Either in my configuration the DS has no noise, or, as also a vinyl listener, I’m out of competition because used to a much higher noise level
In many systems the DS is silent, one cannot hear its quiescent noise floor. In others, one can hear a low level hiss.
Yea mine either - it’s dead quiet. I did have the Brooklyn for over a year and while it was a big improvement over anything I used before I found that my ears felt like they were burning after a while. The DS has made it possible to enjoy the majority of my Redbook collection for the first time. If there’s a small amount of perceived noise for some what does it matter?
You mentioned the EMM DAC2X, tho I don’t have the DAC2X I have the EMM DAC6e whose close brother the DAC8 was used for mastering most of the early SACDs. It’s noisier in my system than the DS (for multiple DS releases now.) As Elk mentioned some noise is inherent in any SDM approach and more in a single bit SDM approach. If the noise is white it doesn’t get in the way of details, if it’s colored it does much more readily obscure details. The best way of lowering the noise from the DS is to have the DS’s volume up pretty high, e.g. 80-100 when listening to the loudest material you’d normally listen to. At that level, if the system is fairly well level matched to that (which often required a preamp, but not always) there shouldn’t be noticeable noise at the listening position, of if there is any when the system is idle it’s gone when any music is playing. That’s why the noise that’s there isn’t a problem when mastering, they are very adept at optimizing S/N for each step of the process.
Still lowering noise matters and every time I lower noise in the FPGA code we all hear the difference, even tho those changes are not measurable by myself, John Atkinson, etc.
The TSS lowers noise in as many ways as I can think of. Here are some, but not all: By doubling the number of digital switches the noise is lowered by 3dB, the TSS uses more expensive digital switches which have a lower noise floor, the power supplies for the analog (and the whole unit in fact) are much quieter over a broader frequency range than those in the DS, the isolation of the digital from the analog cuts down on both conducted noise and RFI from the digital processes, but it also keeps the ugly step brothers away from the analog: the display, things like USB and Ethernet… There are also some slight architectural changes in the analog section that should allow more noise tweaking in the digital signal processing. Since I don’t have a running system I don’t know how much I’ll be able to get out of that in the first release, but the either in the first or second software release I’ll be able to get some noise out that wasn’t possible to squeeze out in the DS or DS Jr.
Ethernet isn’t a panacea, the higher the frequency of any digital process the more noise but no one want’s to use 10MB Ethernet these days. Ethernet is great for mastering houses, studios, etc. But it’s not yet ready for general use in the consumer audio world: there’s no standard for audio transmission protocols so you need devices on both ends of the ethernet link that talk the same language. In a pro environment you can have all parties using the same hardware (at some level) and side step the standardization issues. Right now with bridge like devices the standard is using a computer on each end to transfer files (or some other computer to computer speak.) This isn’t the best for audio: having a computer at each end of the wire… Note also that Ethernet by design has to pass noise that exists in the same frequency range as the signal. There’s common mode and differential mode noise rejection in the Ethernet connectors as well as down stream, but it’s not the case that Ethernet is fundamentally better isolated than other interfaces.
I have trouble with the term voicing: In the DS, DS Jr and TSS I simply strive to do the least damage to the audio. I don’t pay any attention to tailoring the sound, I just work towards the technically best processing (both in hardware and software) that I can do and I know that that does the least damage to the audio. The only way to make a DAC work well in all systems is to use the interfaces that are there as faithfully as possible. As a not so obvious example, in each release of the software I attempt to lower the ultrasonic noise as much as possible: since not all audio related hardware is designed with ultrasonic noise in mind some of it will react badly to too much ultrasonic noise. Whether a given preamp will negatively affect the audio because of the ultrasonic noise more than a given amp isn’t obvious. But in general the wider bandwidth the preamp, the amps, the speakers handle linearly the better: I like my speakers that are +/- 3dB to 50k, the amps to 200k, the preamp to 100kHz. Just 20k at each of these steps is not near good enough for audio unless the input is filtered to 20k…
You mention a compromise in harmonic integrity: there’s no such compromise in the DS or the DS Jr, pre amp or no preamp. Or at least there’s no compromise bigger than any other audio component that doesn’t have infinite linear bandwidth. The real issues with needing a preamp are filtering ultrasonic noise that your amps or speakers react badly too, allowing shorter interconnects if the interconnect capacitance is noticeably higher than “standard” consumer audio interconnects, and most importantly helping with signal to noise issues if your amps (and speakers) aren’t designed to give you the sound levels you want with an input sensitivity that matches the DAC’s output voltage levels. So (ignoring bandwidth issues) if your amps deliver exactly the power that your speakers need to be as loud as you need when driven by the voltage that the DAC delivers you won’t need a preamp and there will be no compromises.
