Superdad
Member of the Trade: UpTone Audio
- Joined
- Apr 14, 2015
- Posts
- 238
- Likes
- 176
Chill guys. The reality is in the middle.
While there is a great deal of immediacy in skipping the silicon-resource-constrained (and frankly sometimes poorly designed) filters that are stock wth many DACs, the very real aliasing artifacts from playing 16/44.1kHz Redbook material through entirely non-oversampling/filterless DACs is in the long run fatiguing and a form of distortion.
Playing at least 24/176.4kHz material through the same NOS DAC is far more musically truthful.
I know this first-hand as I have an all custom PCM1704K NOS/filterless DAC. And while I listened to it for years with straight Redbook, the availability of high-quality sample rate conversion software (Audirvana+/iZotope and HQ Player) combined with an upgrade to an USB>I2S input board capable of 384kHz immediately proved how much the aliasing artifacts were a distortion I was glad to leave behind.
Computers have far more horsepower to run ultimate-quality SRC/SDM filters than what is affordable in a DAC (unless you are a DAC manufacturer willing to incorporate a giant $230 wholesale cost FPGA and do a lot of work, ala Chord's Rob Watts).
I actually think the trend for some DAC makers to offer an NOS mode is a good thing--as long as the DAC will also accept high sample rates at its input. This both assures that users can feed it high rates without overlap from the DAC's own filters (though typically if you feed 384kHz you are pretty much over-riding the built-in filters), and it frees the designer to concentrate on other areas (PS, isolation, clocking, output stage, etc.) where what they do is not so much a compromise as it is with attempting a great built-in filter.
That's my $0.02
--Alex C.
While there is a great deal of immediacy in skipping the silicon-resource-constrained (and frankly sometimes poorly designed) filters that are stock wth many DACs, the very real aliasing artifacts from playing 16/44.1kHz Redbook material through entirely non-oversampling/filterless DACs is in the long run fatiguing and a form of distortion.
Playing at least 24/176.4kHz material through the same NOS DAC is far more musically truthful.
I know this first-hand as I have an all custom PCM1704K NOS/filterless DAC. And while I listened to it for years with straight Redbook, the availability of high-quality sample rate conversion software (Audirvana+/iZotope and HQ Player) combined with an upgrade to an USB>I2S input board capable of 384kHz immediately proved how much the aliasing artifacts were a distortion I was glad to leave behind.
Computers have far more horsepower to run ultimate-quality SRC/SDM filters than what is affordable in a DAC (unless you are a DAC manufacturer willing to incorporate a giant $230 wholesale cost FPGA and do a lot of work, ala Chord's Rob Watts).
I actually think the trend for some DAC makers to offer an NOS mode is a good thing--as long as the DAC will also accept high sample rates at its input. This both assures that users can feed it high rates without overlap from the DAC's own filters (though typically if you feed 384kHz you are pretty much over-riding the built-in filters), and it frees the designer to concentrate on other areas (PS, isolation, clocking, output stage, etc.) where what they do is not so much a compromise as it is with attempting a great built-in filter.
That's my $0.02
--Alex C.
