• 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.

Review and Measurements of Budget ($20) DACs

SiW

Active Member
Joined
Oct 24, 2018
Messages
140
Likes
42
#21
I got a D03K but get a lot of background hiss - could this be usb power supply related?

My CYP DAC is totally silent.

Simon
 

SiW

Active Member
Joined
Oct 24, 2018
Messages
140
Likes
42
#23
Toslink (waiting for a coax cable to arrive from Amazon) and both off the same usb power plug which is rated at 4.8amp.

Will try different usb cable and power plug incase it’s a ground issue.

I’m using my iPhone with Hifime Sabre toslink out adaptor. Then DAC into a Fiio A3 amp.

Waiting for my Topping D10 to arrive :D

Simon
 

edechamps

New Member
Joined
Nov 21, 2018
Messages
2
Likes
0
Location
London, United Kingdom
#24
Thanks for the review, and sorry for resurrecting this thread, but I found this part very intriguing:

Next I tested 14 Khz tone using 48 Khz:

View attachment 5378

Good grief! Clearly neither SIGNSTEK or FiiO have proper 48 Khz clock. SIGNSTEK's distortion products are just 50 db below its signal peak!
I found this quite shocking, as I had no idea the performance of such devices could vary so much between 44.1 and 48 kHz. Out of curiosity, what makes you say that this is distortion is specifically caused by running these devices at 48 kHz? I mean, between the measurements at 11.050 kHz @44.1 kHz, and 14 kHz @48 kHz, two variables have changed: the frequency of the test signal, and the sample rate. What makes you say the horrible distortion is caused by the latter and not the former? I'm asking because I'd like to come up with a test methodology that covers such pathological behaviour.
 
Joined
Sep 24, 2018
Messages
38
Likes
5
Location
Somerset, England
#25
BTW, should have said - I went back to directsound with Kodi (even though it will resample to a specified sample-rate with Wasapi) so I could keep digital volume control , which you lose with digital-out via Wasapi (Foobar has its own software attenuation).

Got to say though - Sox is noticably smoother and less 'gritty' down-sampling hi-res files than Windoze. I was just listening to Fleetwood Mac Rumours 24/96 via Kodi, and when I got to "Songbird" I kind of winced a few times. Stopped and switched to Foobar (WASAPI 'event', Sox) and found it much more relaxing.
 
Last edited:
Joined
Sep 24, 2018
Messages
38
Likes
5
Location
Somerset, England
#26
...

I found this quite shocking, as I had no idea the performance of such devices could vary so much between 44.1 and 48 kHz. Out of curiosity, what makes you say that this is distortion is specifically caused by running these devices at 48 kHz? I mean, between the measurements at 11.050 kHz @44.1 kHz, and 14 kHz @48 kHz, two variables have changed: the frequency of the test signal, and the sample rate. What makes you say the horrible distortion is caused by the latter and not the former? I'm asking because I'd like to come up with a test methodology that covers such pathological behaviour.
As I said earlier - "no such thing as a free lunch". This is a pared-to-the-bone box sold for absolute peanuts, and I think we should be grateful for something that will play our Red Book / 44.1 files so sweetly and so cheap.

Whatever the cause, the measurements don't lie - it makes a mess of 48 kHz and up, which I noticed pretty quickly when listening to 96/192 files, and is the reason I found this thread.
 

