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

[fgribas]

Hello, can someone confirm if it's a bug or I'm a lucky one with a hardware problem?

 

Does your Distortion Compensate setting remains ON if you power off and power on the unit?

On my device, all the settings are retained (input, filter, volumes), expect for Distortion Compensate, which always comes back to OFF.
 

[fgribas]

10 minutes ago, Miska said:

 

Yes it stays enabled. Do you have the latest firmware?

 

Yes, I have 2.12 

 

All the other settings are retained, that's strange.

 

[fgribas]

Thank your for you report guys. At least 2 other users from another forum confirmed the same problem.

 

But I have some good news!

 

1- Pro-ject support kindly helped me to fix that. I have contacted them through the support site using Bermuda as my country (thanks @left channel for the tip!) They answered my message 15 hours later with the solution. Much better than my expectations.

 

2- It's not a firmware bug nor a hardware problem. I was another victim of the misleading user manual. The thing is that we need to push the volume knob to confirm the selection of the Distortion Compensate setting. I was changing the setting and pressing Menu again.

 

I replied them saying that an update to the user manual would be nice, as I was not the only one that faced that problem. I indeed remember someone posting about the misleading user interaction with the configuration menus some time ago. 

 

The sound I get from this DAC is amazing, specially by upsampling to DSD512 with HQPlayer. Great price/performance ratio!

[fgribas]

1 hour ago, Jamesroy said:

Do you have any upsampling tips for A+

Sorry, not using A+

 

All I'm listening to is Tidal Hifi in Roon, upsampling with HQPlayer.

 

Some months ago I bought a Macbook to use A+ believing it would give me superior SQ but got pretty disappointed. I even did the hack on MacOS Sierra to enable DirectMode (https://audirvana.com/?p=4048) but still worse than my Windows environment.

 

Back in the days, macOS used to be (or at least was considered) the best OS for audio. The scenario has changed. A properly optimized Windows with ASIO sounds much better to me. Sold my pretty brand new Macbook  

 

Still to try Linux  when my miniPC arrives (it will be a NAA).

 

I heard A+ will be released for Windows. I might give it a shot, but I already paid lifetime subscription for Roon + license for HQPlayer: unbeatable combo for me.

[fgribas]

5 hours ago, left channel said:

Are all digital volume controls weird, or just the one on this product, or even just my unit? I've mentioned this before but nobody else has, so I'm wondering. Sometimes when I turn the volume knob, nothing happens for a few notches (the little clicks or detents you feel when you turn it). No change in sound or display for at least three of those. After that, now that it is "awake", the same few clicks will immediately change the volume. But turning back the other way may require extra clicks. And after not being turned for a while, it's back to requiring at least three extra clicks in either direction. It's really strange.

Mine is the same. It's a little bit annoying...

[fgribas]

@Miska do you know if the internal filter is completely bypassed using any of these sample rates?

 

352.8 / 354

705.6 / 768

DSD256 44.1/48

DSD512 44.1/48

 

I found out that the internal filters are quite good, specially Hybrid Filter and Fast Roloff. But since I'm using HQPlayer to upsample, I want to bypass internal filters.

 

And I hear a noticeable change with different filters when using 44.1/48 source, but I cannot hear any difference using DSD256 and DSD512 changing the internal filters.

 

[fgribas]

22 hours ago, Miska said:

 

Those filters selectable in this DAC don't apply to DSD inputs. ESS chips have a separate filter selection for DSD inputs, but this DAC doesn't provide means to switch those, so it is always using one static choice for all DSD rates. So your listening experiences are correct that changing the filters don't affect output when using DSD.

 

For PCM inputs at rates 352.8k and higher, the on-chip digital filter should be bypassed. I didn't check those 8x (352.8/384k) rates, but they should work similar to the 705.6/768k.

 

Thanks! I'm loving this little box. First heard of it on your HQP's recommended hardware page.

Still to try a linear power supply. For now, I'm using a 5v power bank and it sounds much better than the included wallwart SMPS.

[fgribas]

On 2/18/2018 at 8:09 PM, Miska said:

 

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

 

The THD levels visible in his plots look like the ones with distortion compensation disabled.

 

But this plot looks looks somewhat like expected:

 

"Best" has the distortion compensation disabled, and "Test" has it enabled.

 

 

I've been reading all Amir's thread on his forum and now I'm quite confused. 

 

@Miska the image you pasted has the following text:

 

We see that "best" sharply raises the noise floor by some 25 dB! In doing so the noise now masks all of those distortion products. Seeing how there is some 32 dB of gain in the above measurements, the "best" mode likely has less signal to noise ratio than needed for 16 bit audio. When I have time I will do some listening and see if there is an audible difference. For now, I would probably use the DAC in "Test" mode. 

