• 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!

A new Dac to improve sound quality from Khadas Tone Board 1

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
Maybe the 0dBFS being reached a lot has something to do with it.
My FiiO X3 didn't like 0dB either. The X5-II has no issues with it.

Try this: Switch on the built-in equalizer and set it to 'flat' this way the DAC and or subsequent post filter/amp will not reach 0dBFS.
 

b4nt

Addicted to Fun and Learning
Joined
May 29, 2021
Messages
803
Likes
270
Which equilizer? APO? The DAC is connected via ASIO. And the D50s features only modes from 1 to 7.

I may try with my head sets and the PC sound card, plus APO equilizer, to avoid any additionnal clipping.
Or with the DAC, but lowering the level in file using Audacity.

But I shouldn't have to do that. I do more suspect a real track issue, in the audible spectrum, or the spectrum moving speakers. Then it might pop or not in settings depending of their sensitivity or precision.

The head set I'm using is a common and basic studio mixing/recording pro model, sensitive wide band (5-30kHz), and would reproduce any audible clipping/distortion/noise at recording time or whilst mixing. Maybe that track wasn't checked so. On 100-16k PC speakers (wide spread), it seems clean, and all people in the comments section on YT enjoy that track (a title which I enjoy also, so I focused on).

Edit: Lowered the signal by 2dB in Audacity (so signal away from clipping region). Same issue. It is related to spectrum in the track, on some positions;
 
Last edited:

b4nt

Addicted to Fun and Learning
Joined
May 29, 2021
Messages
803
Likes
270
But we may discuss again about DAC differences.
And you may consider a/ I'm trolling or b/ I open again those discussions :)

Maybe the 0dBFS being reached a lot has something to do with it.
My FiiO X3 didn't like 0dB either. The X5-II has no issues with it.

Did'nt we agree all sound the same? :)

X3 and X5II are now old generations. From what I noticed, main differences where in output amps qualities, and so in pricing/range.

I had now hit several issues with my X5 and specific tracks, close to those I hit now with the D50s and Hush. But for Hush, I beleive it is inside the track. I'll try to make some comparative lines out records, to check. I'm now interested in that also. Maybe, I'll find a fix like for Hush (where I would basically need to EQ some positions for a better sound).

The X5 has an EQ feature which I never used (found that sound not so good). I may try to record also with EQ active, because I noticed EQ lowers the output levels (so away from clipping area). Or to check how output signal changes with EQ, for a problematic track.

Try this: Switch on the built-in equalizer and set it to 'flat' this way the DAC and or subsequent post filter/amp will not reach 0dBFS.

I use ASIO, for direct streaming towards the DAC, without any rate/bits conversion by Windows (or Linux on my other setup). I do not EQ either.

Doing so, the DAC shall also be working with 24 bits when in stream. The D50s is capable to reproduce up to 19 to 20 bits? I do not know what my Aune S16 is capable of. Maybe the D90SE is a nice candidate for me to replace that Aune, according to reviews, here...

At a previous page, one mentionned hires may contain a lot of noise in hifreqs, susceptible to alter the audio in audible range (causing distortion). Later I remembered I have a DSD file. Then I found that post, here, about such hifreq noise and that sample DSD file, plus conversions techniques:
https://www.audiosciencereview.com/forum/index.php?threads/digital-filter-game.23795/

That only DSD (sample) file I have is Mahler 1st IV finale, you shall listen to it. Samples are available in different bitrates to compare how they sound. With a particularity: high volumes differences inside the tracks. A 44.1/16 DAC would be limited with, where a 24bits DAC would be able to produce a better sound for low levels of such a track.

If you quickly want the already converted FLACs or the DSD files to listen and compare, you may directly go here. Care about your volume settings, don't raise it too much for the begining of the track:
https://dsdmaster.blogspot.com/p/blog-page.html

A 24bit or a DSD capable DAC may so sound very different I think, compared to 44.1/16 DACs. Depending of files/formats one is listening to.
 
Last edited:

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
Which equilizer?

Talking about the FiiO X5-II DAP. It has a built-in equalizer. That's what I meant.
It helped prevent known clipping issues for the the original X3.

The head set I'm using is a common and basic studio mixing/recording pro model, sensitive wide band (5-30kHz),

please specify the headphone (the 5-30kHz spec, when it happens to be a Sony, is given at -20dB cutoff points.

I really can't say why you are suddenly hitting issues with several DACs.

