How do you master a DSD recording?

[gregorio]

The theory of how only one signal would fit into these data points thus producing a perfect signal I am not understanding yet.

It’s fundamentally the same as the concept of the printer printing perfect circles.

I might’ve to dig into the actual math of how this works, let’s see how far I can go.

If you’re able, that’s the way to go. I can give you a link to Shannon’s original paper if you want but Wikipedia breaks it down and makes it a bit easier.

G

 

[Chimmy9278]

It’s fundamentally the same as the concept of the printer printing perfect circles.

If you’re able, that’s the way to go. I can give you a link to Shannon’s original paper if you want but Wikipedia breaks it down and makes it a bit easier.

G

Wikipedia was also what I have in mind. But one more paper to refer from would not be unwelcome!

 

[gregorio]

A mathematical theory of communication - Claude E Shannon

G

 

[71 dB]

“We know mathematically how the signal must move between the sample points, so adding more sample points don't give anything new.”

It is this part that boggles me. I understand quite clearly how adding one more sample point to the wave does not make it more accurate.

The theory of how only one signal would fit into these data points thus producing a perfect signal I am not understanding yet. I might’ve to dig into the actual math of how this works, let’s see how far I can go.

Don't stress about. It is a lot already that you want to understand this stuff.

Maybe this helps: In digital audio, the signal is not "created" from sample pairs. The reconstruction filter uses typically a few dozen sample values to shape the signal (the reconstruction filter has a small effect on the signal at frequencies very near Nyquist frequency while the ideal filter would be infinite sinc-filter.)

 

[SoundAndMotion]

The theory of how only one signal would fit into these data points thus producing a perfect signal I am not understanding yet. I might’ve to dig into the actual math of how this works, let’s see how far I can go.

You have multiple people helping you, so I won't toss in anything just yet, but if you get bogged down in the math, I can help. If all you want is to accept the validity, so you can "accept PCM redbook with all my heart and the love it can give", I think that's great. If you want to understand why it works, you'll need some math. It's really quite cool how it works, if you want to go that way.
Hint: you will never work with 2 points. As Gregorio mentioned earlier, you need an average of more than 2 points. How many more depends on stuff (math stuff), e.g. how many cycles of that frequency are in the signal. Happy delving!

 

[Chimmy9278]

So I watched these two videos:



Only one signal can fit in a sample set without exceeding the nyquist frequency. Got that.
Any slight alteration speeding it up will exceed that and render the reproduction invalid; any slight alteration to slow down the signal won’t fit all of the points.

Even though this might not be an entirely correct summary, it is how I look at it and understand it for now. You don’t have to correct this unless it is critically incorrect.

Anyways, because of this You have come to accept the counter intuitive idea of how a signal could be perfectly reproduced when band limited. Part of it is the nyquist theorem actually has a proof for this (per the video). Not digging into that proof without supervision, so this will do for now.

I have yet to formulate my concern on that blind test paper, so will need more time for that. Otherwise without ado, nativeDSD and other such sites will only be used if I want those septic masters; CD should be pouring in soon; just gonna buy 44.1/16 FLAC files. One because I will just let the oversampler work for itself, second the 192khz might actually introduce more quantization noise when it is converted from its DSD master(at least for Sony most of the time).

 

[Inspector Gadget]

So there brings up the problem of DSD. Since it has a higher sampling density, it should be able to recreate the wave form more accurately ~ analog. And since I did find them to sound better than a file played by converting to PCM, I believe there is a difference and therefore wanted to know of the mastering process.


Thank you for your input.

This is a misnomer. It isnt how sampling works. hires music does not use more points to describe the same waveform. a 1khz wave sampled at 1MHz produces exactly the same 1kHz signal as if it were sampled at 10khz. hell, even ~2.000001kHz

 

[Inspector Gadget]

Well a quick update. Just asked a friend to help me do a pseudo AB blind test.

Did my best to match levels, and tracks start randomly. Quiet environment with iems. Source is my dap with mango player which does native DSD and power amp which converts it to PCM for play back.

Out of 12 runs I got 7, 58.33%.

But there are interruptions and gear was not optimal. Still I am pretty happy with this result.

I'm afraid this is only marginally better than 50%. It would not be considered any better than guessing in any sort of study.

 

[Inspector Gadget]

apologies, I should have read the whole thread to see others explained this. I also appreciate your attitude @Chimmy9278 and dont take me too harshly.

 

[gregorio]

Even though this might not be an entirely correct summary, it is how I look at it and understand it for now. You don’t have to correct this unless it is critically incorrect.

Well done, for two reasons: Firstly, you seem to have basically got it and secondly, you’ve linked to some factually accurate videos for a change

Although not “critically incorrect”, a slight alteration in sampling speed won’t render the reproduction entirely invalid. There will certainly be some amount of error in the reproduction but you’d need a relatively large alteration of sampling speed to invalidate the reproduction.

Otherwise without ado, nativeDSD and other such sites will only be used if I want those septic masters; CD should be pouring in soon; just gonna buy 44.1/16 FLAC files.

The first part is certainly worth baring in mind, as is also mentioned in the summary of the study to which I linked, it is relatively common to find better quality masters on SACD than you will find on CD.

G

 

[Chimmy9278]

Just saw this video:

While that allowing a filter at the recording stage more space to attenuate might reduce artifacts, what matters are the individual gear.

I think that some ADCs might already be doing this, when you pick per say 48 or 44.1, it would be sampling at twice or thrice that, and downsamples for the output. Meaning the artifacts at the filter is already avoided as much as possible.
In the DAC, what we should compare is how much does those artifacts affect the signal when oversampled versus natively hi res. Which I have not seen much if any discussion at all.

So I mean, I don’t know. Maybe hi res is worth it, forgot to consider its higher bit depth earlier too. Some orchestra pieces can reach 130dBs is what I heard.

 

[sander99]

Sorry, that video is again marketing BS.

About bit depth:
Very enlightening video, compare 16 and 8 bits and listen to the difference:


https://www.head-fi.org/threads/24bit-vs-16bit-the-myth-exploded.415361/

 

[bfreedma]

Hans is one of the worst of the charlatans. He puts just enough technical language around his subjective opinions to bait audiophiles.

This man has produced videos claiming that Ethernet Cables and Ethernet Switches make an audible difference. It takes an amazing lack of technical knowledge to make such a claim.

Best advice - avoid all Hans B. videos - nothing but snake oil wrapped in a phony layer of "science"

 

[Chimmy9278]

Hans is one of the worst of the charlatans. He puts just enough technical language around his subjective opinions to bait audiophiles.

This man has produced videos claiming that Ethernet Cables and Ethernet Switches make an audible difference. It takes an amazing lack of technical knowledge to make such a claim.

Best advice - avoid all Hans B. videos - nothing but snake oil wrapped in a phony layer of "science"

Right, thank you for telling me that. I did not like how he did not provide us with any measurements in terms or noise increase when using recording equipment under different formats, but it seemed just sciency enough to convince people. Snake oil.

 

[Chimmy9278]

Sorry, that video is again marketing BS.

About bit depth:
Very enlightening video, compare 16 and 8 bits and listen to the difference:


https://www.head-fi.org/threads/24bit-vs-16bit-the-myth-exploded.415361/

I do know about this, the higher bit depth yields a lower noise floor, thus a higher dynamic range, it is true that 16 bits is already pretty sufficient, but like I mentioned, maybe, just maybe, loud orchestra pieces might benefit with a dynamic range over 96dB since it can reach 100 or more.