 

What I understand by that is Distortion Compensate Enabled suppresses harmonic distortions by raising the noise floor. Kind of a trade-off. Am I right?

 

After that post, Amir said:

 

I need to do more testing but in preliminary measurements, Enabling distortion compensation seems to raise distortion, not lower it! Since this is backward I like to perform more tests and only post them when I am sure.

 

So he ran more tests. Measuring a 1kHz residual noise and distortion, it seems that Distortion Compensate Enabled raises 2nd harmonics distortion, but lowers 3rd harmonic distortion.

 

 

What we see here is that when we turn off Distortion compensation, we actually get a reduction in distortion in 2nd harmonic. There is a trade off elsewhere though with third harmonic actually increasing. We also see some small spikes with compensation enabled that are not there when we turn it off. So the overall picture is a mix. 

 

 

And then a new THD+N test that really confuses me, showing that "when we sum the total distortions+noise, distortion compensation is doing the opposite of what its name indicates."

 

 

 

Here is the specific post:  https://audiosciencereview.com/forum/index.php?threads/review-and-measurements-of-pro-ject-pre-box-s2-digital.2370/page-4#post-66480 

 

After that, other users asked him to run the tests with different filters (Fast roll off is one of them), but always getting the same conclusion: Distortion Compensate Enabled actually raises the noise floor.

 

 

 

After scratching my head a little, I tried to listen the differences. I always used Distortion Compensate Enabled, but I made some A/B enabling and disabling it. I don't have measurement equipment, that's only my listening impressions. And the impression is that disabling Distortion Compensate actually gives me a darker background (lower noise floor), specially playing only vocals with no other instruments.

 

Still scratching my head...

[fgribas]

7 hours ago, Miska said:

But those measurements you refer to are self-contradictory.

 

That's exactly what's confusing. On his first measurement, THD+N was better with Distortion Compensation Enabled. On the following tests,  THD+N was better with Distortion Compensation Disabled.

 

I don't use Test or Best, as I don't like their filters. I'm using Hybrid Filter, but that doesn't matter cause I'm upsampling everything to DSD256 and DSD512. Used to have Distortion Compensation Enabled, but now I'm using it disabled. 

[fgribas]

 

16 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.

 

@Miska apart from the confusion regarding user settings, could this be the reason for what I'm talking about? It seems that on firmware 2.12 the flag inverted, and using Disabled apparently turns on this ESS chip feature. Someone mentioned it on Amir's forum also.

 

What FW version you had when you measured Dist. Compensation On/Off ?

[fgribas]

Amir's hadrdware teardown https://www.audiosciencereview.com/forum/index.php?threads/hardware-teardown-of-pro-ject-pre-box-s2-digital-dac.2393/

[fgribas]

6 hours ago, Miska said:

 

Probably I tested it with the shipped firmware (2.11?) because I did it pretty much straight when I got it. Measurement files are dated October 23/24 2017. I didn't write down the firmware version though.

 

I'll re-check again with the 2.12 firmware next time I measure something on this desk. (it takes two hours or something like that to set up things and do the measurements)

 

Great, thanks! Opinions are conflicting across different forums regarding this.

[fgribas]

Thank you @Miska

 

So you get less distortion with the feature enabled, as it really should work. As per my understanding, it's the opposite result of Amir's tests.

[fgribas]

2 minutes ago, Miska said:

 

Depends which one of his tests you check. Because they are self-contradictory. His tests show poor results with profile "Best" (dc disabled) and better results with profile "Test" (dc enabled). While elsewhere he notes otherwise.

 

 

Yes, his second round of tests were when the confusion started. I think I'll stick with your results

[fgribas]

There are a lot of SQ impressions on this thread. I will write down my opinion.

 

Well, this DAC really shines on DSD256 and DSD512. PCM 768/705 is also very good. 

 

MQA is nice, but the quality of the MQA Mastering is different on every MQA album available on Tidal. With some albums I prefer (by a small margin) the MQA decoded version than upsampling to DSD512. On other albums, upsampling to DSD (therefore not decoding MQA), sounds much better than decoding MQA. It's well known that MQA uses poor/low performance filters (even an Audioquest Dragonfly has processing power to use these filters). Sometimes the poor filter can sound nice (master engineers/record labels supposedly use the best available MQA filter to fit each album they encode to MQA), but sometimes it sounds simply not good.

 

For me, this DAC sounds amazing! Before mine arrived here, I ordered a Schiit combo: Magni 3 + Loki Mini. In a nutshell: I will sell them. I prefer the ESS headphone amp embedded on S2 than the Magni 3. The M3 colors the sound, loses details and changes the soundstage (in a bad way). The Loki equalizer is fun to play with, but can't rectify the M3 mutations on the sound.

