[mk351e]

I've spent 20 years tinkering with home stereo, HT, and car audio. I spent a LOT of time and TONS of money designing and building home speakers. With all that stuff, when you're designing, there's enough published specs to make accurate educated guestimates and computer models about how things will sound. Best part is, you get pretty close. I've discovered with the world of high end headphones, NONE of this really seems to apply!! It's wild! It feels like everything you do here is little more than a shot in the dark! I went to a HP meet 3 years ago, and was SHOCKED at some of the things I heard, and they were so opposite to what I expected! I expected to like the T1's, and didn't, I expected to hate the HD800, and yet I've been thinking about it for 3 years now!
 
So, I'm in the market for a new DAC/Amp, and ideally, I'm looking for an all in one box. That aside, like the title says: can we draw ANY conclusions about how something will sound, or, how good it will sound (which is entirely subjective itself) by looking at the internals? For example, when I look inside say, an Audio GD Compass 2, it looks pretty bad ass. There's all kinds of business going on in there. Problem is, I really know nothing about circuit design, except big caps and transformers are good, and so is Class A. Another unit on my mind is the Fostex HP-A8c, and I did manage to find an internal shot on the web, and it looks considerably less impressive.
 
I've attached pics of both below. Can we draw any conclusions? Anything?? If so, what should I look for?? Also, I know about the obvious price difference, so just ignore that for the sake of this discussion
 
 
 

 

 

[johnwmclean]

Paging Spritzer...

 

[Operakid]

One might argue that he could draw some very tentative ideas about the possible upside based on quality of components but in reality a design full of great components can sound horrible.  For instance, in only one of many possible scenarios, we could have a potential great design that oscillates, creating horrible distortion.  Changes in values of certain components of that design could eliminate the oscillation making it sound great.  
 
I have heard messy, proto looking designs sound fantastic and beautiful looking design with fantastic components sound nearly unlistenable.
 
   

 

[Currawong]

Quote:

One might argue that he could draw some very tentative ideas about the possible upside based on quality of components but in reality a design full of great components can sound horrible.  For instance, in only one of many possible scenarios, we could have a potential great design that oscillates, creating horrible distortion.  Changes in values of certain components of that design could eliminate the oscillation making it sound great.  
 
I have heard messy, proto looking designs sound fantastic and beautiful looking design with fantastic components sound nearly unlistenable.
 
   

 
^This I reckon. It's probably relatively easy to make judgements on the design of a point-to-point wired tube amp, but much harder with digital gear on the other.  I own basically the same DAC as I did in 2009, but as the quality of the digital input and processing has been improved, even though what I owned in 2009 doesn't look radically different to what I have now, the improvement in resolution has been vast. There is also the Meitner DAC and EMM Labs DAC2X which are basically the same DAC but in the latter, better components are used for exactly the same design with, reputedly, significant improvements in the sound as a result.

 

[Maxvla]

No.

 

[obobskivich]

You might be able to make a few guesses here and there (as currawong and operakid have said), but your best bet is going to be to listen to (or measure, or otherwise "observe") the final product. Especially in the era of VLSI - a lot of the overall "sound" or "flavor" is going to be as much influenced by software, as it is by hardware (this has become quite common with AV hardware, and I've started to see it popping up here and there with hi-fi gear too (e.g. the HK990)).

 

[wuwhere]

No EE or audio EE would even speculate. Software or firmware? Not even close.

 

[obobskivich]

Software or firmware? Not even close.

What do you mean by this?

 

[wuwhere]

Quote:

What do you mean by this?

 
Most real time software is written in C language. If you see the code, how would you know what its doing?

 

[obobskivich]

Most real time software is written in C language. If you see the code, how would you know what its doing?

Oh, okay now I understand what you meant. (I think it's officially past my bedtime too...)

I think we're on the same page - that it's easier to just observe the outputs, than to try and reverse engineer the device out of a black box.

 

[wuwhere]

Quote:

Oh, okay now I understand what you meant.

(I think it's officially past my bedtime too...)

I think we're on the same page - that it's easier to just observe the outputs, than to try and reverse engineer the device out of a black box.

 
Reverse engineering is not easy but doable. I remember about using a Motorola assembly code and reversing it back into C source code. That was back then.

 

[yage]

Most likely not. The proof is in the end result. That's why you can't only look at measurements to determine sound quality, you need to audition it for yourself.
 
For example:
 

 
or this:
 

 
Tells me nothing about how it will perform in my system, to my ears. They both look cool, though.

 

[spritzer]

With simple overview pics such as those you can draw basic conclusions about the performance of something.  Look at the parts used, how they are used etc.  Basic opamp buffer is never going to outperform a good discrete buffer and stuff such as that.  I've been doing this over the last couple of days when looking for a new preamp for a friend.  You'd be amazed what has been written off simply based on a simple over view pic revealing sub par engineering, nasty compromises and sheer stupidity.  10$ volume control in a 8K$ preamp? 
 
Go in closer and you know a lot more when you can see the part numbers and how they are used.  It's easy to draw up a complete schematic from a few pics and some experience. 

 

[mk351e]

Ah, great, so it looks like I'm screwed then?! Good thing there's a meet every year! Ugh, this hobby is brutal. And no matter how many times I say it, truth is, it NEVER ENDS!! =]

 

[wink]

Quote:

Ah, great, so it looks like I'm screwed then?! Good thing there's a meet every year! Ugh, this hobby is brutal. And no matter how many times I say it, truth is, it NEVER ENDS!! =]

And, your wallet is always the loser.....