scaprile

New Member
Joined
Jan 10, 2019
Messages
3
Likes
0
#27
between the measurements at 11.050 kHz @44.1 kHz, and 14 kHz @48 kHz, two variables have changed: the frequency of the test signal, and the sample rate. What makes you say the horrible distortion is caused by the latter and not the former?
(hello everyone!)
I'd have to do my maths, and I'm getting older too, but those peaks look like digital artifacting caused by sample rate conversion or multiplier-accumulator saturation. If I had to bet, I would surely bet high on Mark's claims. Then, loosely bet on saturation somewhere, maybe unnoticed while generating the test but again it looks like the device is not able to handle this. I would not place my bet on 14 vs 11 because it is way below Nyquist, even when sampling at 44.1, and I would expect to see less peaks, both ar harmonic and aliased frequencies.
In my opinion, odds are this cheap device is designed to handle 44.1 and when driven at 48 it either does a lousy job by poor sample rate conversion or simply can't cope with it (but I guess this would probably generate more hearable effects)
It may not, it may just be 14 vs 11, but I won't bet on that.
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
5,930
Likes
1,876
Location
Riverview, FL
#28
Here's an unsmoothed sweep of the signstek digital output (USB -> Optical) sent to and decoded by a real DAC, at 44.1 and 48khz for your amusement.

1547189968291.png
 

scaprile

New Member
Joined
Jan 10, 2019
Messages
3
Likes
0
#29
What is that artifacting beyond 5K on the red channel ? I'm curious.
Can you get an impulse response or an idle channel noise measurement ?
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
5,930
Likes
1,876
Location
Riverview, FL
#30
What is that artifacting beyond 5K on the red channel ? I'm curious.
Can you get an impulse response or an idle channel noise measurement ?
I have old measures handy. Will update this post shortly.

No, not now. Maybe later...

 
Last edited:

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
5,930
Likes
1,876
Location
Riverview, FL
#31
What is that artifacting beyond 5K on the red channel ? I'm curious.
Can you get an impulse response or an idle channel noise measurement ?
Signstek analog out -> PC soundcard

Detail of 19 to 20kHz sweep tone at 48khz. 44.1kHz sweep is "normal".

1547258320712.png


Impulse is normal, symmetrical.

Noise not measured, but I don't remember it being a concern.
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
5,930
Likes
1,876
Location
Riverview, FL
#32
What is that artifacting beyond 5K on the red channel ? I'm curious.
Can you get an impulse response or an idle channel noise measurement ?
Impulse on the Signstek looks normal, "at" the impulse with 48kHz:

1547278222509.png


Zoom out in time, there's something...

1547278633922.png


Zoom way in in amplitide and it's strange. Same on several different 48kHz impulses. Not present on 44.1khz impulses.

1547278465349.png


Clueless...
 

scaprile

New Member
Joined
Jan 10, 2019
Messages
3
Likes
0
#33
Interesting!
Trying to explain what can cause such a ringing on a frequency response measurement, not knowing how the "instrument" operates... I thought that one possibility is that there is something for real in the chain that is causing such big amplitude variations at adyacent frequencies. I don't think that is likely. Intermodulation perhaps ?
Other thing I thought of, was probably an extra sound/noise appearing when excited at those frequencies, not rejected by the measuring instrument filtering (if any), or always present at those freqs and just captured when sweeping there; or both...
Impulse measurements can't usually be done on a direct form. I don't know many methods and don't know what this program uses. The portion of waveform you mention looks like a chirp used for that purpose. It shouldn't be there anyway.
I guess either your device or your board are outputting this high-frequency interference and the software captures it and wrongly interprets it as what it shows. My 2 cents.
Since it is present at 48K and not 44.1... I would bet on the Signstek.
Since nothing strange shows up on noise measurents... perhaps that is because of sample rate conversion or so at 48K... just guessing out loud...
Thanks!
 
Last edited:

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
5,930
Likes
1,876
Location
Riverview, FL
#34
The first two clean (44.1khz) and ugly (48khz) sweeps were performed within two minutes of each other, nothing changed except the REW settings.

The signstek is the only thing here that creates such interesting results on the analog output.

Its digital output doesn't create problems downstream, so, that being my intended use for it (USB->Optical Conversion), it didn't go into the trash.
 

bigx5murf

Active Member
Joined
Nov 6, 2018
Messages
156
Likes
48
#35
I have a SMSL SD-192 pro gathering dust. It's very similar to the fiio's form factor, but with significantly better specs. I wonder how it compares.
 
Top Bottom