As has been mentioned I’ve mentioned my home system on several threads
I have JMLab (Focal) Nova Utopia Be speakers with the inverted beryllium domes, silver speaker wire (and until recently silver interconnects), sand amps, but still there’s no exaggerated hiss, etc. I’ve had many complements that my system sounds like Maggies but with better base, or it sounds like it has tubes, or “is that sub on?”, “where is the center speaker?”, etc. It’s not perfect by any means and these days my ears are further from perfect but I still double check with my LCD-3’s and with my wife’s or daughter’s ears.
Wow. Great stuff, Ted.
My nerd side is tingling.
“…ears felt like they were burning…” : )
That was my issue with the Manhattan, which (just prior to buying the DSJ) I had hoped could be a centerpiece of both studio and stereo, as I still had multitrack sessions on archaic FW drives and so forth at that point. The Mytek just hurt after a while ; )
There have been points over the past couple of years where I’ve used the DSJ direct to my amp, and I could hear it if I went up to the speaker. But I don’t do that ordinarily, so I didn’t care. The bigger issue was that it didn’t sound particularly “musical”. With a quiet preamp in front of it, I don’t hear it. I’m now listening with a pre with dying tubes that are making noise which is audible from the listening position when music ends (if you listen for it). Still sounds more (koff, koff) like music than without the pre.
During the noise shaping process, is it possible that some of the ultrasonic noise can aliase into the audible audio band?
No, the analog output filter takes care of that. At a 11.2896MHz sample rate the analog filter has a lot of room to filter out any aliasing, almost 7 octaves so it doesn’t take a very steep filter to get down to -120dBFS by 5.6448MHz (a 18dB/octave filter is sufficient.) As a check the noise from the SDM process on the DS and DS Jr measure the same as a C++ program doing a SDM with the a similar (but digital) output filter.
@watchdog507 You’re absolutely correct. I misstated which Mytek DAC I was using.
Please forgive my repeated mistake: the DAC I compared to the DS was the latest model Mytek Manhattan II (with their network card), NOT the Brooklyn. I have never had close experience with the Mytek Brooklyn, but obviously had it on the brain (I stayed in Brooklyn in November?) …I’m blaming holiday drinking.
Apologies all, please re-calibrate my erroneous post with this correction.
I doubt they sound as good as Tannoy Westminsters. The Verity’s belong in a Kardashian house.
I don’t know anything about either…as to what belongs in a Kardashian house, apart from Kardashians, I know even less.
I’m late to the party on the recent TSS happenings, have we seen any photos of the casework? Now there is mention of a VU meter in the analog box. There has been previous mention of a capacitive touchscreen in one of the boxes…is the VU just a computer generated graphic like in the P20, or a real VU like in a D’agostino amp?
And you’ve actually heard the Monsalvat or any Verity loudspeaker?
Mike and Paul on AS both owned them and both have moved on to the AG Duo Mezzo HD so not an end all speaker while very good. The Monsalvat is a complete multi-cabinet speaker system including a double stack of powered subs so not in the same ballpark. The Westminsters would compete with the Amadis S or maybe the Sarastro IIS. The Monsalvat or the big Kharma would require almost a ballroom sized space.
The VU meter will be on the digital box display. It should be more realistic since the display has finer detail and the display controller is faster. It will have the correct ballistics: “0 VU is defined to be a level of +4 dBu for an applied sine wave. The VU meter has relatively slow response. It is driven from a full-wave averaging circuit defined to reach 99% full-scale deflection in 300 ms and overshoot not less than 1% and not more than 1.5%. Since a VU meter is optimized for perceived loudness it is not a good indicator of peak performance.” (http://www.aes.org/par/v/#VU_meter) Tho the display isn’t completely laid out I suspect that we’ll allow PPM (http://www.aes.org/par/p/#PPM) as well. There’s plenty of bandwidth to pass a decimated stream of each channel from the FPGA to the display. (At least on my laptop the AES links show material just below the real targets, scroll up to see the entry if you don’t see it when you chase the link.)
Cool if the ballistics/levels can be accurate and realtime. Would need to be REAL purty, though.
Something on the order of how the best pro audio plugins look on a computer.
Why do a VU meter? Unnecessary eye candy that has no real function other than to impress people that don’t know better.
If you don’t want VU, select some other display. Personally I’d choose a peak level meter with peak hold something like:
or simulated LEDs: