[MRC001]

Mine isn't for sale!

The WM8741 isn't SOTA anymore (not based on the numbers) given new offerings from Sabre & AKM. But some of these DAC chips can be finicky to implement, as evidenced by the Sabre "IMD Hump" that many DACs using this chip had, and some did not. But I think we don't hear the DAC chip as much as we hear how it is implemented on the board and the downstream analog stage. The Soul has some nice DSP features, but IMO the analog stage is what sets it apart from other DACs and where it really shines, with its switched power supplies, balanced topology, frequency-shaped internal feedback, and unique volume control.

 

[Tekunda]

Has the Corda Soul ever been tested against the DCS Bartok? Any impressions so far?

 

[wantan]

Question to the SOUL owners, which filter settings do you use (linear phase or minimum phase)? And do you hear a difference between them or do they sound the same?
Do you hear a difference between redbook and hires on minimum phase settings?

 

[MRC001]

To better understand what these filters do, I've measured the Soul's response and I've done extensive comparison listening including blind testing. Details here and here.

In summary, with most music I can't hear any difference between them. That's as expected, since (A) there's no real difference between the filters below 15 kHz and most music has little or no energy above 15 kHz, and (B) most people can't hear above 15 kHz anyway. It's not that unusual not to hear any difference; I'm in good company with Keith Howard from Stereophile.

That said, I can hear a difference with certain sounds. The easiest (relatively speaking) sound to hear a difference is a square wave. Play that while switching back & forth between the filters, and the minimum phase filter (switch upper position) has just a tad more "zing" to the sound. Or at least that's how I perceive it subjectively. Other signals where I can hear a difference are: a high quality close-miced recording of jangling keys on a keychain (a trick recording engineers sometimes do to test HF response), and 1 or 2 very high quality recordings of castanets.

All that said, even when I can hear a difference, I can't say which I prefer, or which sounds more natural or realistic. Because of this, I leave the switch in the linear phase (lower) position. This is the "correct" response from an engineering perspective.

Interesting note specific to the Soul. At 44.1 and 48 kHz, the switch changes 2 things. The upper position is minimum phase, and it is also slow rolloff. At higher sampling rates (88.2, 96 and 192), the switch changes only 1 thing: linear vs. minimum phase. At these higher rates, the Soul always uses a slow roll-off filter. This in my view is good engineering design because slow roll-off filters are more transparent in the passband than sharp filters. At low rates, you might not want to use slow roll-off because it starts attenuating below 20 kHz. But at high rates, there is no drawback to slow roll-off, so you might as enjoy its advantages.

The filter selections are just one example among several others showing the careful attention to engineering detail that went into the Soul.

 

[ujiko]

I just purchased a Soul from another head-fier, Sound Eq. I am having trouble installing the driver for windows 10. I downloaded it from Meier's website, then opened device manager and the Soul is not showing up except as SPDIF output. I test the output and get an audible indicator that that output is going to my speakers thru the Soul but that is the only sound I can get (the test signal). When I click on the driver folder it just opens. I am new to using PCs, have been a Mac user all my life.

Any advise?

 

[MRC001]

I suspect this is an issue with Windows, not with the Soul. That's because every device I've plugged into the Soul's USB input: my phone, tablet, desktop & laptop computers, all work without any need for drivers. But I run Linux, not Windows or Mac. The Soul's USB input actually works better with more different devices than the USB input of the Oppo HA-1 that I owned prior to the Soul.

Stupid question: did you actually install the driver? You said you downloaded it and opened the folder, but you didn't say that you actually installed the driver. Also, with Windows you often must reboot after installing drivers. You say you're a Mac user - do you have a Mac that you can connect to the Soul's USB input?

 

[ujiko]

I can't figure out how to install the driver. When I click on it it opens but that's all it does. It doesn't open an install program. I did reboot the computer. The Soul works great on my Mac. I will email Jan.

 

[MRC001]

A driver is often delivered as a ZIP file. After downloading it, you open it and you can see the files it contains. One of these files is usually the installer program. Double-click it to run it. Usually these programs have a file extension of EXE or MSI, but Windows sometimes hides extensions so you might not see that.

 

[ujiko]

Thanks I will give that a try.

 

[wantan]

To better understand what these filters do, I've measured the Soul's response and I've done extensive comparison listening including blind testing. Details here and here.
...

