I currently have a number of fantastic products on my desk, and given this interesting opportunity to directly compare some products that are usually not something many will have heard, and if they have, will likely have heard on different systems, I wanted to do just that. Compare in a direct AB test.
I had a few questions I wanted to answer:
- How much difference does MScaler make?
- Qutest + MScaler combo vs Dave (Can you get DAVE performance for half the price?)
- MScaler performance as DDC only (no upsampling)
- Difference MScaler makes on a chord vs non-chord dac
- HQPlayer vs MScaler
- Holo May vs Dave
(If you have any other questions you want answered, ask and i'll do my best to answer)
To test this, I used an XLR switch to quickly switch between DACs with 0 delay. All DACs were grouped together in Roon so that they were all playing the same thing at the same time. Volume matching was done with roon DSP to keep things even and fair. (Volume matched to 0.1dB using motu M2 ADC)
I try to avoid DSP volume control wherever possible, but as you'll see shortly, its clear that it didn't hold the bigger dacs back so it didn't interfere with the test. And was necessary for a fair test.
When using a DAC without MScaler it was connected via an SMS200 Ultra Neo with SPS500 PSU.
The MScaler was connected to my PC using an ifi Igalvanic 3.0.
All of these devices have galvanic isolation anyway, but hey, might as well just in case.
When connecting to the MScaler, Dual-BNC was used on the chord dacs with shawline digital BNC cables.
Equipment used:
- Roon
- 3-in 1-out XLR switchbox
- Goldpoint SA2X Attenuator
- Benchmark AHB2 Amplifier
- Hifiman Susvara headphones (Also used Arya and HD800S but for the most part susvara)
Video review of DAVE + Mscaler here:
---- How much difference does MScaler make? ----
So, first thing I wanted to do was of course check that the MScaler wasn't intentionally or unintentionally harming quality without any upsampling (which the cynical side of me thought that this could potentially be done to give the illusion of improvement when actually the MScaler's "bypass" mode was simply worse).
To check that, I just connected dave by USB, and also Dual-BNC to the mscaler.
I played a few tracks and swapped between USB and MScaler input, and the first thing I noticed, the MScaler was quieter. By 3dB as it turns out.
This was unexpected, but also makes sense. Oversampling requires digital headroom to prevent clipping. Most DACs will have this as part of their internal oversampling. But the mscaler will need to attenuate before oversampling and then passing the signal to the DAC. I imagine the reason it does this even when NOT oversampling is exactly to prevent what I mentioned, having people accuse them of "faking" improvement by increasing volume when swapping modes.
I adjusted the volume, and carried on.
Absolutely identical. With perhaps a slight edge to the mscaler even. So, with that concern out of the way, how big of a difference is there between the MScaler vs native input?
Quite a big one....
The biggest difference is (and this will be a common theme in this post) staging.
I couldn't quite test going from 44.1khz to 768khz, because the DAVE takes a couple secs to recognise and swap to dual BNC, playing mono for a few secs, and making an 'instant' a/b impossible. But going from 768khz to 44.1khz is instant. No biggie, its easier to hear when something is missing rather than added anyway.
The MScaler makes a massive improvement to size of stage, and layering within it. With every element much more clearly defined in terms of its position, all whilst having everything so clearly separate, and sounding all the more resolving.
There really were no drawbacks here, and as to all the good things it's doing, that's something i'll discuss in a proper review, as the improvements and theory behind it is too much for this post, but the simple answer to this question is yes, the MScaler makes a quite definitive and impressive improvement.
---- Qutest + MScaler vs Dave----
The result for this was fairly clear. I compared this using both roon's headroom management, and the DAVE internal volume control for volume matching, and in either case, the DAVE pulled ahead of the Qutest by a generous amount.
The qutest is a slightly more laid back and warm presentation compared to the dave, though they both share the same "house sound", namely chord's signature separation and almost immunity to being fatiguing. And part of this is simply due to the DAVE's resolving power over the Qutest.
The qutest when enabling the mscaler's upsampling provided a considerable improvement to the stage. The qutest is not a dac that on its own stages particularly well, though the separation and imaging is excellent. With the MScaler though this changed and it became quite good.
Switching to listening on the DAVE though there was a noticeable improvement in air and upper treble transparency, and staging itself. Things no longer sounded in or near my head and were quite clearly staged outward.
The DAVE also beat the qutest in just about every other area. Macrodynamics and 'slam' particularly, though this is more due to the qutest being a little soft in that department, rather than the dave being better.
In fact lowend on DAVE was considerably better. The final part of Daft Punk's "Contact", the deep rumble on the DAVE was distinctly separate from the rest of the mix, whereas on Qutest it sounded a touch like it was overlaying with other stuff too much, and being a bit obstructive. The drums at 1:45 were also much less lifelike, with the kick drum timbre being much better on dave, and the snares a lot snappier.
