlar. What else is there (in terms of detail or transparency?) 

Really, what else am I missing, someone tell me? I am not an expert, but I just cannot believe in something that has no evidence behind it.

The "it" being that a more expensive amp can be more transparent than an amp designed to be completely transparent with numbers to back it up. BTW, the whole point of his blog is to disprove this "it" and as far as I know he does not recognize anything of the sort. 

"Put it this way, if there were no audible upgrades available to the current O2 design, why would his challenge be to amps costing only up to $500 and not any amp of any value? Clearly NwAvGuy recognises there is an audible difference (even if it is slight) the higher up the ladder you go." - Naim.F.C 

Quote:

Originally Posted by purrin 

That's a good question. Open up an amp with discrete components and you see all the parts and layout - pretty simple. I'm not sure that's the case with op-amps. I've noted some manufacturers hide the design / schematics for their top-tier op-amps (e.g. LM4562) . Even when the designs are indicated in the data sheet, it's my understanding it's a very simplified representation of what's actually implemented on the die. I would have to defer this question to those who design semi-conductors or are experts in op-amp design.

Yes, the schematic given for a more complicated op amp circuit is just a user's guide, a simplification of the "equivalent" key parts of the topology.  It tells you general system characteristics and maybe how it will interface with components you use with the op amp, not so much about the entire internal circuit composition, which would be much more difficult to read and understand.  For some of these systems, if you printed out the entire schematic, even the ones who designed them would struggle to understand what's being represented.

I love basketball. It's my favorite sport in the whole world, and I think it's the perfect blend of excitement and intensity. it has the simplest statistics to dissect, and you can get a good feel of who the best players are in the world just by glancing at statistics.

Well, basically there's this guy named Bill Russell. By every measurable metric, he shouldn't rank as one of the greatest players in NBA history. He was a subpar scorer, a mediocre shooter. He was an excellent rebounder and defender, but nothing about him stood out. This man won 5 MVPs and 11 championships and by the time he retired in 1970, nearly everybody agreed that this is the greatest player ever (until MJ arrived).

40 years later, I see a large gap between the opinions of those who saw him and those who never did. Those who didn't point out the statistical inadequacies he had compared to contemporaries like Wilt Chamberlain or Oscar Roberston. Those people who saw him, of the older generation, try to point out that what made Russell great cannot be measured in statistical value. They pointed out that such things like leadership, team defense, nitty gritty plays cannot be measured statistically, that one had to see him play to realize why the man won 11 championships - by far the greatest winner in any sport. How could a man who wasn't outstanding statistically, win so much with a team that wasn't that much better than everyone else's? That's why they call it intangibles, because they go beyond what you can see in a boxscore. That's why it's so hard to quantify Bill Russell in a modern world.

The debate rages on today.

Thanks for correcting me above. Seems I may have miss-read. I assumed the challenge was only for amps up to $500 in value. Didn't realise it wasn't, nor that it was subject to any op amps.

I can't wait to conduct my own listening tests between my amps. I actually own the O2, which currently I use exclusively with the LCD-2, since it doesn't seem to give as much warmth as the ALO Continental with my T1's. I find the T1's are a little dryer with the O2's. The high's have a sort of grain to them.

If the O2 is neutral, this leads me to believe the Continental is intentionally coloured or smoother. Likewise, I find the LCD-2's aren't as quite as clear with the Continental as the O2. I wonder if this could have anything to do with sensitivity or ohm's of the cans in relation to the amps?

EDIT: Just to clarify, the above testing was iPhone > LOD > amp, so it's possible once I add one of my DAC's to the equation it could change things.

Quote:

Originally Posted by Shike 

Really?  It sounds like your claim is similar to  "go chase ghosts for yourself because I know they exist/saw one".  We aren't going to do the chasing for you, I've heard more expensive amps besides the O2 and couldn't tell them apart (after the review I've had hands on with M3 and gilmore amps + some others).  This leaves the excuse that the problem is my ears.  Thus, the proof of burden is on the one claiming a difference, because otherwise we have a conflict of interest or the "tin ears" scape' goat.

So I assumed you heard the amps in question and heard no difference? That's OK and fine by me. I do think it's a little self-righteous to tell many others who've designed, built, and listened to the discrete designs mentioned that it's all in their heads.

Quote:

Originally Posted by Satellite_6 

Why are we measuring things that can't be heard? To ensure they can't be heard in order to ensure absolute transparency. :|  

 

The O2's THD is ~.002%, IMD is simlar. What else is there (in terms of detail or transparency?) 

 

Really, what else am I missing, someone tell me? I am not an expert, but I just cannot believe in something that has no evidence behind it.

What you are missing is that THD or IMD does not measure detail and transparency. Do you know what 2% THD sounds like for a 100Hz signal at 90db? I don't - because that's not enough information. In this case for example, I would then ask: What comprises the THD? Is the second order distortion higher than third order (or vice versa), because these types of distortion sound totally different to the ear.

