• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Delta-sigma vs “Multibit”: what’s the big deal?

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
Me too.

They can just look so ordered, and professional, and complicated...
It just has to be good!

That may be so, but some folks keep telling us that its the least complicated devices that sound best.
I am confused now o_O
 

graz_lag

Major Contributor
Joined
Oct 13, 2018
Messages
1,296
Likes
1,583
Location
Le Mans, France
That may be so, but some folks keep telling us that its the least complicated devices that sound best...

Indeed ...
finish_2.jpg
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
But... this design uses an electrolytic coupling cap... and does have an ordinary potmeter.
How can that sound any good despite being simple ?
... And ... where are the tubes or power (MOS or whatever other) FETs ?
Don't tell me these are tucked away inside the little black box I see on the PCB !
At least it doesn't use an SMPS and is mono... it is even battery fed so no ground loops either.
You may be on to something ... let's hype it on other forums.
There is bound to be someone that prefers it over other designs.

wait ... isn't this thread about multibit vs DS ?
(Someone must have derailed it... must be me... I'll stop it right here)
 
Last edited:

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,596
Likes
12,036
I've been listening to multi-bit NOS DAC loan from my friend. A kitsune holo spring.

I have to say. I know this stuff measures worse, but I'm a sucker for NOS. I can listen for 7-8 hours without fatigue. Plus I don't hear the images/artifacting from not oversampling anyway. I really wanted to hate this thing..... :D it sounds great to me. The DX3 has more clarity but I get listening fatigue over time. The iFi black label had the same effortless fatigue-less sound but the holo spring is a much better DAC than the iFi.
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
I've been listening to multi-bit NOS DAC loan from my friend. A kitsune holo spring.

I have to say. I know this stuff measures worse, but I'm a sucker for NOS. I can listen for 7-8 hours without fatigue. Plus I don't hear the images/artifacting from not oversampling anyway. I really wanted to hate this thing..... :D it sounds great to me. The DX3 has more clarity but I get listening fatigue over time. The iFi black label had the same effortless fatigue-less sound but the holo spring is a much better DAC than the iFi.

Is it this one?
 

Frank Dernie

Master Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
6,445
Likes
15,781
Location
Oxfordshire
I've been listening to multi-bit NOS DAC loan from my friend. A kitsune holo spring.

I have to say. I know this stuff measures worse, but I'm a sucker for NOS. I can listen for 7-8 hours without fatigue. Plus I don't hear the images/artifacting from not oversampling anyway. I really wanted to hate this thing..... :D it sounds great to me. The DX3 has more clarity but I get listening fatigue over time. The iFi black label had the same effortless fatigue-less sound but the holo spring is a much better DAC than the iFi.
IME the reconstruction filter may well have a notable effect on the sound. I have a dCS P8i with a choice of filters and some of them sound different.
Which did you use? Some roll the treble off quite early which mellows the sound quite a bit, particularly if your tweeters aren’t too smooth...
I am not sure how much difference is attributable to non oversampling and how much to any filter curve chosen.
The staircase effect without a filter is indeed at inaudible high frequencies but whether that is a problem depends on whether it creates a problem for your amp/speakers not whether you can directly hear it.
 

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,596
Likes
12,036
That's the one!

IME the reconstruction filter may well have a notable effect on the sound. I have a dCS P8i with a choice of filters and some of them sound different.
Which did you use? Some roll the treble off quite early which mellows the sound quite a bit, particularly if your tweeters aren’t too smooth...
I am not sure how much difference is attributable to non oversampling and how much to any filter curve chosen.
The staircase effect without a filter is indeed at inaudible high frequencies but whether that is a problem depends on whether it creates a problem for your amp/speakers not whether you can directly hear it.
On the DX3 I use(d) filter 6 :) on the holo spring filterless 44/48Khz. Objectively I know filterless is a travesty but subjectively I actually like it.
 

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,596
Likes
12,036
Looking at the review that is the setting which starts to roll off at around 2kHz. Like a Koetsu cartridge :)

My hearing is bad though. I just like the roll-off.