Great blog articles, thanks! I like your approach discussion of the measurements of the Sooul and also the Corda Jazz (which I happen to own, too).

When I had the RME ADI 2 I did some blindtests on filters, it's easy to do with the programmable remote controll, and I can usually hear the difference. I'm using a DACCORD DAC, this doesn't have the option to use the linear phase filter.The two units should sound fairly simillar. I can hear the difference between Redbook and hires on the DACCORD and attribute this to the filters as the difference in the files shouldn't be audible. I't's just guessing though, so I was curious what the SOUL users would hear.

I also have experimented with upsampling (SoX upsampler) but this seems to change the SQ in itself, hard to confirm this with a blindtest though because changing the settings need a moment.

I'm using a DACCORD FF with JAZZ FF and CLASSIC FF. All very nice sounding units. I envy you gus SOULS.

 

[jokostyle]

I'm using a DACCORD FF with JAZZ FF and CLASSIC FF. All very nice sounding units. I envy you gus SOULS.

Hi,

For curiosity's sake : which Daccord output do you use ? The variable or the fixed one ? If you didn't try the variable output, can you make a test ?
If you tried the two, what were your findings ?

Thanks

 

[wantan]

Edit:
I withdraw the comment about the Daccord preamp because there are some technical variables I hadn't considered. So I've done some further comparisons with a Phonitor e on the fixed and the Jazz FF on the variable (variable at 3o'clock, Jazz high gain at 9o'clock and HD650) and there's no loss of anything at the Corda Jazz FF site.
(It's impressive enough that the Jazz FF holds it's own against the Phonitor....)

original comment

Spoiler

Hi,
I've used both frequently. The fixed output has slightly better microdynamics and a clearer, blacker background. The variable output is a bit smoother. These are small differences and if you want to attentuate the signal of the Daccord the variable output is the best way to go, it's a very good solid state preamp and close to tranparent.

 

[MRC001]

...The fixed output has slightly better microdynamics and a clearer, blacker background. The variable output is a bit smoother....

To compare the fixed vs variable output, you'd have to get them to the same level, which means attenuating the fixed output. How did you do that? And how do you know whether the differences you hear is due to the fixed vs variable output of the Daccord, or due to the method you used to attenuate the fixed output?

For example, if you use a preamp to attenuate the fixed output of the Daccord, then any differences you hear could be due to the preamp, not to the Daccord itself.

 

[wantan]

The differences are not related to volume and it's not necessary to level match when comparing audio tech. It helps with blind tests to confirm differences but even there variable volume can be used because any error will average out over multiple tests.
The variable output has higher max volume than the fixed, so you can just attentuate this for level matching. I've actually compared the Schiit Sys to the variable output and it's pretty much the same. The Sys has a very good reputation for excellent sound (or hardly any sound) and good measurements. The variable output is excellent, I use it quite often, for example to listen to amps on high gain.
There are plenty of options how to compare audio tech if you want real objective results.

If you like some experiments try to compare redbook to highres on the SOUL using the minimum phase filter. You can also try upsampling with SoX. I'd be curious to learn what you hear.
What do you use as a digital transport in your system?

 

[MRC001]

... the differences are not related to volume and it's not necessary to level match when comparing audio tech. ...

Level matching is essential to critical comparisons, since even slight differences in level mask other differences. But that's a whole 'nuther subject, no need to rehash it here.

... if you like some experiments try to compare redbook to highres on the SOUL using the minimum phase filter. You can also try upsampling with SoX. I'd be curious to learn what you hear.

Subjectively, at higher sampling rates (88.2, 96, 192) I must admit I can't hear any difference between the filters. Objectively, they also measure nearly identically, at least up to 20 kHz. IMO, which digital digital filter one uses only makes a difference at lower sampling rates like 44.1 and 48. Unless the filter is a really bad implementation, but that doesn't apply to the Soul.

What do you use as a digital transport in your system?

Most of the time, the SPDIF output of an Oppo BDP-83 playing CD and DVD-A. Sometimes, the SPDIF output of this PC's sound card (an ESI Juli@). Sometimes, the USB output of my Android phone or tablet (using USB Audio Player Pro).
That said, I don't think the digital transport matters much, as I understand the Soul's digital SPDIF inputs reclock the data, and its USB is async. Bits are bits. I suppose a really bad digital transport could drop bits or have such poor timing it loses sync, but that seems to be the exception not the norm.