Edit: Lowered the signal by 2dB in Audacity (so signal away from clipping region). Same issue. It is related to spectrum in the track, on some positions;

So not a DAC issue. crappy recording, case solved ?

I just listen to music for enjoyment. Indulged you with the YT video just to see if your observation made sense.
Whenever I encounter something that repeats itself in the recording I just ignore it.

Already found out my hearing limits and abilities decades ago so am not anal about these things anymore.
This allows me to enjoy music. It must sound good. Crappy recordings I just don't listen to.
Don't do DSD nor anything over 96/24 as this already exceeds my hearing by many factors.
Also auditioned expensive high end stuff with TOTL headphones, DACs, amps etc in case you wonder.
 

b4nt

Addicted to Fun and Learning
Joined
May 29, 2021
Messages
803
Likes
270
please specify the headphone (the 5-30kHz spec, when it happens to be a Sony, is given at -20dB cutoff points.

See the attachment. It is also well known that headsets aren't that flat over the range.

I really can't say why you are suddenly hitting issues with several DACs.

Who said several "DACs"?
- the entry range Fiio I gave to my daugther had less bass in his output (the amp, compared to the X5II)
- with my X5II, some tracks are bad, seem to be crappy, but the exact same tracks with other DACs are Ok...

So not a DAC issue. crappy recording, case solved ?

I never suspected the D50s, only the Hush and Arpo track contents, which is now confirmed.

Whenever I encounter something that repeats itself in the recording I just ignore it.

I try to get the best from my hardware and files, especially when I notice impacts on audio or in tracks I enjoy. Now I discovered that crappy tracks may be fixed using Audacity and basic EQ. Cutting out or silencing stuff inside a track not being a so good option.

Don't do DSD nor anything over 96/24 as this already exceeds my hearing by many factors.

The major limiting factor I do see being price of tracks or albums. Plus the storage space that could eat up.
 

Attachments

  • DT770.png
    DT770.png
    234.8 KB · Views: 72
Last edited:

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
See the attachment. It is also well known that headsets aren't that flat over the range.

Ah DT770, owned that one for a while.

The major limiting factor I do see being price of tracks or albums. Plus the storage space that could eat up.

Yes, while offering no sonic benefits

the entry range Fiio I gave to my daugther

X1, M3 ?
 

b4nt

Addicted to Fun and Learning
Joined
May 29, 2021
Messages
803
Likes
270
Don't remember, and that Fiio is gone.
 

b4nt

Addicted to Fun and Learning
Joined
May 29, 2021
Messages
803
Likes
270
The linked YT video: I can' t hear nor see anything ' unexpected' at 0:47

please specify the headphone (the 5-30kHz spec, when it happens to be a Sony, is given at -20dB cutoff points.

So you didn't see neither hear something unexpected there.

I now made another test. Connecting more basic Yamaha NX-50 to my PC audio jack, used them instead of the Topping, amp and speakers. Their specs are here:
https://fr.yamaha.com/fr/products/audio_visual/desktop_audio/nx-50/specs.html#product-tabs

Listening to that YT video, those NX-50 pop at 47". Then later again, and crackled also at around 1:00".

For no pops or crackles in that YT track, I need to switch to laptop entry range mini speakers.

Ah DT770, owned that one for a while.

Could you please describe what the specs of your speakers or headsets to evaluate an audio track and others opinions about sounds currently are?
 
Last edited:

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
I used HD800 on Kameleon amp with E30 on Linux Mint 20.1 and on Dali speakers using refurbished and improved Quad amplifiers.
The linked YT video sounds poor quality to me overall.
 

b4nt

Addicted to Fun and Learning
Joined
May 29, 2021
Messages
803
Likes
270
Maybe check how you improved your Quad. Mine is not, and I get that pop at approx 47".
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
I do not hear any pop not on speakers nor on headphones.

On the quad FR was simply extended in both extremes. This is not a FR response issue though.
At which sample frequency and mode was the D90SE plot made ?
 

b4nt

Addicted to Fun and Learning
Joined
May 29, 2021
Messages
803
Likes
270
I saw nothing more there for that FR plot. If I would have to plot a DAC outputs frequency range (for RCA or XLR), I would use any high sample rate like 192k or more. It shall so not be flat up to 88.2k-100k, but limited, depending on the output stage and line load (line/output buffering capabilities).

Somehow Amir plots this, partially and for the audible band, when he cheks DAC linearity from 0 to 20k, using 44.1k.

