Upsampling by dac, can't hear an improvement.

[barrows]

On 8/21/2017 at 9:33 AM, Speed Racer said:

 

People on this site, in general, are in love with discreet upsampling. They like to say that the upsampling code in a DAC must be worse than what can be achieved in the DAC. Just as an example, how can you call the upsampling code in Yggdrasil "resource constrained" when it upsamples in real time all the time without fail. It has a filter that is "a time- and frequency-domain optimized digital filter with a true closed-form solution".

 

My listening acuity must suck too since I disagree.....

I did not see anyone saying that the oversampling filters in EVERY DAC suck.  But it is true that hardware based filters often are at a loss to what can be accomplished with a more powerful device.  In my experience though, it is possible to implement very good filters with a powerful FPGA or other type of separate DSP chip these days, as is done in a lot of better DACs.  But the filters built in to most DAC chips do seem to be a bit weak...

[barrows]

9 minutes ago, Speed Racer said:

 

But they do.....

This is sarcasm, right?  Or are you suggesting that the only dAC with a good onboard oversampling algorithm is the Yggi?  

[barrows]

5 minutes ago, mansr said:

The filters in any high-end DAC from the usual vendors (AKM, ESS, TI, etc) are perfectly adequate. While it is possible to construct mathematically superior filters, the differences in the actual analogue output are too small to be audible

Respectfully, that is not my experience, not even close.  I can easily compare the ESS 9018 chip's first stage of built in oversampling and the the custom intermediate minimum phase filter in the FPGA in my DAC and there is no contest between them.

As with the constant harping about confirmation bias, listening is a learned skill, and some people are much better listeners than others, and indeed some people will be bothered by digital filtering artifacts more than others.  Psycho acoustic testing done mostly with lay persons are not (much) applicable to audiophiles who have spent years developing critical listening skills.  

[barrows]

3 minutes ago, Speed Racer said:

There are plenty of people here that think the only way to go is discreet upsampling because DACs couldn't possible do it right themselves.

We are in agreement here.  There are certainly some DACs which can, and do, apply more complex than usual oversampling algorithms and exhibit superb sound quality.  In every case I am aware of these DACs utilize a separate chip (DSP/FPGA) with quite a bit of processing power to do this.  Chord does a fairly complex algorithm, even in the tiny MOJO.  More and more processing power in chips is making onboard filters, for the DAC designers with the chops, better and better.

For those who want to play around with different filters themselves though, I can see the attraction of OSF in the computer for all the flexibility programs like HQP and Audirvana offer.