SONCOZ SGD1 Audio DAC Review

[WolfX-700]

Since this thread is a Review within a Review, could we at least agree to to have each review in its own thread?
Who’s to respond to questions on measurements, original reviewer or second reviewer?

This is my problem. When I comment in this article, I just want to enhance the credibility of the device. Because it's incredible to use 9038Q2M to achieve such a measurement result!
At the same time, I may have considered too much. I think that the measurement results are consistent, so I can quickly respond to some additional measurement requirements, which I think can reduce the pressure on Amir's work.

 

[somebodyelse]

Amirm, I'm not familiar with the term rotary encoder, can't recall you calling that out before. What is significance of that component? Beautiful unit.

In this use it's functionally equivalent to up and down buttons and an enter button, but with better ergonomics.

 

[ElNino]

0dB is maximum value of a sampled point, but "true peak" or "intersample over" of the recording could be much higher:
- https://techblog.izotope.com/2015/08/24/true-peak-detection/
- https://www.audiosciencereview.com/forum/index.php?threads/intersample-over-test.3730/
- http://www.analog-precision.com/Dow...rsample_Overs_using_the_Intersample_Guard.pdf

For example 20kHz square wave at 0dBFS sampled at 44100kHz contains intersample overs with +8.15dB amplitude. These will be generated during oversampling in the DAC.

Unfortunately there's no upper limit on intersample peaks; it's possible to construct pathological signals with arbitrary amounts of intersample peaking.

There's a really good thread over at HydrogenAudio where one person measured intersample peaking on his actual library of several thousand songs. You cover the vast majority (over 99.8%) of actual recordings with 3.5 to 4 dB of headroom, although this person did find one heavily-processed song that required just over 11dB of intersample headroom.

 

[Noob]

I just came to acknowledge that this is the best looking DAC I have seen. The enclosure for this is a work of art. Beautiful.

 

[Alexanderc]

Hey all. I have searched through several threads looking for an answer to the, “is the ESS hump audible?” question, to no avail. Lots of people have asked and I’m wondering if someone who knows can give a reply for us non-engineers. Apologies if this has been discussed; I didn’t find the answer.

 

[gvl]

I don't think there's a definite answer to that question, but the numbers suggest it can be.

 

[ElNino]

Hey all. I have searched through several threads looking for an answer to the, “is the ESS hump audible?” question, to no avail. Lots of people have asked and I’m wondering if someone who knows can give a reply for us non-engineers. Apologies if this has been discussed; I didn’t find the answer.

We'd need an ABX test to know for sure. The hump is measurable, so it does provide at least a plausible mechanism for the claim subjectivists sometimes make that some Sabre-based DACs have a "sound", as opposed to voodoo differences that are not even measurable. But the hump is not enormous and relatively low in level, so it's also entirely plausible that it isn't audible.

 

[pozz]

I'm kind of confused by the backlash against posting the ESS hump. Sure it looks bad for the manufacturer, but then are we supposed hide that information?

Unfortunately there's no upper limit on intersample peaks; it's possible to construct pathological signals with arbitrary amounts of intersample peaking.

There's a really good thread over at HydrogenAudio where one person measured intersample peaking on his actual library of several thousand songs. You cover the vast majority (over 99.8%) of actual recordings with 3.5 to 4 dB of headroom, although this person did find one heavily-processed song that required just over 11dB of intersample headroom.

Can you please link the thread?

 

[gvl]

So this is why the KTB balanced went cold.

Seems they have another cheaper DAC in the pipeline - LA-QXD1, that looks more like the KTB balanced.

 

[ElNino]

I'm kind of confused by the backlash against posting the ESS hump. Sure it looks bad for the manufacturer, but then are we supposed hide that information?

Can you please link the thread?

I think it's this one: https://hydrogenaud.io/index.php/topic,98753.25.html Be sure to read all the way to the end, because the first attempt at doing this had some mistakes in the methodology. The worst album was Merzbow Venereology, where one track had 11.3dB of intersample peaking.

 

[qwerty88]

@amirm , @Thomas savage & @WolfX-700 ,

why don’t you touch base with each other, sync your planned devices up for review. As much as I appreciate your reviews seems a bit waste of resources measuring same device twice.

n of 2 is better than n of 1, which serves to bolster the reliability of each of their results.

 

[gvl]

Dear friends with APs, and I mean @amirm or @WolfX-700, can we please see the 32-tone test done at -27dB level?

 

[suttondesign]

What’s the point of the coax digital out? Because there is a streaming board in it? Is it post volume or (assuming digital) pre?

I've been waiting for a good switchable-source DAC with a digital out so that I can route the signal to my active DSP speakers (Dutch 8C). The REM pro version of the ADI-2 has it, but there's no remote control.

This unit was obviously designed for me.

 

[RayDunzl]

The worst album was Merzbow Venereology, where one track had 11.3dB of intersample peaking.

I don't think I'll worry about that one...

https://merzbow.bandcamp.com/album/venereology

 

[Soniclife]

@amirm , @Thomas savage & @WolfX-700 ,

why don’t you touch base with each other, sync your planned devices up for review. As much as I appreciate your reviews seems a bit waste of resources measuring same device twice.

My view is the people doing work on their own time get to chose what to do, with whatever criteria they chose, they are not employees to be managed efficiently.

I've liked the dual contributions in this thread FWIW.

 

[WolfX-700]

Seems they have another cheaper DAC in the pipeline - LA-QXD1, that looks more like the KTB balanced.

The developer told me that this unit is not ready, and I will measure this unit when it is ready

 

[WolfX-700]

Dear friends with APs, and I mean @amirm or @WolfX-700, can we please see the 32-tone test done at -27dB level?

 

[gvl]

Thank you. Interesting, that SMPTE suggests there can be IM components above - 100dB at this level, yet there's none visible in the multitone test. Perhaps I'm misreading SMPTE.

 

[Soniclife]

View attachment 41953

So at the worst part of the hump the bad stuff is still -116 down?

 

[bequietjk]

Oooo, at 220W x 250L it will make for a good bottom stacker.