And then it gets more complex. What about non-linear distortion (THD + IMD) for more complex signals (say 3 tones, or 5 tones). Then things get really hairy and complex. Music of course is even more than 3 or 5 tones, and certainly not steady state.

That being said, I do hear the lack of distortion of the O2. It's very very clean sounding up and down the volume range.

Quote:

Originally Posted by idletime1213 

I love basketball. It's my favorite sport in the whole world, and I think it's the perfect blend of excitement and intensity. it has the simplest statistics to dissect, and you can get a good feel of who the best players are in the world just by glancing at statistics.

 

Well, basically there's this guy named Bill Russell. By every measurable metric, he shouldn't rank as one of the greatest players in NBA history. He was a subpar scorer, a mediocre shooter. He was an excellent rebounder and defender, but nothing about him stood out. This man won 5 MVPs and 11 championships and by the time he retired in 1970, nearly everybody agreed that this is the greatest player ever (until MJ arrived).

 

40 years later, I see a large gap between the opinions of those who saw him and those who never did. Those who didn't point out the statistical inadequacies he had compared to contemporaries like Wilt Chamberlain or Oscar Roberston. Those people who saw him, of the older generation, try to point out that what made Russell great cannot be measured in statistical value. They pointed out that such things like leadership, team defense, nitty gritty plays cannot be measured statistically, that one had to see him play to realize why the man won 11 championships - by far the greatest winner in any sport. How could a man who wasn't outstanding statistically, win so much with a team that wasn't that much better than everyone else's? That's why they call it intangibles, because they go beyond what you can see in a boxscore. That's why it's so hard to quantify Bill Russell in a modern world.

 

The debate rages on today.

That's pro basketball, which is worlds apart from the input/output relationship of an audio amplifier.  "Who's the best basketball player?" is a much different question than "how different is one amp compared to another?"  If the output is supposed to be 1.93571 at time 4.351 and it's 1.93589 instead, we can score that.  Saying something is the "best" is always subject to a lot of interpretation over what "best" means, context, and so on.  

What some people are saying is that the traditional stats aren't everything that matters.  As one example, huge points and rebounds guys with lots of production can actually be quite mediocre players because of low efficiency, stealing actions that others would have gotten, and so on.  That's the primary reason why I'd say that traditional basketball stats are much less indicative than traditional audio metrics.  Probably +/- and other more advanced metrics probably correlate better with how good a player actually is.  For many legendary players, the teammates are probably better than they are given credit for, at least relative to the other players in the league at a given time.  Also, some people act as if championships and success are perfect indicators of ability.  It's random.  11 championships could have been 8 or 12 or something else, with different luck.  This doesn't really have much to do with the audibility of certain effects and measuring amplifiers.

I'd almost never trust listening impressions from people who build things or design them, since they've got to be attached.  But there's definitely reasons to go discrete, such as (1) more power output possible even though most people probably won't need it, (2) better marketing.  In theory you can also get measurably better performance which some of us would clearly argue is way past the point of audibility for humans.  If you want to best the O2's performance using stock op amps and buffers, you end up with a design like The Wire.  If you want something even better for a particular application, maybe it's possible with your own topology, rather than more general-use parts.

Even more importantly, from a headphone amp designer's point of view, using op amps is relatively boring and easy.  In general, particularly if there's not a confounding challenge like portability, size, or price, doing it the easy way is no fun, with no feeling of accomplishment at the end.  People want to innovate.

Can anyone knowledgable on output impedance etc educate me on what sort of specifications would best suit the T1's? And whether by default design, the O2 is more likely to match with the LCD-2 over the T1 based on such factors. I'm genuinely curious.

Quote:

Originally Posted by Shike 

The Continental if I'm not mistaken is an OTL design, which means it's likely to have a larger output impedance.  This can definitely modify frequency response of headphones by creating a voltage divider depending on the headphones impedance in regards to frequency.  Read NwAvGuy's write-up on why output impedance is important, or if you don't trust him I'm sure there's articles on Google that can explain it by now ^_^

I'm pretty sure its a hybrid with a transistor output stage.

Quote:

Originally Posted by Shike 

 

This isn't a valid comparison

Yes, it only makes sense to talk about ghosts or Santa Claus in relation to audio electronics.  You should stop appealing to logic in the deductive sense you are misapplying it because the fallacies are racking up quickly.

Quote:

Originally Posted by maverickronin 

I'm pretty sure its a hybrid with a transistor output stage.

Yup.  Still, the Continental is largely irrelevant to the real discussion that purrin and mikeaj have been having.  But it's nice to have more data and input to evaluate.  That's how I feel, I'm sure others disagree.

I haven't read every post on this thread so I apologize if it was answered but I couldn't find it with the search. Can anyone compare the  Zo2 vs the O2 vs. maybe even a maverick D1. I have been thinking about the O2 but I don't have a dedicated dac yet and I'm not exactly sure if I need/want a portable rig at this moment so that's why I added the D1 to the list. I can make a new thread if it is too off topic.