• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Distortion Listening Test

stevenswall

Active Member
Joined
Dec 10, 2019
Messages
132
Likes
58
#62
This song sounds like a poorly recorded analogue tape master digitized years later, but I did okay. They have better options for 6" speaker when that is selected.

1579426584451.png

1579426612464.png


Edit: None of these are particularly clear. Owl City "Fireflies" seems like it would work better. Also, the bass is distorted during recording and doesn't sound anything like a modern recording on most of these. Bit of noise floor it seems too. Cool test though.
 
Last edited:
Joined
Feb 2, 2020
Messages
6
Likes
1
#64
So if we get a result like -30, does this mean that we won't really benefit from DACs that go much lower in dynamic range?
 

solderdude

Major Contributor
Joined
Jul 21, 2018
Messages
4,122
Likes
7,559
Location
The Neverlands
#65
It means you need transducers with better performance, room treatment and/or training in listening.
That will probably get you in the -51dB segment, as you can see there is a statistical oddity there.

This is not caused by DACs nor can it be blamed on amplifiers (unless they aren't at least decently designed) or by cheating.
 
Last edited:

JEntwistle

Member
Forum Donor
Joined
Oct 25, 2019
Messages
46
Likes
46
#67
After watching a portion of the video: Does this mean that you could have a DAC that performs relatively poorly in Amir's SINAD measurements, but by "sheer luck" happens to have its noise in frequencies that don't stand out to our hearing?

I'm not sure if any of the tested DACs actually meet those criteria. And if they do, I'd bet not by intentional design...

If this is correct, then it seems to me the idea is to purchase a DAC with minimal SINAD across the frequency spectrum so as to have all bases covered, as opposed to hoping some DAC designer is able to put the noise only in frequencies where it doesn't matter. And with so many excellently performing lower priced DACs available, why would you take a chance on a noisier DAC these days?

On a side note, I really like his characterization of "accuracy" vs. "preference"; let the user decide their preference, but don't argue with the math of accuracy.
 

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
517
Likes
214
Location
London, UK
#68
After watching a portion of the video: Does this mean that you could have a DAC that performs relatively poorly in Amir's SINAD measurements, but by "sheer luck" happens to have its noise in frequencies that don't stand out to our hearing?
"Noise" may well be at less sensitive frequencies and not by sheer luck... noise shaping. (Dither can also be spectrally non-flat.)

as opposed to hoping some DAC designer is able to put the noise only in frequencies where it doesn't matter.
That's fundamentally how delta-sigma DAC's work.

1581793919204.png


Source: Texas Instruments.

To quote from the linked page:

"Any DAC will produce noise that extends well above audio frequencies, and this effect is strongest in delta-sigma DACs. These converters use noise-shaping techniques to improve in-band signal-to-noise ratio (SNR) performance at the expense of generating increased out-of-band noise."

The page then discusses the need to filter "out-of-band" noise for automotive application regulatory compliance.



In the "multi-bit" world, even the first CD player from Philips used noise-shaping. So, you'd be stuck with off-the-wall "No Oversampling" (NOS) "multi-bit" designs. (=POOR measured performance.)
 
Last edited:

JEntwistle

Member
Forum Donor
Joined
Oct 25, 2019
Messages
46
Likes
46
#69
"Noise" may well be at less sensitive frequencies and not by sheer luck... noise shaping. (Dither can also be spectrally non-flat.)



That's fundamentally how delta-sigma DAC's work.

View attachment 50307

Source: Texas Instruments.

To quote from the linked page:

"Any DAC will produce noise that extends well above audio frequencies, and this effect is strongest in delta-sigma DACs. These converters use noise-shaping techniques to improve in-band signal-to-noise ratio (SNR) performance at the expense of generating increased out-of-band noise."

The page then discusses the need to filter "out-of-band" noise for automotive application regulatory compliance.



In the "multi-bit" world, even the first CD player from Philips used noise-shaping. So, you'd be stuck with off-the-wall "No Oversampling" (NOS) "multi-bit" designs. (=POOR measured performance.)
Yes. But I thought in the video you posted he was saying that at certain frequencies - even in the lower ranges - high noise levels would not be noticeable, even before noise shaping. So if you have a poor implementation that does not take advantage of even the most basic noise shaping and other signal processing techniques, conceivably not all of the noise would be audible to most people. And that leads to arguments like "well, it doesn't matter" or "it sounds better to me", which are all just excuses for poor engineering efforts.

Anyway, thanks for posting this video and for the additional information. I am still learning about this and reading all I can, and I greatly appreciate it.
 

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
517
Likes
214
Location
London, UK
#70
Anyway, thanks for posting this video and for the additional information. I am still learning about this and reading all I can, and I greatly appreciate it.
You're welcome. Your questions deserve a far more elaborated response than the one I gave, but I'm too tired at the moment to do so.

Perhaps @j_j might notice this thread...
 

JEntwistle

Member
Forum Donor
Joined
Oct 25, 2019
Messages
46
Likes
46
#71
Wait a minute - is the j_j here on ASR the same as James D. Johnston in the video? I didn't put that together....

One of his threads sent me down a rabbit hole to learn more about dithering. I read the excellent master's thesis by Cameron Nicklaus Christou that someone posted here. While I can't follow all of the math, he really does a great job of explaining what is happening.
 

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
517
Likes
214
Location
London, UK
#72
Wait a minute - is the j_j here on ASR the same as James D. Johnston in the video? I didn't put that together....
Indeed he is. :)

One of his threads sent me down a rabbit hole to learn more about dithering. I read the excellent master's thesis by Cameron Nicklaus Christou that someone posted here. While I can't follow all of the math, he really does a great job of explaining what is happening.
Have you a link, please?
 

JEntwistle

Member
Forum Donor
Joined
Oct 25, 2019
Messages
46
Likes
46
#73
Indeed he is. :)



Have you a link, please?
Here is the link to the paper. It is really about dithering as used in image processing, but the first section has an overview of dither as used in audio processing:
Optimal Dither and Noise Shaping in Image Processing by Cameron Nicklaus Christou
https://uwspace.uwaterloo.ca/bitstr...d=25572771711789E8989DB2878126304E?sequence=1

I don't know if this will be advanced enough for you, but other than some of the higher level mathematics, it was a good intro for me.

I can't find the thread where I was referred to this paper, and now I wonder if I just googled it on my own. But I became interested in the topic from this thread:
https://www.audiosciencereview.com/...l-process-not-a-psychoacoustic-process.11169/
 

j_j

Addicted to Fun and Learning
Audio Luminary
Technical Expert
Joined
Oct 10, 2017
Messages
606
Likes
860
Location
My dining room.
#74
Wait a minute - is the j_j here on ASR the same as James D. Johnston in the video? I didn't put that together....

One of his threads sent me down a rabbit hole to learn more about dithering. I read the excellent master's thesis by Cameron Nicklaus Christou that someone posted here. While I can't follow all of the math, he really does a great job of explaining what is happening.
https://ieeexplore.ieee.org/document/823976 (take a deep breath and have a cup of coffee (tea, whatever) first!)

http://users.eecs.northwestern.edu/~pappas/papers/pappas_ist94.pdf

I'll be quiet now. :D

The irony of dither is that it's been a known thing for many, many years, and the need for linearization is very, very old, and can be implied from Shannon's original paper.
 

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
517
Likes
214
Location
London, UK
#75

j_j

Addicted to Fun and Learning
Audio Luminary
Technical Expert
Joined
Oct 10, 2017
Messages
606
Likes
860
Location
My dining room.
#76
Top Bottom