 

Unless my RCA cables are really bad (Schiit cables by the way), I don't consider the M3 as a neutral amp, and consider a downgrade compared to the S2 internal headphone amp.

 

If someone has the initial impression of S2 as being something like "anemic", O.S. optimizations and good quality upsampling can bring a lot of weight and life to the sound. And every system is unique.

 

Every little detail changes the sound. My laptop has 2x USB and 1x USB-C ports. The sound changes drastically when I connect the S2 to each of these ports. USB-C is the best in my system. I think it is the most isolated (less shared) bus internally.

 

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.

 

When using my laptop on batteries, the sound is great, even with the noisy fan on the laptop running on to cool the 3.7GHz fixed overclock. When I plug the SMPS power supply (dell standard) on the laptop, the sound gets very worse.

 

Even by making simple optimizations on a standard Windows 10 like core isolation (processor affinity) for the player process, changing process priority (real time or high priority) of the player process and lowering the O.S. Timer Resolution can bring noticeable improvements on the sound.

 

Digital Audio is tricky and complex. You gotta deep dive and study... Or spend more money to get the same SQ.

 

I'm still waiting for a fanless mini-PC to arrive here, it will be a NAA/Roon-Endpoint. Also waiting for one Linear Power Supply for the DAC and another one to the mini-PC. They should improve even further my SQ.

 

The thing is: the DAC (or any other system component) can sound very different on different systems. And there's also personal preference... For me, it sounds great! Specially for its price and size. If you can buy it having the chance to return, you can give it a shot. I could not return it and I'm happy I liked it.

[fgribas]

2 hours ago, exdmd said:

 

Torq over at SBAF considers the headphone amp built in to the Pre Box S2 to be not so good:

 

 

I happen to have both a Magni 3 and Vali 2 and would be glad to compare a review sample of the Pre Box S2 using both amps.  No interest in buying one though.

 

Yeah, I just read that on SBAF. For me, Magni 3 changes the tonal balance (so not a neutral amp) and messes with sound stage and instrument positioning.

 

Some might like the altered tonal balance, depending on the headphone used and personal preference. In fact I tested S2 alone and S2 + Magni 3 + Loki with a Philips Fidelio L2 I have around, and I preferred S2 + Schiit stack than S2 alone because the frequency response of the Fidelio is so messed up. And a decent and balanced frequency response is more important than sound stage.

 

But I'm an IEM guy. With my Triple Fi 10 (with Null Audio Arete MK II cable), the tonal balance, correct sound stage and all the details are there with the S2 alone.

 

But this is just my subjective findings. As it also seems to be Torq's findings (I see only measurements of the frequency response using different filters there).

 

Or maybe it's just that my RCA cables are a piece of Schiit. They are Schiit Pyst, a low cost cable indeed...

[fgribas]

On 3/2/2018 at 12:17 PM, left channel said:

Also, for the record, I like what my Magni 2U does to the sound signature when I feed it with this DAC. And not just because the Magni is much more powerful.

 

 

 

@left channel almost everyone seems to love Magni 2U and specially Magni 3. I've read tons of posts on the forums around and decided to give it a try. It's a  matter of personal preference and matching your audio chain. For me, the  ES9602Q amp on S2 is a better match for my IEMs. But when using my semi-open Fidelio heaphones, I prefer Magni 3 amping. What's your headphone?

 

Oh, but one thing is for sure: the  ES9602Q of the S2 lacks some power. Even with my IEM ( 32ohm impedance and 117 dB/mW sensitivity) sometimes I reach -8 dB on the S2 (older and quiet recordings most)

[fgribas]

On 3/17/2018 at 11:51 AM, Mark Dirac said:

Thanks left channel.

 

left channel reported that he could get his S2D to work from Android if he inserted an Uptone Audio Regen. Has anyone discovered a cheaper way to inject power into the USB connection?  I imagine that a powered USB hub might work, except that (IIRC) it has been found that S2D will not work through a USB hub?

 

Has anyone tried a powered USB hub on Android?

 

 

I use a USB power-bank (1.0A) feeding S2D on its mini-USB 5v port. 

 

With that in place, I use a Samsung S8 directly connected to S2D USB. The result is very good indeed (for mobile source without upsampling). I also have been using the same setup with an iPod (and one Apple Lightning to USB 3 Camera Adapter) as my travel setup.

 

 

[fgribas]

@trl nice findings, thank you for sharing.

 

I'm aware that the voltage regulators inside these cheap power-banks are crap.

 

I bought 1 LPS for the DAC and 1 for the mini-PC (NAA). Both power supplies have not arrived yet. And so the mini-PC. It's painful to buy from overseas in my country...