Resolution wasn't a competition, DAVE won here hands down.
So no, unfortuantely Qutest + MScaler isn't "Dave lite", but the MScaler is 100% an improvement over stock. However, the issue is that it then becomes a ~$5k combo, and is competing with dacs like the M1SE, Holo May, Denafrips terminator and many others, none of which it does compete well with in my opinion, and so it is not a combination i'd recommend. MScaler is great, but you shouldn't be spending more on it than you are on your DAC.
---- MScaler vs SMS200 Ultra as DDC only (no upsampling) ----
This was a bit tricky to test, because all three of these dacs feature galvanic isolation on the USB inputs, putting them on a more even playing field than most other dacs would be. However, the may features "full" isolation, whereas the dave and qutest it is power lines only.
Once the schiit modius a friend is sending me arrives, i'll test using that to get a better picture of how good the mscaler is as a DDC. (Didn't want to use ADI-2 as it has a good PLL that would make differences less apparent than on a dac without a PLL.)
The difference on the may between USB, USB to SMS200, and the MScaler (no upsampling) was so subtle that I'm honestly not confident enough to draw any definitive conclusions. The combination of the may's excellent USB implementation with FULL isolation, and the PLL system is clearly ensuring that regardless of what input you use, the result will be great.
On the Qutest, USB to the SMS200 Ultra vs PC was noticeable. There was a slight improvement to both staging and imaging when using the SMS200 Ultra.
I could hear no definitive difference between the SMS200 Ultra and the MScaler, so clearly there is a decent USB implementation, and the MScaler is doing a good job as a DDC too.
The same went for the dave. USB sources did show a difference, but USB vs MScaler was identical.
It stands to reason that without upsampling, the USB implementation in the MScaler and the DAVE are going to be similar, so I guess it makes sense that there would be no difference.
As mentioned, once I have a less "fully featured" dac in terms of the quality of the digital inputs i'll test again. But its clear that the MScaler is good as a DDC alone.
---- Difference MScaler makes on a chord vs non-chord dac ----
This one is a bit tricky to test, as I would inherently be using a different DAC and therefore it is NOT a fair, direct comparison. Testing aspects like improvements in staging or timbre is difficult when they're starting from different points to begin with.
The other option I had was to simply use a chord dac, and swap between 192khz (the max SPDIF input rate for the vast majority of dacs) and the full 768khz, to compare the difference.
The simple answer here was that going to 192khz gets you about 80% of the 'full' improvement of the MScaler. There are no 'changes to the changes' if that makes sense. You simply get a little bit more of all the same improvements. There is definitely a difference, but definitely not as much as the initial 192khz upsampling. And the improvement at 192khz is more than enough that I'd say the MScaler is worth getting for non-chord dacs.
I did of course test the MScaler with the Holo May, and there was a big difference vs running it stock OS or NOS. I should mention though that this isn't a particularly fair comparison. NOS (with or without analog reconstruction) is going to be quite different from OS, and for a NOS capable R2R dac, I personally treat it as simply different, not better/worse. The holo may's NOS also sounds different to most other NOS dacs due to how it handles analog reconstruction. I'll update once i've tried the MScaler with a different delta-sigma dac however.
I could not decide on if I preferred running it NOS or with MScaler. It was track dependant. And i've found the same when using HQPlayer.
I have to say though, it is a shame that chord dacs do not have I2S input, and the MScaler at its pricepoint does not include either I2S output, or USB output, so that the sample rate limit for non-chord dacs wouldn't be an issue. Having only SPDIF outputs on a product of this price is a shame. I'm sure plenty of people would pay an extra £500 or so for an I2S or USB output capable version to take advantage of the full sample rate on a dac of their choosing.
---- HQPlayer vs MScaler ----
This I tested first on a chord dac. I wanted to give the MScaler the absolute best possible chance, with both MScaler and HQPlayer running to the DAVE.
I set HQPlayer to output to the USB of the DAVE, and the MScaler running into the Dual-BNC (fed by roon @ native sample rate).
For the filter, this is an immediate advantage to HQPlayer, the ability to choose. HQPlayer has a variety of filters, whereas the MScaler you have only one. And in fact HQPlayer offers filters with even higher tap counts, Sinc-L has up to 2 million taps, and there is even a 16 million tap closed-form filter.
After playing about a bit though, I went with the Sinc-M million-tap filter, as it was closest to the sound of the MScaler filter. In fact, it was very close....REALLY close....so close I'd be fairly happy to call them identical. Its almost like that "M" in the filter name was hinting at something.........
Its worth noting though that there is still an inherent benefit to using the mscaler in that as mentioned, it is a fantastic DDC.
So MScaler vs HQP over direct USB may be a different story to using an HQP network streamer like the SMS200 Ultra or a pi2aes. The MScaler will likely be better in most cases than connecting your DAC directly.
EDIT: Just wanted to update this post as after some further testing, I no longer feel that the above statement is true. (Am leaving it up for transparency, but please do read this edit).
At 192khz, which is what I was initially comparing to (because of the DAVE's delay when switching to dual-BNC making AB'ig difficult), sinc-M sounded near identical to the MScaler. However, since then i've been playing about at 768khz, and also to put another dac into the mix, testing 192khz from MScaler vs 768khz from HQP. And at this point things do NOT sound identical. In fact sinc-L then sounds closer.
As to which filter sounded 'best' it depended on the music, but I almost always found myself picking HQP over the MScaler.
Jussi Laako described the filters as follows:
This somewhat explains what I heard.
I'm guessing that the fixed 1m taps at 192khz was causing the faster cutoff and therefore making it sound closer to the MScaler, whereas sinc-L is a faster rolloff filter by design, so at higher rates it will sound more similar rather than sinc-M.
Also interesting to note how far the filters go in terms of bit-accuracy to the ideal sinc function. WTA filter in the MScaler by comparison is accurate to 18 bits according to RW in an interview with 'passion for sound'.
This was perhaps the most interesting result of everything I tried throughout this 'exploration'. As HQPlayer is something that you can use with ANY dac and with no 192khz limit. (And at <10% the cost of an MScaler)
And in fact, on the holo may, you can even go to 1.536mhz. Twice the sample rate that the MScaler outputs to a chord dac on dual BNC.
HQPlayer also has the interesting advantage of being able to do full delta sigma modulation in software, bypassing the DS modulator in your dac entirely if it is capable of handling DSD/PDM natively. Which can bring some impressive results. (not so much on high end FPGA dacs like the DAVE, especially if you're paying specifically for a "house sound", more something that can bring improvements to midrange and off-the-shelf stuff).
It is worth saying however, that barring all questions of quality, with the MScaler you're also buying convenience, and a great DDC. Which are both things worth paying for.
HQPlayer's native interface is not great, though roon has integrated support for it that allows you to effectively use it as an output device, and use only the roon UI.
---- Holo May vs Dave ----
So then, the battle of the titans! I've spent some time with some other high end dacs. And with delta sigma, my favourite dac is the dCS vivaldi, followed not too far behind by the Bricasti M1SE.
That changed today. And I can say that I prefer the Dave + Mscaler combo to the M1SE. Though the dCS i'd need to get a chance to A/B properly with it as its been a bit too long to make a conclusion from memory. (Dave alone however I'm not sure, and I think I might pick M1SE over a dave without mscaler).
I have not yet had a chance to try the Denafrips Terminator or MSB. But given as just about all comparative reviews between Terminator and May choose the may, and the fact that MSB is "remortgage your house" money, i'm in no immediate rush to try either, though would love to when the opportunity arises.
First, given the price difference, I thought it'd be fair to give the may a shot against Dave without the mscaler.
That didn't last long. The first three tracks I tried, May bested the Dave in several areas in a not particularly subtle manner. Resolution, speed and impact of the lowend, timbre of vocals and instruments, and most of all staging.
So, I gave the dave its clothing back and put it on the mscaler.
It was a lot closer now.....but not close enough.
The dave had a slightly warmer presentation, but much of this was due a slightly looser lowend. The May was still faster, with timbre that was addictingly real and convincing, and a spatial presentation so clear and present that it wasn't anymore a question of distance, space or imaging. You can literally HEAR how the violinist has their instrument oriented. Its a bit hard to put into words how good the sense of "presence" the may gives is. New Record Day on youtube put it quite well in his recent review.
Resolution was a close call at this point, and it was a bit tough to judge given how much more "convincing" the may sounded, leading me to lean towards picking it, but it really is close in raw resolution, and the differences in presentation elsewhere are going to be a much more important factor.
With the mscaler on both at 192khz though....it really wasn't all that much of a competition anymore.
Both of these are absolutely world-class dacs, there is no doubt about that, and either of them are going to be able to do a true summit-fi system absolute justice. But the may really is something quite incredible. And given as it costs half of what the DAVE does, its hard to find a reason not to go for it instead other than space (May is THICC)
I'm going to be doing a full video review of each of these products, which you'll be able to see on my youtube channel (
https://www.youtube.com/channel/UCJ0oW8D5z_IiFc7w46JJEuA ), as well as a written version here too.
But its been good fun trying all these and comparing all sorts of fun combinations.
If you have any specific questions about, or things you'd want me to test/compare between these products, let me know and i'll do my best to answer.
![]eep](https://cdn.head-fi.org/avatars/m/139/139747.jpg?1273084439){kind=avatar}
wink
)