About those pops (and later also smaller crackles in that same Hush track), you shall get them, as I get them here, even on basic speakers (not on my laptop mini speakers, but they are very limited).

Maybe you listen differently than I do? Especially for such a track, you do not like? I'm sensitive to any clics, pops or noises, even smaller ones. Whatever the general track quality is. Those I hear in that Hush track don't seem to be part of the artistic project, that is the point, so they are/seem disturbing to me.

Listening the way I do, I noticed I was sometime getting additionnal noises with my D50s setup, noises not being in tracks. D50s setup: not only the DAC, but PC, amp, DAC and DAC PSU. I cleaned up my groundings (potential source of EMC/mains noise), my USB VBUS, and moved away my phone (high power and RF signals too close to DAC/RCA wires).

For pops in Hush track, one could use the D.zip attachement I added here, will be easier:
https://www.audiosciencereview.com/...-khadas-tone-board-1.24596/page-8#post-833718
- the Original, with the 47" shape/pop (not so loud, but can be distinguished)
- the Different file, in which I aplied EQ to the track, but only on that pop
- the track shape is different in EQed file, at some part, see below picture
- looping them with Audacity, the Original shall sound different, poping at around 2"280...
 

Attachments

  • Difference.PNG
    Difference.PNG
    23.6 KB · Views: 96

b4nt

Addicted to Fun and Learning
Joined
May 29, 2021
Messages
803
Likes
270
That is another nice noisy sample track. I think this is an artistic choise. It features common noises (like clics), but also a low weird noise.

If you do not hear the clics, neither the lower noises, you probably do not need to upgrade from a Khada Tone Board :)

 

b4nt

Addicted to Fun and Learning
Joined
May 29, 2021
Messages
803
Likes
270
I think this is an output line buffer (or filter) of the Khada. It isn't a basic non inverting op-amp:

KhadaOut.PNG
KhadaRCA.PNG
 
Last edited:

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
It is a (filtering) I/V converter.
The output goes to a differential (also filtering) opamp which is also the line driver (100 ohm out).
The I/V converter part is not connected to the RCA out.

Designers like those of Nagra may have cared about this effect on signals.

Most designers care and simply follow the suggestions made by the DAC manufacturer. Nagra is in no way different from other designs in this aspect except in pricing perhaps.
 
Last edited:

b4nt

Addicted to Fun and Learning
Joined
May 29, 2021
Messages
803
Likes
270
Yes, filtering, I see caps in there. That part being static, defined by components, not by any rate/DAC filter settings.

I do not know how the D50s is designed. Thats is from its review thread. I see there C71, C72, .., C73 around U4, the output IC, I guess.

This just to say output stage circuits may vary between DAC models or designs.

D50sU4.PNG
 

b4nt

Addicted to Fun and Learning
Joined
May 29, 2021
Messages
803
Likes
270
furthermore with the Khadas approaching the ear to a speaker I heard a rustle (inaudible from a distance and therefore irrelevant), I tried to do the same with the Loxjie but the rustling was gone; perhaps because the Khadas was powered by the Raspberry while Loxjie has its own built-in power supply.

Did you check and maybe swap your Raspberry 5V PSU?

All new DACs come now with very similar performances (same chipsets, and very low distortion). Difference being mostly their PSU (switching or linear) and their case and shielding (the Khada comes with no case).

My D50s came with a case, is so shielded, my RCA cable is correct, but I still get low level supperposed noises if my phone is at less than a meter from my DAC and desk....

The differences you describe could very well be because of the 1.2dB higher output level of the D30.

In any next response, you added you consider also your auditive limits, you know. And that you won't go above 24/92.

I'm wondering now what 24 bits really add. How many bits DACs reproduce in their output, and how this improves the sound or not compared to 16/44.1.

And I'm not sure I'll find the answers in the reviews here (neither elsewhere). That is from the D90se, where we see 130dB for SINAD (but human ear has a limit of 100bB if I'm right):

T90se.PNG



I found that page, about materials (CD, hires, SACD) and distortion they come with, due to quantization noise: https://www.hi-fiworld.co.uk/index.php/cd-dvd-blu-ray/71/162-distortion.html


LowLevelDistortion.PNG


The reviews here for multitone are given for 0dB. Is this enough, or which plots should I read for what could distinguish DACs for signals at -60dB?

Would that d90se (or any other DACs) reproduce 24 bits sources at 0,04% for -60dB, or at 0,02% at -60dB for SACD source?
 
Last edited:
Top Bottom