 

So my main setup with the S2D is a laptop as source using Tidal over Roon. I don't have any measurement instrument, so all my tests are purely subjective.

 

So far I have tested this setup using 3 different power supplies: Included SMPS wall-wart, 1.0A port on power-bank, 2.1A port on power-bank. And the 1.0A port on the power-bank was the best. Without a power supply on the mini-USB port, my Windows won't recognize the DAC.

 

Every week I'm on a different city, and sometimes Internet/Tidal is not available. So my alternative setup is to use an iPod with offline Tidal content as source.

 

The laptop, being a noisy computer that has fans running on and is always connected to the AC thru a noisy SMPS, I always though it would be worse than the already bad power-bank.  So It didn't occur to me to test the USB ports on the laptop to feed power to S2D with this second setup. I just tested it. And hey, your right! Sounds much better! Thanks

 

Actually the 3 available USB ports results in different sound. But all of them are better than the power-bank.

 

So when I'm back to my main setup, I will test 1 laptop USB port for the S2D USB-B data + 1 laptop USB port for the S2D mini-USB power. The LPS should take 3-5 weeks more to arrive...

[fgribas]

New driver available, an updated Thesycon driver 4.47

 

https://www.project-audio.com/wp-content/uploads/2018/12/UNI_Project_v4.47.0_2018-09-05_setup.zip

 

The link will change when they update the S2D page, but this link will work for some days.

[fgribas]

"Pro-Ject Audio Control Panel" is a part of the driver. Thesycon is a company that makes the driver for Pro-Ject any many other DACs based on XMOS chip.

 

For the firmware I'm not sure,  but I think is also made by Thesycon.

[fgribas]

8 minutes ago, left channel said:

 

The earlier drivers and firmware were developed by John Westlake. John has promised us one final firmware bug fix update, and a future optional firmware upgrade that to reach the full potential of his vision for this product. But John is no longer associated with Pro-Ject, Thesycon also does firmware, and I'd be happy enough if Thesycon takes over the bug fixing for the S2D and other DACs in the S2 line.

 

Pro-Ject is aware of the need for a new firmware. I've been talking with their support about firmware and drivers. Hopefully it will be released soon.

 

Thesycon won't take over the product, unfortunately. Not for free. As far as I know, the manufacturer (Pro-Ject) pays Thesycon to develop new drivers.

 

By looking at the 3 versions of the driver we have so far (looking at the inf files, control panel etc of each driver), it seems that all of them were made by Thesycon.

[fgribas]

On 1/19/2019 at 4:39 PM, Mark Dirac said:

Has anyone noticed any improvements, or even differences, with the new (v. 4.47) driver?

 

For me, only the compatibility with JPLAY.

 

SQ is the same.

[fgribas]

It's just a matter that Tidal app doesn't support ASIO output, only WASAPI.

 

This is the main reason I use Roon, it supports output to ASIO drivers.

 

WASAPI (wth Exclusive Mode) and ASIO are both bit-perfect. But ASIO sounds better. Not every bit-perfect chain sounds the same. ASIO is better than WASAPI (many will disagree, many will say "bit-perfect is always the same" but that's not true).

 

Kernel Streaming sounds even better than ASIO, but there's a compromise with stability and compatibility. Almost no app supports KS.

 

Well, to use Tidal with ASIO you need another player that supports Tidal and ASIO output. Roon is great, but expansive as hell. You can use the default Tidal Windows app with the ASIO4ALL app: it's a virtual driver routes a WASAPI output to an ASIO driver. I used this chain before buying Roon.

[fgribas]

22 hours ago, left channel said:

 

When I play Tidal and other stuff through Squeezelite clients I'm using ASIO. Not sure I hear an improvement, and also there's no MQA in that app. But at this point I'm mostly streaming Qobuz anyway — and the Radio Paradise FLAC streams. I'll try switching between the ASIO and WASAPI options in the Qobuz app again as time allows.

 

Good to know we have a cheap alternative to Roon to get Tidal + ASIO. 😊

 

Speaking of which, yesterday Roon released a new version that's integrates with Qobuz. So I started my trial account. And OMG! Their HiRes releases are awesome! Thanks for the tip @left channel My Tidal account is in serious danger.

 

Now my chain is:

 

Roon playing Qobuz -> Upsample + EQ -> ASIO -> JPLAY -> Kernel Streaming-> Pro-Ject S2D -> Shure SE846

 

Sometimes I prefer upsampling in Roon, sometimes in HQPlayer.

Still using an iFi iPower to feed S2D, waiting for a chinese linear power supply to arrive. But the SQ is already delightful 😎