RETRO-AUTOPSY: Museatex Melior BS-1 Bitstream S/PDIF DAC (circa 1995?). A few words about external clocks for audiophiles. And a speaker cable blind test?

The company Museatex was formed around 1990 by Ed Meitner and John Wright based in Calgary, Alberta, Canada. Over the years, the brand was assumed under A/D/S which eventually closed in 2001. There's still a web page for information on Museatex products here. Ed Meitner has moved on these days to EMM Labs since 1998 and given his work with SACD, his name has been associated with "bitstream" type technologies since (some more info/video).

By the end of the 1980's, 1-bit Philips "Bitstream" PDM DAC chips started to show up in the market with the Philips SAA 7321 chip as one of the 1st generation devices (see Stereophile reporting back in 1989). Some of these chips were incorporated in CD players like the Meridian 208 that came out in early 1990, well-reviewed by Paul Miller in High Fidelity.

By 1991, the "second generation" Philips SAA 7350 "bitstream" DAC became available, capable of taking up to 20-bit PCM (from 16-53kHz sample rate) and upsampling to 1-bit SDM applying 3rd order noise shaping (some more history here). It is this DAC chip that's at the heart of the Museatex Melior Bitstream DAC, I believe first released in 1992. Here are the specs from the company.

The exact history of this particular unit I had access to has been shrouded in the sands of time, but was probably purchased around 1995. The original suggested price on one of these was US$899 which would be about equivalent to over US$1500 in 2023. Even to this day, I've seen comments from audiophiles who have a special affinity to the sound of this vintage DAC. Claims to fame included anti-jitter input circuitry (supposedly reducing "glare and harshness"), a custom digital filter allowing for "more accurate signal conversion", and high quality power supply with line noise filtering.

Thanks again to my friend linnrd for digging this unit out from his audio device archives. While externally it looks to be in very good shape, unfortunately, it didn't quite power up normally for me which sadly resulted in this article being more of an "autopsy" rather than actual measurements. I guess having a look inside is still better than nothing for historical documentation if nothing else.

I've seen some pictures of this DAC online and some models look like they have natural colored wooden side panels. This one is painted a glossy black. As you can see in the picture above, there's a single "INPUT" button on the front and red LEDs light up when in use indicating samplerate (32/44.1/48kHz), a de-emphasis indicator, and whether S/PDIF Coaxial or TosLink input is being used.

The rear consists of a very basic layout (compared to some modern devices with USB, ethernet, Bluetooth antenna, etc.). In stock form, the power cable is not detachable, we have 2 S/PDIF inputs (TosLink and coax), and a pair of single-ended RCA outputs. This one was "Made in Canada". Some others out there are "Made in USA".

Here's what it looks like on the bottom. Again, nothing fancy, some rubber semicircle footers for non-slip vibration reduction I suppose.

Although it's obviously much larger as you can see in the image above compared to a modern Topping D10 Balanced or any USB dongle DAC (like the E1DA #9038D6K, as recently reviewed/measured), this is a nice looking, solid ~7lb box. There's what looks like a slab of glass at the top. The front has a matte finish plastic membrane very reminiscent of the old membrane keyboards of very early computers (like the Atari 400 or Timex Sinclair) with a single "INPUT" button to switch between the S/PDIF inputs.

I would bet if one did not know this was a DAC these days, some might wonder if it's a linear power supply or maybe an amp.

While I would have loved to measure this device, alas, although it turned on, audio output was sporadic and I was not able to confidently get usable results. After a few minutes, it looks like the power supply would give out.

As such, let's look inside and see what US$899 audiophile dollars would have bought back in the early '90s.

It came apart quite easily with just removal of a few screws. For orientation I indicated the front and back of the DAC. AC input enters from the rear and runs across the board to the front where the transformer/power supply section is located, leaving the DAC and output circuitry at the rear.

Here's a closer look at the transformer with markings for primary and secondary windings. These days we often see toroidal transformers for linear power supplies in "high end" audio devices, not so much back in the '90s.

As per the PCB silkscreen text, the board is designed in 1992 out in Calgary, Alberta. Museatex (Meitner) implemented the "C-Lock" circuitry back in those days to address poor jitter performance through the S/PDIF.

Another angle, we can see those 2200μF/35V Nichicon power supply filtering capacitors.

Here's the DAC portion. Note 2 pairs of pots presumably for fine tuning of the 2-channel outputs. We also see a pair of 8-pin AD845 op-amps for each channel. These op-amps have a reputation of running hot and some have recommended putting heatsinks on these. Modern DACs sometimes have socketed opamps for "rolling" which could (usually subtly) affect sound quality.

A closer look at the DAC chip and "C-Lock" circuitry. We've got a 20.000MHz crystal (no fancy OCXO part of course). There's the large square Philips SAA7350 DAC chip.

I would have loved to measure something like the J-Test on this to see how well that C-Lock worked.

Here's the large Yamaha YM3623B rectangular chip. This is the "digital interface receiver" (DIR) which decodes the S/PDIF digital stream, including detection and application of de-emphasis. Remember, these days it's better to use high quality software de-emphasis; unclear what the quality of this process would have been with old chips like this.

The Yamaha DIR is capable of 16-bits only although the Philips SAA7350 can accept up to 20-bit input. Here's an interesting DIY discussion about this Yamaha chip.

Historically, the YM3623 doesn't have a good reputation for jitter performance compared to the Crystal CS8412 in the early '90s. Perhaps this is even more reason why EMM needed to implement the C-Lock. I see that Stereophile measured the C-Lock performance and showed that it did improve temporal performance with the Meitner IDAT back in 1993.

Another important chip we see on the board is the Yamaha YM3434 just above the crystal. This performs the 8x oversampling using linear phase settings, with 18-bit precision. This part was released back in 1988.

As per the title of this post, this is just an "autopsy" unfortunately. As mentioned above, after a few minutes, the device seems to poop out going into some kind of shut down mode. I didn't see any bloated caps, electrolyte leak, burned resistors, nor blown fuses. I suspect the issue is with the power supply, possibly those old Nichicon electrolytic caps need replacing after ~30 years.

Perhaps in the days ahead, linnrd might want to "resurrect" this old DAC. If that happens, I'd certainly be happy to measure the performance of this "standard resolution" 16-bit, 44.1/48kHz DAC and especially have a look at that C-Lock jitter performance.

As a fun comparison, here's a look inside a modern <US$100 Topping D10s from 2021:

Not exactly the same feature set of course, plus this is USB-powered so no linear power supply. As a 24-bit hi-res DAC, the Topping D10s would perform significantly more accurately than the old Museatex at something like 1/15th the asking price when new.

--------------------

Over the years, I've measured older products like these '80s-early '90s CD players, this 1994 Laser Disc player, and even this old Sony PS1 as CD player also dating back to somewhere around 1994. To be honest, none of these old devices demonstrated significant jitter issues. I find it amazing how the audiophile press has obsessed over this word over the decades when it came to straight forward CD playback.

Articles about jitter back in the day from the likes of Stereophile like this one is simply referring to jitter over the digital S/PDIF interface. I don't think there was ever really a problem with single-box devices like CD players, so if you didn't use S/PDIF, don't worry! These days, asynchronous USB and ethernet transmission are typically excellent, jitter-free, with reputable products. Besides, even when present at moderate levels, I've discussed why I don't think jitter is generally audible.

Yeah, even though some people continued to raise all kinds of bother about the "importance" of jitter into the 21st Century, as if there are problems, I seriously hope that honest, educated audiophiles who have used their ears to listen will know by now that most of this had always been just hype and FUD.

Speaking of Meitner and jitter/temporal performance, I want to highlight this video from Doug Schneider of SoundStage! visiting EMM Labs:

I recommend jumping to the latter half of the video (~4:35) where Meitner / Schneider is making an important point about the best-practice of keeping the clocking mechanism as close to the D-to-A conversion circuitry as possible for optimal temporal performance.

Indeed "What is it with these external clocks?" Stuff like the US$10,000+ dCS Rossini Master Clock. Or maybe the equivalent-priced Auralic LEO GX.1. The MUTEC REF10 SE120 asking for €5.5k. Or Silent Angel Genesis GX Word Clock which appears to be a comparative bargain at just over US$3500. ;-)

Even though I do not doubt that the fancy clock must be performing very well inside those boxes, I do wonder what happens to the timing precision once you send it down cables into the DAC component and ultimately what effect it has on the D-A conversion itself! I would love to see an instance where something like the J-Test is actually improved by the outboard clock.

As much as this Stereophile subjective reviewer suggests that the sound improved with the Rossini Master Clock, notice that in the objective measurements section, there was actually no test to see if that master clock made any difference! How hard would it have been to disconnect the clock for a bit and show us if it was worth throwing another $10k into this system on top of the >US$25,000 for the Rossini Player!? Apparently too difficult to bother measuring. ;-)

Yes, there is a time and place for using a master clock in a digital audio system. And that place is in the studio. As per the linked Sound-On-Sound article (bold mine):

"There's a widespread notion that adding a high‑quality master clock to a digital system will somehow magically improve its overall performance. While that might possibly have been the case in the very early days of semi‑pro digital converters where, frankly, some of the internal digital clock designs were pretty ropey, it certainly isn't the case today. As I've explained above — and will prove below — today's converter designs generally work best on their own internal clocks, and most will deliver a slightly poorer performance when clocked externally. The very best devices will show no change in performance at all, because they have superb clock-extraction circuitry that can remove all traces of clock jitter and other external clocking artifacts, so they work just as well as when running on their internal clock."

After this statement the article goes and show how various ADC's actually may perform less well with an external clock (higher noise, more jitter). So it is with DACs running off an external clock where you now have another wire going between boxes and extra circuitry to implement this feature. Notice that the SOS article came out in 2010. These days with DACs featuring asynchronous interfaces, chips with very high jitter reduction, and fancy units featuring "femtoclocks", IMO, there's simply no market for external clocks other than in the studio. The only potential role these boxes have are as some kind of "look at me" luxury items to impress audiophile friends with a significant chance that it's actually reducing playback fidelity if the audiophile had bothered to objectively test!

Be careful about "reviews" that insist that DACs "sounds better" with these add-ons. With high-fidelity systems, very subtle changes might sound "different" but this doesn't mean it's necessarily higher performing (assuming the listener actually heard a difference in the first place - expensive placebo and all). When it comes to clocking, I can't imagine not using objective measurements to prove that temporal performance has improved! It really doesn't matter what some subjective guy believes he hears with the sound (like this or this), unless the J-Test or even a simple 1kHz THD+N FFT can show us benefits - literally the only thing of benefit these boxes might be able to do for us, again, assuming they don't make things worse!

I bet there's a good amount of margin for companies to profit from selling these external clocks. Here's a review article on the US$7,500 dCS Lina Master Clock for example. Notice how empty that box is inside! Most of the expense is probably in the fancy metal enclosure.

Rather silly IMO to have a master clock in a home audio system when all you're doing is playing music from your player/DAC. In a way, it makes an audiophile wonder why a company like dCS has such trouble getting timing correct when others are achieving technically perfect performance already without asking for another $10k! :-)

“First you make people believe they have a problem, and then you sell them the solution. That's how advertising works. Every snake oil salesman knows that.”
    --- Oliver Markus Malloy, Bad Choices Make Good Stories: Finding Happiness in Los Angeles

--------------------

One last thing. Check out this "blind test" of speaker cables:

Why is it that they're comparing a 16AWG speaker cable of unknown provenance with GR Research's effectively 8AWG cable (which is basically this Asian-made cable with some kind of 1/2" cotton rope in the middle)? I suspect the GR Research formula is something like: Fatter Cables (with cotton rope inside) = Higher-End = Better Sound. IMO, the technical hand-waving is just a marketing distraction, justification to ask US$290 for 21' of cable, enough to make 2 x 2.5m/8.2' lengths before cost of connectors or construction time (vs. maybe US$100 for the Asian cable fully constructed with banana plugs).

For this test, couldn't they just spend a few bucks and terminated a new, decent OFC 12AWG zip cord with some copper banana plugs and see if Danny could still achieve 8/10? How about using Jay's AliExpress cable (0:45) that he thought sounded "broken" - should not be hard, right? (Maybe double check with a multimeter that it's not literally broken!)

It's important to let viewers know what the "materials" are used in tests like these if others want to replicate the results. The opaqueness of this test, especially the fact that this 16AWG cable was not described and viewers cannot actually go buy one of these for their own listening is problematic. I suspect the measured LCR would be quite different between the cables, especially R would have an impact on a direct A/B switching test!

Anyhow, I don't think there's anything all that mind-blowing here since I don't know of anyone recommending 16AWG cables for very high performance audio, nor claiming that this is supposed to sound exactly like an 8AWG snake! Seems to me like they're trying to disprove their own straw man. IMO, simple and inexpensive 12AWG OFC speaker cables with good connectors are more than adequate for reasonable lengths (say less than 30') at home with most systems except for maybe the most current-hungry speakers.

Okay everyone, off again to get some Spring Break R&R. I see Pink Floyd's Dark Side of the Moon is getting exposure again this week with the Live In Wembley and 50th Anniversary Deluxe editions, the latter with an Atmos mix on BluRay. Hopefully dynamic range compression is not excessive in the remastering. I think I have enough versions of this old album (LP, CD, 2003 30th Anniversary hybrid SACD with multichannel, and 2010 Immersion) that I'll give this a pass; maybe just a listen to the Atmos mix at some point. Surely, there aren't many more ways to get consumers to buy yet another version of this 1973 album after this, is there!? After half a century, doubtful there's much sound quality to extract even if they went back to the master tapes!

Remember, get me your 24-bit vs. 16-bit Blind Test results. Hope you're enjoying the music.