Does DSD sound better than PCM?

[Kal Rubinson]

Harrumph!

As repeatedly demonstrated on this forum there are those whose lesser grasp of English does not prevent them from conveying information better than others who are native English speakers.

Granted but it can range from being trivial to becoming a deterrent. In this case, it was the latter for me.

 

[mansr]

Granted but it can range from being trivial to becoming a deterrent. In this case, it was the latter for me.

He also has some rather, shall we say, unorthodox ideas. I wouldn't put too much faith in anything he writes. Some of it is of course accurate, you just have no way of telling which bits, unless you already know the subject.

 

[dc655321]

The total amount of quantization noise remains same if you double samplerate but its spread over larger bandwidth.

If all of our quantization noise is distributed over a 20 khz bandwidth, and we increase the bandwidth via higher sampling rates to 40 khz bandwidth, then the quantization noise is the same amount spread over twice the bandwidth which means the first 20 khz has 3 db less quantization noise.

Thank you both.
I was not considering the effect on the floor of the signal (duh?!?).

 

[bennetng]

Why don't we see what Merging, creator of DXD (hi-res PCM) and big supporter of DSD, developer of both worlds, has to say about the formats?

https://www.merging.com/highlights/high-resolution

and practitioner of both world, 2L (Lindberg Lyd)
The advantages of DXD for SACD
http://www.lindberg.no/english/collection/004.pdf

How about "micro" dynamics, aka impulse response?

A slightly late reply but I still want to write about it.

The Adobe Audition 1.5 I use was released in 2004 and it supports up to 10MHz and 32-bit so I wonder what's so special about the "creator" of DXD. It is just a natural process that hardware evolved to support higher rates, just like x86 processors started to achieve 1GHz 18 years ago.

The pdf mentioned SACD so it means DSD64. I would use a real example in 2L's website, the last one in their list, 2L-053 since this music contains more real high frequency content than other demos. Also, there is a very long quiet part at the end.
http://www.2l.no/hires/index.html

In the DXD version, noise begins to show up at about 40kHz.

In the 24/96 version, that noise got filtered since 24/96 can only resolve up to 48kHz.

In the DSD64 version, after converting to the "SACD edit format" (DXD), the noise is so huge that I can't even see where the music ends. The spectrum closely resembles those DSD promotional materials. What does it mean? The illustration that DSD has a sharp impulse surrounded by silence is a complete lie, in reality any and all impulses will be buried into noise and invisible in time domain.

How about removing some ultrasonic crap? Sure it is possible, but the impulse will also be filtered. I used a cutoff point at about 40kHz. Once filtered, impulse response will be similar to 96kHz PCM, but not as good since there is still noise above 24kHz. Remember, even the lowest DSD64 has a bitrate of 5644.8kbps and 24/96 is only 4608kbps.

Now, let's see what an inexpensive modern DAC can achieve.
https://www.audiosciencereview.com/...-su-8-version-2-balanced-dac.5433/post-120602

Flat noise floor up to at least 90kHz unless people decide to use some exotic filters. Even with those filters the noise floor only increased by a small amount.

Even a cheap old soundcard in 2005 has lower ultrasonic noise than what SACD can achieve.
https://www.audiosciencereview.com/...dx3pro-dac-and-headphone-amp.4967/post-123644

 

[cjfrbw]

I am glad that this thread supports all of my subjective biases so far. Nice work, objectivists!

 

[Sal1950]

I am glad that this thread supports all of my subjective biases so far. Nice work, objectivists!

We're happy when your happy.

 

[krabapple]

Does DSD sound better than PCM?
Do you get better sound?
I am so confused.

1. No.
2. No, and neither do you. It's the mastering.

Next question?

 

[dragonspit4]

Another sincere post asking us to believe "what I can hear".

Personal anecdotes. We need more authoritative affirmation.

I have to ask why ASR keeps getting this sort of input given that its masthead clearly explains ASR's purpose? View attachment 19023

I am so confused.....
I am getting even more confused...
Why are people hearing difference between sampling rate of 44.1kHz vs 192 kHz? (converted using software: up-sampling from 44.1 to 192 khz)
Do 192 kHz actually sound better than 44.1 kHz?

Just a refresher of some analysis I performed on DSD content in video format: https://www.audiosciencereview.com/...gh-res-music-naxos-jazz-at-the-pawnshop.2004/

@amirm
Help!!!

 

[Sal1950]

I am so confused.....
I am getting even more confused...
Why are people hearing difference between sampling rate of 44.1kHz vs 192 kHz? (converted using software: up-sampling from 44.1 to 192 khz)
Do 192 kHz actually sound better than 44.1 kHz?

People hear what they believe they should be hearing when using sighted conditions. Take away the knowledge of what they're listening to and most all reports of differences disappear. Either that or the reason for the change is easily identified.

 

[jsrtheta]

Forgive my cynicism, but it is interesting how many manufacturers of decent professional kit now find that an excursion into the high-end domestic market is recommended by their accountant. At one time, dCS made the best professional converters available. But it behove them to completely abandon their core clientele in the pro world in favour of the domestic market. "Why have you stopped making them", I asked about 10 years ago (I had been a good customer until then) and was told that there was no way they could still sell the pro kit (at the prices they had charged hitherto) while getting away with the astronomical sums they were asking for the domestic stuff; it was simpler to just pull out of the pro market.

"Behove"?

 

[jsrtheta]

1. No.
2. No, and neither do you. It's the mastering.

Next question?

Stepping into this late, so my apologies.

Meanwhile...

Does DSD sound better than PCM? No, as Krabapple said.
Do you get better better sound? Not so's you'd notice.
Is it worth it? Hell no.

 

[bennetng]

I am so confused.....
I am getting even more confused...
Why are people hearing difference between sampling rate of 44.1kHz vs 192 kHz? (converted using software: up-sampling from 44.1 to 192 khz)
Do 192 kHz actually sound better than 44.1 kHz?

@amirm
Help!!!

What if he told you both PCM and DSD are crap when compared with MQA?

 

[Pluto]

"Behove"?

transitive. To have use for or need of, to require; to be in want of.

“But it behove them to completely abandon their core clientele in the pro world”

It suited their commercial purpose. In the US, “behoove” is the more common spelling.

 

[JJB70]

Even Benchmark made a balanced headphone amplifier after having had a page on their website providing the arguments why balanced headphone cables are not necessary. The market spoke and Benchmark decided it was better to give people what they wanted.

 

[andreasmaaan]

In the US, “behoove” is the more common spelling.

Is behove not the correct past form, even in the US?

 

[andymok]

Even Benchmark made a balanced headphone amplifier after having had a page on their website providing the arguments why balanced headphone cables are not necessary. The market spoke and Benchmark decided it was better to give people what they wanted.

I thought it reads.

// We do offer a 4-pin XLR connection on the HPA4 headphone amplifier, because this 4-pin connection offers lower contact resistance than a traditional 1/4" TRS connection. For performance reasons, this connector is not driven with a voltage-balanced signal. Instead, it provides isolated left and right ground returns that are individually connected to the ground reference points in the left and right amplifiers. //

 

[Sal1950]

I though it was behoof,

 

[jackenhack]

Dsd performs good on measuring but if you calculate the resolution of high frequencies in dsd is lower thant 16 bit

I think you mean lower than 24-bit because a DSD64 comes to about 18-bits of information.

 

[jackenhack]

I haven't studied DSD, but to my understanding, the use of one bit is used as an accumulator. The zero or one determines if it should increase or decrease the previous values. This is contrary to PCM, where it samples the entire value of the signal at a specific time, and stuff it into an 8, 16, 24, or 32-bit value. So you can't make a direct comparison.

 

[Bluespower]

I think'some are sensitive to the noise shaping.
To'my hear dsd sounds bad no'dynamic. Maybe some prefer it cause it's asceptic. Like'some'prefer heavily compressed records.
Pure delta sigma dacs converts pcm to'dsd and then the delta sigma'modulator makes its job. So'there is not any reason to'listen pcm converted'in'dsd whith'foobar (except'if the frequency'is much higher that the'one on ds dac)
Also i don't like the sound of pure delta sigma dacs and prefer the'one that have miltibit delta sigma technologie and also think i prefer true multibit but don't ever listen to high' quality multibit'dac.
So'i'guess it's the process of oversampling and noise shaping that can have some effect that we don't like.
You get maybe 20bit resolution if i'trust you but it's artificial by the noise shaping technic and maybe that's why some'don't like'it.
I'remeber having found good articles that explains why dsd was way below'pcm but cannot find them anymore. So'it's fine to see posts like number162.