Excellent document not only from a technical aspecy, but also very well written.
It’s a breath of fresh air at a moment that the thread is getting filled with all kind of postings that do contribute little to answerring the OP’s question.
Hans
Being one of the readers and not a participant is because I think I've already compare the two kinds of DACs several years ago. One was a DAC with a PCM1798, as far as I know an OS DAC, the other was a so called PIANO DAC, as far as I know a NOS DAC. For the comparision I made a small box with a potentiometer (volumecontrol) to even the levels as good as possible and a switch to be able to quick compare. We used a homemade (my own design) class A amplifier, unfortunately I forgot the loudspeakers we used (it was with a audiofriend), but this were not simple ones. I was able to hear a very small difference between the two DACs, but that was only a tiny shift between the placing of the instruments. So tiny that I blame the volumecontrol to be not exact synchrone. In quality I was not able to hear a difference. After about 20 minutes I decided that for me their sound was equal. I do not want to disturbe this thread and I'm still interested in all these findings.
Unfortunately there's a lot of FUD out there regarding digital volume control. I never understood that, never heard any detrimental effects of it.
In my case I simply never run the output signal to the DAC hotter than around -6dB. Problem with intersample overs solved.
No fancy plugin required and actually not available for cheap but even for free
Absolutely, provided you have a digital volume control and there is no sample rate conversion done before the signal reaches the volume control. When I play a CD or another old-fashioned physical digital carrier, I have no such control. When you use an SRC4392 to reduce the volume, it first converts the sample rate and then reduces volume.
In my case I use pulseaudio with sample format configured to 32 bit float, I don't care if the SRC causes values >1.0, that's a big advantage of running your audio middleware with float values because they don't clip. The volume is then reduced by 6dB which will make samples <1.0 again, hence no clipping if decimated down to 24 or 16 bit integer before being sent out to whatever DAC you use.
I can see why this might not apply to lots of people, lots of them might use windows with very little control to the audio backend, have fear of computers etc...
But if you only play digital media I can encourage everyone to use a small media center PC for that. Setting up a desktop linux distribution is not hard nowadays and completly free. Playing CDs is also no problem if you add a CD/DVD drive.
I can see why this might not apply to lots of people, lots of them might use windows with very little control to the audio backend, have fear of computers etc...
But if you only play digital media I can encourage everyone to use a small media center PC for that. Setting up a desktop linux distribution is not hard nowadays and completly free. Playing CDs is also no problem if you add a CD/DVD drive.
Let me take a hard standpoint.
In my terminology you are just a troll pretending to have all wisdom but just doing nothing else but stir up unrest.
Sorry for that.
Hans
Aside from the fact that this is NOT the original question, what artifacts are introduced aside from the obvious ZOH based treble droop, which is easily EQ'd flat? Since you use the plural, artifacts. Do enlighten.
Last edited:
You see confusion - except your own. The unsupported suppositions, and flatly wrong assertions you make are quite entertaining, so don't let me stop you.
Marcel,
I had occasionally noticed editing being done on older posts as well, and also wondered how. Where did you find instructions?
Thanks.
Apparently a lot of people are overthinking this stuff.
Apparently, a lot of other people haven't thought about it enough.
Nowhere, I just turned out to have an edit button for the very first post (not for the other old posts).
What do I need to perform a NOS/OS test?
Obviously, I would need a DAC that accepts 44.1 kHz to 176.4 kHz in 16 to 24 bit format, and its reconstruction filter (if there is any) can be disabled.
But what about the source? If the sample tracks are played on a PC through USB output, it is operation system dependent what is happening behind the curtains (WASAPI, Direct Sound, ASIO, upsampling, truncating, there is little visibility of all these parameters). For example I can play all sample tracks using Audacity to my humble external sound card having only 16 bits / 48 kHz resolution. I suppose Windows does the downsampling/truncating. But the Windows sound control panel still permits setting the audio playback to 24 bit / 96 kHz, and Audacity happily plays a 32 bits / 176.4 kHz file... So I am confused.
Perhaps a recommended methodology would be useful for most of us.
Obviously, I would need a DAC that accepts 44.1 kHz to 176.4 kHz in 16 to 24 bit format, and its reconstruction filter (if there is any) can be disabled.
But what about the source? If the sample tracks are played on a PC through USB output, it is operation system dependent what is happening behind the curtains (WASAPI, Direct Sound, ASIO, upsampling, truncating, there is little visibility of all these parameters). For example I can play all sample tracks using Audacity to my humble external sound card having only 16 bits / 48 kHz resolution. I suppose Windows does the downsampling/truncating. But the Windows sound control panel still permits setting the audio playback to 24 bit / 96 kHz, and Audacity happily plays a 32 bits / 176.4 kHz file... So I am confused.
Perhaps a recommended methodology would be useful for most of us.
Ken,
It’s very convenient and because you can only add text, you can neatly pack attachments into dropbox files or use normal links.
See here as an example where I added a link at a much later stage.
Designing a universal diff-in/diff-out Head Amp
Hans
It’s very convenient and because you can only add text, you can neatly pack attachments into dropbox files or use normal links.
See here as an example where I added a link at a much later stage.
Designing a universal diff-in/diff-out Head Amp
Hans
Of course not.
First of all, scientific literature can at best provide evidence, not proof. 'Proof' exists only in mathematics.
Second, no reason to believe that everything that can be researched must already have been researched. So, why should we care if some literature can't be found? It means nothing one way or the other.
Last edited:
Hi Laszlo,
When using software like JRiver, Foobar, Roon or Audirvana to name a few, you have much more control on what’s going on.
Hans
To the contrary, it is known that some mastering engineers find their clients prefer the sound of clipping to the sound of look-ahead (or earlier types of) peak limiters.
Last edited: