Pro-Ject Pre Box S2 Digital: MQA HW decoding at reasonable cost

[asdf1000]

17 minutes ago, Miska said:

 

IIRC, it has 3rd order IIR filter with adjustable corner frequency of 50, 60 or 70 kHz and I believe newer versions added something like 42 kHz too. Datasheets are not really clear on how things exactly work, but I have some thoughts about how things are implemented internally. With DSD the on-chip processing is quite minimal (this is also why it has only 1:3 clock ratio requirement).

 

 

It has typical slight analog filter as function of the D/A stage.

 

The on-chip filtering corner frequency moves as function of sampling rate. So the quoted "50 kHz" for example is relative to DSD64. For DSD128 that means "100 kHz".

 

 

I haven't seen indication of such, but probably depends on the DAC architecture.

 

I didn't make jitter measurements at other DSD rates than DSD512. But at least at DSD512 the result is practically identical to that of 768k PCM. While 44.1k input rate is slightly worse. But even this is not due to clock's phase noise, but rather related to digital domain activities.

 

44.1k/24 PCM input:

 

768/32 PCM input:

 

DSD512 input:

 

 

Nice, thanks for sharing.

 

So even 768/32 PCM up-sampling performs really well.

 

From all the measurements you've done with this DAC (not only the J-test), in what areas does DSD512 up-sampling perform better than 768/32? Or do they measure essentially the same, for all your measurements?

 

[asdf1000]

@Miska thanks for sharing! For the geeks like me that find that stuff interesting :-)

 

[asdf1000]

53 minutes ago, Miska said:

only other peaks visible are the UAC2 protocol USB packet ticking at 8 kHz harmonics (8, 16, 24, etc).

 

You've never used or considered something like the Intona, to take care of these? I'm using one now and it has no problems with PCM768/32 . It only has issues with USB DACs that are fully bus powered, limited to ~300mA at 5Vdc

[asdf1000]

1 hour ago, Miska said:

have not measured other DSD rates than DSD512 yet. I can try to do those other rates next time I'm measuring something on my office desk...

 

This would be cool, when you get the chance

[asdf1000]

On 17/02/2018 at 10:16 PM, Miska said:

For sake of completeness, here's Jtest-24 figures for coaxial S/PDIF input.

 

44.1/24 input:

 

192/24 input:

 

These measurements seem to show similar jitter performance between SPDIF and USB:

 

https://audiosciencereview.com/forum/index.php?threads/review-and-measurements-of-pro-ject-pre-box-s2-digital.2370/

 

Different measuring equipment and methods I'm assuming.

 

His initial comment/results are also than enabling distortion compensation makes things worse but he says he wants to test this some more.

[asdf1000]

1 minute ago, Miska said:

Looks like they had the distortion compensation disabled (default setting) in their measurements...

 

His initial comment/results are that enabling distortion compensation makes things worse but he says he wants to test this some more.

[asdf1000]

3 minutes ago, Miska said:

 

I don't use "Best" or "Test" mode. I have "Audio Quality" = "User"... Distortion Compensation = Enabled, Filter = Sharp Rolloff

 

 

Yeh let's see what he reports after more testing of it.

[asdf1000]

1 minute ago, Miska said:

Anyway, "Best" is certainly all but "Best".

 

 

Ha, my ears agree.

 

Most importantly, overall, he also found it measures well! Not surprising given our listening impressions and Miska's measurements and recommendations.

[asdf1000]

@Miska what is actually happening with the DSD512 signal inside this DAC's ESS chips, from the input to the analogue output, what happens in the "in between step" in the chip. I understand there's no further oversampling of the DSD512 dignal internally? So does this 1bit signal just go straight to analogue conversion?

 

And what happens internally with the PCM768/32 signal? This gets converted to a 1bit signal? At what sample rate? before getting converted to analogue.

 

ESS used to have lots of info available explaining this stuff on their website but now it's hard to find.

 

I don't know if/how this particular 9038PRO mobile chip has changed from their past designs, in how conversions work internally.

 

Cheers

[asdf1000]

34 minutes ago, Em2016 said:

@Miska what is actually happening with the DSD512 signal inside this DAC's ESS chips, from the input to the analogue output, what happens in the "in between step" in the chip. I understand there's no further oversampling of the DSD512 dignal internally? So does this 1bit signal just go straight to analogue conversion?

 

And what happens internally with the PCM768/32 signal? This gets converted to a 1bit signal? At what sample rate? before getting converted to analogue.

 

ESS used to have lots of info available explaining this stuff on their website but now it's hard to find.

 

I don't know if/how this particular 9038PRO mobile chip has changed from their past designs, in how conversions work internally.

 

Cheers

 

Here it's mentioned : "the Sabre uses asynchronous sample rate conversion on all input data and converts into a clock rate of 40MHz"

 

So does this mean both DSD512 (1bit) input and PCM768/32bit inputs, are 'converted' to 1bit 40MHz by each ESS9038Q2M chip? ?

 

[asdf1000]

2 hours ago, left channel said:

Thanks @Nihilnisibene. If the response tells you to contact them through the website, note that your messages will go to a local dealer unless you select as your location a country that has no dealers. Like Bermuda.

 

Very true. Although I like the Facebook channel. Perhaps applies a little more pressure with more eyes watching, in the case of bug fixes etc.

[asdf1000]

8 minutes ago, left channel said:

 

Yes, especially since their "tech support" is the same PR/marketing staff in charge of the Facebook page. Maybe even the same one guy.

 

LOL. With 'more eyes watching' I meant more public (current and/or potential customers) eyes watching.

 

Agreed though, it's probably a one man band at the Pro-Ject end.

[asdf1000]

3 hours ago, 57gold said:

S2 has more technology than the Liberty...and my son has grown up with devices (mandatory laptop in 5th grade), studied advanced math at U and writes code for commodity trading as part of job.  

 

So, the S2 with fancy chips, MQA, accepts high sample rates...a techie barrel of monkeys for not a lot of $s.

 

Let's not forget the late Charles Hansen's (Ayre) post on what he considers the most important features of a DAC's design.

 

Fancy chips are down the list. Having said that, Myek make really great stuff and this S2 DAC is no slouch.

 

 

[asdf1000]

28 minutes ago, mfsoa said:

FYI I ordered a unit from my dealer (my brother) in the US (NJ) and was told it won't be available until late May.  I told him to get a bunch...

 

-Mike

 

Audio Advisor in the US say they expect stock in both silver and black, in April.

 

[asdf1000]

3 hours ago, exdmd said:

• The “Distortion Compensation” option works backwards currently. Turning it “On” in the menu increases distortion. Turning it “Off” reduces this. It is known to be a simple flag-inversion and will no doubt be fixed in the next update.

 

 

Hmm more different observations about distortion compensation.

[asdf1000]

4 hours ago, left channel said:

That's just firmware. Don't get me started on their parts sourcing.

 

Please do share :-)

[asdf1000]

2 hours ago, fgribas said:

I have a 5v USB power-bank. When I connect the S2 to the 2.1A port of the power-bank, the sound is better than using the included wall-wart. When I use the 1A of the same power bank, the sound is even better. Must be something related to the low quality voltage regulators on the cheap power-bank.

 

I noticed the same. I measured and the external 5Vdc input of the S2 DAC only draws a maximum of 0.5Amp, so a 1Amp PSU or powerbank is all you need.

 

I thought that even though the DAC can work powered by USB 2.0 only, I thought maybe with external power it draws a little more power but it doesn't - it's still 0.5Amps max drawn.

[asdf1000]

Pro-Ject are coming with their own battery based PSU soon - a battery box known as Accu Box S2 USB

 

Might be worth waiting for this.... 

 

http://pinkfishmedia.net/forum/threads/pro-ject-pre-box-s2-digital-dac.201691/page-24

 

No word yet on it's battery life but if it's anything over 8 hours continuous playback at max current draw, with ultra low noise regulators, I'd be happy. Just let it re-charge overnight.

[asdf1000]

9 hours ago, PunkRiot said:

So probably there is no need to spend too much time and money thinking about power supply as the small DAC with ES9311 chip is designed in a way that at least can give good performance with standard in box SMSP or in spec USB power.

 

I'm not sure the recommendation for a linear PSU (or batter) power is related to noise output - the S2 DAC has nice regulators inside as you say.

 

I think it may be more due to AC mains leakage (currents) associated with SMPS's. Linear PSU's have lower levels.

It all depends on what else connects to the DAC of course and how any ground / leakage 'loops' are formed. If only using the S2's headphone output, then none of this matters of course. But if using it as a DAC you may (or not) hear differences with a linear PSU.

 

Just a thought. The DAC sounds great with just the stock PSU. But you may (or not) be able to squeeze even better performance with the PSU side.

 

The designer himself (John Westlake) has recommended this on the pinkfishmedia forum...

 

[asdf1000]

It's a great DAC. They took the long and scenic route to get to the current stage (bumps along the way) but there's not many major issues left. And they confirmed they are working on these last minor issues.

 

Switch to linear phase + fast rolloff and enable distortion compensation (as per Miska's advice long ago) after you update the latest firmware updates and you shouldn't see any major issues in use...

 

Buy with confidence I say! Feed it with a linear PSU and you'll be surprised and pleased I think.

[asdf1000]

6 minutes ago, left channel said:

It includes a comparison to the iFi iDSD Black Label

 

Minor thing but I believe it's a comparison to the Nano model.

 

Great review and in line with what a growing number (including myself) have observed in use.

[asdf1000]

Just now, left channel said:

Thanks @Em2016 I still had time to edit that post. More proof I shouldn't be posting from my phone. Or at least that's my story, and I'm stickin' to it.

 

Ha! Let's just say (hypothetically) I'm also thankful for the 'edit' function on the forum.

[asdf1000]

6 hours ago, trl said:

Why do you want to use a power bank to power your DAC for? I found a couple of USB ports on my desktop computer with less than 1mV RMS of AC ripple & noise, which is quite acceptable.

 

One thing to consider, which may (or not) apply.

 

The S2 DAC has nice low noise linear regulators inside, if the noise output / ripple of a powerbank is an issue.

 

But the S2 DAC won't block ground loops / leakage currents / mains power RF from a mains connected computer USB port supplying bus power, which a powerbank is immune from.

 

Pro's and con's with nearly everything in this hobby, or as they say, there's no such thing as a free lunch.

[asdf1000]

1 hour ago, Miska said:

 

For ground loops, as long as you have the computer USB connected, it doesn't matter where you pull the power from...

 

But I USB cable with ferrite.

 

If the S2 DAC is being used to drive headphones (as per the photo above) then if the DAC is powered by battery then there will be no 'loop', even if the computer USB source is mains connected.

 

I know I'm not telling you anything you don't already know. Just let me have this little win ! :-)

 

[asdf1000]

@Miska what's your favourite HQP settings for DSD512 up-sampling to this Pro-Ject S2 DAC?

 

Are they the same settings you like with the iFi iDSD for DSD512?

 

Can you share a screenshot showing the exact settings so I can copy your favourite as a starting point.

 

If I don't have enough horse power after trying I'll then tweak these settings to reduced CPU needs.

 

Cheers