Interestingly though, I have tried to mimick hard roll-off via system-wide EQ and it never really has the same effect :/ :/ then again I did not ABX so I'm not necessarily testing or comparing the right way :)
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
My hearing is bad though. I just like the roll-off.

Interestingly though, I have tried to mimick hard roll-off via system-wide EQ and it never really has the same effect :/ :/ then again I did not ABX so I'm not necessarily testing or comparing the right way :)

Was just about to ask if you’d tried that :)
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
My hearing is bad though. I just like the roll-off.

Interestingly though, I have tried to mimick hard roll-off via system-wide EQ and it never really has the same effect :/ :/ then again I did not ABX so I'm not necessarily testing or comparing the right way :)

The roll-off isn't constant. How much it rolls-of depends on the frequency and whether or not it happens to be in phase with the sampling frequency.
So the roll-off isn't constant as a filter would do.
At a given point in time there may not be any roll-off while at another point in time the roll-off can be severe.
 

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,596
Likes
12,036
The roll-off isn't constant. How much it rolls-of depends on the frequency and whether or not it happens to be in phase with the sampling frequency.
So the roll-off isn't constant as a filter would do.
At a given point in time there may not be any roll-off while at another point in time the roll-off can be severe.

That's actually quite insightful. Makes sense my one-line filter does not have the same effect!

UaoW1DE.png
 

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,596
Likes
12,036

graz_lag

Major Contributor
Joined
Oct 13, 2018
Messages
1,296
Likes
1,583
Location
Le Mans, France
It looks to be a very nice built machine, indeed, with a price, which is not offensive compared to other ones that have been tested, and failed, around here ...
 

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,692
That's actually quite insightful. Makes sense my one-line filter does not have the same effect!

UaoW1DE.png
You may try foobar's MultiResampler, choose Zero Order Hold and upsample to the maximum multiple of 44.1k sample rate that your DAC supports. It should be closer to the "real" thing than a EQ.
https://www.foobar2000.org/components/view/foo_dsp_multiresampler

The original file has a white noise at -4dBFS and a 19kHz tone at 0 and -2dBFS.

Here is an analog recording using SoX Resampler mod with these settings to simulate a DAC with a typical filter. The default settings are too good and I think it is unfair to use that setting to demonstrate things.
sox settings.png

sox.png



Here is MultiResampler. The tone at 0dBFS is clipped, but at -2dBFS it still has quite a number of spikes.
multiresampler.png

zoh.png
 

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,596
Likes
12,036
You may try foobar's MultiResampler, choose Zero Order Hold and upsample to the maximum multiple of 44.1k sample rate that your DAC supports. It should be closer to the "real" thing than a EQ.
https://www.foobar2000.org/components/view/foo_dsp_multiresampler

The original file has a white noise at -4dBFS and a 19kHz tone at 0 and -2dBFS.

Here is an analog recording using SoX Resampler mod with these settings to simulate a DAC with a typical filter. The default settings are too good and I think it is unfair to use that setting to demonstrate things.
View attachment 21295
View attachment 21293


Here is MultiResampler. The tone at 0dBFS is clipped, but at -2dBFS it still has quite a number of spikes.
View attachment 21296
View attachment 21294

Yep, thanks for the ideas. I had thought of that too, zero order hold is basically what the iFi "bitperfect" stuff does, just hold the samples.
The result is similar enough I think :) but not 100%. Here is kitsune measurement from stereophile, seems a tad less noisy maybe.

518HoloSpringfig03.jpg


The fun thing about the Kitsune is that you can exit out of NOS mode at the flick of a button :)
 

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,692
Yep, thanks for the ideas. I had thought of that too, zero order hold is basically what the iFi "bitperfect" stuff does, just hold the samples.
The result is similar enough I think :) but not 100%. Here is kitsune measurement from stereophile, seems a tad less noisy maybe.

518HoloSpringfig03.jpg


The fun thing about the Kitsune is that you can exit out of NOS mode at the flick of a button :)
My screenshots are from a Realtek:p

The noise floor and spikes depend on the converter, they are not generated by the resampler. Without going through the analog chain it looks like this:

That means despite violating the sampling theorem, filterless DACs can still be objectively measured.
Image1.png
 
Top Bottom