[Danny Richie]

You mentioned things can add up.  I agree and also stipulate they add up quicker when done in a haphazard way.

Math and processing has come a long way.  Math does not suffer from parts tolerance variances and age degradation.

Math and processing potential improves with age as history has proven.

I experimented with math and have a few examples below.

The best measuring DAC today is @ 130dB so to drive that DAC to its potential you need something cleaner than its maximum potential.

DSP can be done today in 16, 32, 64 and 128-bit, yet many tools chose lower precision math.

Here are some examples that I have tested:

Here are 4 FIR XO's processed in 32-bit ( 16-bit has a higher noise floor than -130dB so is already under-performing the DAC )

Note the noise floor of -140dB.  Just marginally better than the best contemporary DAC.

Here are the same 4 FIR XO's processed in 64-bit.

Note we are now down to @ -240dB, surpassing the best contemporary DAC's capability.

Now look at the same 4 FIR XO's processed in 128-bit.

Note we have now exceeded the precision of the display tool.

It is possible with today's tools to do all the math and convolution in 128 and 64-bit until the DAC's doorstep.

I would suggest the difference between good and poor DSP performance today is the choice and consistency of the filter and processing chain (using your favorite DAC du jour).

I have found some software tools to cut corners, thus avoid them accordingly in favor of ones that do not.

It sure would be nice if it were just about the numbers.

[NoahH]

@emailtim - indeed it can be done. I fully believe that some folks who are smart enough and have spent enough time (decades) on it like Theoretica get there. And I would always be delighted to audition new implementations. But the reality is the DSP based gear I have heard has underperformed simplified electronics in controlled spaces. There is a reason that many of the most-belovee DACs are NOS. Ladder DACs dominate many of the rankings.

But *theory does not matter on this*. Trial different approaches and try for yourself. There is no replacement for experience, and trying different approaches is much of the fun of the hobby.

My best suggestion for anyone new is buy used gear that you can sell at a low loss and try different stuff for a long time.

[jmimac351]

My best suggestion for anyone new is buy used gear that you can sell at a low loss and try different stuff for a long time.

BINGO!  I did that with DACs (and preamps) and found out what I wanted to know. That issue, for me, is settled.

Now I play with speakers.



Below is how the preamp / DAC mental gymnastics got settled... but the little "secret" is that little black box, by itself, is almost as good.  I did it because I work hard and could, not because I needed to to be happy. 

Buy gear used.  The only "new" I buy are special speakers / speaker parts... if they can't be bought used.  If you buy good stuff, and buy it right... you haven't actually lost the money - it's just value in a different form to have fun with until it needs to change form... like turning it into track tires.

I will say it's important not to skimp on what kind of cardboard you use for housing a test crossover that you're anxious to hear. 

[NoahH]

@jmimac351 - I am also an Ayre guy now.