That’s interesting and somewhat puzzling that the PIC code could produce the polar opposite in a different system.
Your listening to 3.0.5, because like me you have Uptone pls 1.2 that reduces enough noise to make 3.0.5 work with the matrix. best investment yet, more so than the matrix on its own
3 Likes
I went from 3.0.0 and 3.0.4 to 3.0.6 a few days ago and either I am crazy (always a possibility) or I really do hear noticeable improvements. In my system, 3.0.6 has better bass performance and better overall dispersement/spaciousness than previous versions. It reminds me of what I missed about Red Cloud’s deep bass performance which I thought was better than earlier versions of Snowmass; although Snowmass always had a cleaner bass. With 3.0.6 upper high frequencies are a little smoother but with all the detail. Nothing was lost with this version but things were improved.
Snowmass 3.0.6 is the best yet for me. I’m okay now if no future firmware updates ever come out but I imagine more magic will follow.
5 Likes
Totally agree !
It is most interesting that such sonic improvements in 3.0.6 arise from a PIC code fix for something Linux-based - with no changes in the FPGA code.
The algorithm or process Ted and cohorts use to explore the parameter space to find the best sounding FPGA code variant and hook with some variant of PIC code for best sound must be efficient — and perhaps does not cover the entire parameter space for that reason?
The Linux-fix PIC code coupled into 3.0.6 makes one wonder where in the parameter space the next sonic improvement of similar magnitude hides…
Certainly Ted’s efforts from the understood FPGA programming side make marked steps forward (as from Redcloud to Snowmass), but the seemingly more random/less understood changes
(such as bring the improvements in 3.0.6 beyond 3.0.0 or 3.0.5) are worth “systematically” exploring further if some method could be found to do so, and again it probably has to be “efficiently” in some sense.
3.0.0 just seems so dark to my ears.
1 Like
3.0.6 is the way to go
1 Like
I didn’t mind that, I find the treble the problem. That nice analogy effect you get without the lps in 3.0.0 sounds too much and too overdone with the lps making the treble sound a bit too syrupy. 3.0.5 has a ‘normal’ treble with lps.
1 Like
I’ll try 3.0.6 when my cold has gone
In my system, I have more of the magic with 3.0.0 vs. 3.0.6. The bass is more controlled, the central image is more well defined and yet the sound stage is wider.
I did enjoy the slightly increased ‘energy’ of 06 but I tired of that fairly quickly. I’ve gone back and forth, giving each 4-6 hours of listening before coming to my conclusion.
I should try Redcloud again for a while…nah.
Edit: I should note, I didn’t like 3.05 and find 3.06 much more similar to 3.0.0 than any of the other interim releases. I’m grateful each and every time that I can load a ‘new’ DAC, fantastic.
1 Like
I think it is natural that different people will prefer different versions (presuming that they do actually sound different and we aren’t imagining it). I prefer 3.06 myself - the bass just seems more “right” to me, but if I hadn’t heard it I would be perfectly happy with 3.05!
I think that all the time. My system never sounds as good, to me, when I have people over, unless everyone is in the groove. If I think, even in the slightest way, people aren’t enjoying the music or my system, the sound quality is significantly diminished for me.
4 Likes
I only made the one switch from 3.05 to 3.06 without back & forth (due to not caring so much for small differences).
I”m always doubtful regarding such power off/on comparisons with the same unit as too much influences involved…I also know that the first impression is not always the correct one, but anyway in such cases I rely on it due to laziness.
I played a 2L and a Chesky recording with lots of cymbal sound to check treble differences as well as an old but good Ellington Big Band recording and 2-3 more great recordings with soundstage/ambience.
What I heard with the AMT tweeter was at the very first after power on and loading, a more stable/firm impression of imaging/ambience … less spherical. Didn’t know for sure if I liked this better … also ambience seemed a little shrunk (which turned around after the DS settled again after some time…this makes it so difficult to compare after a power down). Then I mainly recognized a more open top end (very welcome), a somehow little more articulated presence (nice), also I had the impression of a little deeper extended bass. I heard natural sounding bass recordings, no killer bass tracks.
Mainly it turned out as main improvements that the top end really got some more open and ambience/aura between instruments also seemed to have improved compared to 3.05 after some time listening…maybe it also was other environmental circumstances, but I finished this nice listening session with the following track, which never sounded so good and enveloping to me as today for whatever reason. I keep 3.06 and don’t bother about the versions anymore. A switch back might have shown what of all this was illusion 
1 Like
According to PS Audio, 3.0.6 isn’t supposed to sound different, just a coding change to work correctly with Linux. Hence, as someone from PSA stated earlier, it didn’t get near the listening attention that a major firmware edition would have. As with the difference many users found between 3.0 and 3.0.5, there appears to be a disconnect.
In my system, which is a modified Oppo to DS (with iFi spdif purifier) to ATC active floorstanders, there remains three days in a qualitative difference with 3.0, for better and worse. I wrote before about the clarity and delicacy it introduces, instruments and notes being much more distinct in a attention catching way. At the same time, tonal balance (voicing) is not the same as previous versions, to the brighter, forward, “live” side. I wonder if the latter isn’t throwing some people off. In any case, I can even hear it on my cable HDTV, where male announcers with resonant voices now sound a step up, e.g., baritone to tenor.
I have a little trouble with the word “voicing”, we are just trying to get the hardware out of the way of the music. The differences in various versions of Snowmass are purely do to changes in hash generated by the rest of the system probably via the analog outs, but also possibly back thru the digital interconnects and/or power cord. There is no change in the level of any of the frequencies and no changes in the phase of any of the material you are listening to. Changes in the spectrum of the noise in the system can both mask details in some frequency bands and (like dither) increase detail in others. Any changes in ultrasonic noise has the possibility of affecting systems in different weird ways more than the changes in the audio band.
Just a gentle reminder, if you are having any problems, upgrade to the latest version of Snowmass and see if they are still there, but otherwise have fun.
9 Likes
3.06 is better yet IMO! It surpassed my DAVE 4 releases ago IMO. DAVE is on my TV now. It is a good TV though. It just keeps getting better. MR. Smith, did you put anything from the TSS in this one? As long as he is willing to keep going it will get better and better. Until I guess he hits a brick wall with memory but he is very good at small code. He already got a heck of a lot out of it. This is the first truly “almost” non-obsolete hardware! I wish phones were this way. 3 updates then buy a new one.
3.06 is the same FPGA code as 3.00. Just the PIC code has changed and that is something PS Audio, not Ted, does.
2 Likes
Indeed the basic changes between Redcloud and Snowmass are a part of the work that was needed to use the newer Xilinx tools that I’m using for the TSS. And yes I’ll be backporting any new enhancements that the DS and DS Jr hardware support in future releases. (And seegs108 is definitely correct, all versions of Snowmass have the identical FPGA code, as I mentioned above any changes you hear are caused by noise, not by algorithmic changes.)
8 Likes
I haven’t gone through to make a count, but whatever its cause(s) three of us with presumably disparate systems noted to one degree or another straight off that we heard more energy from the upper midrange and less from the lower midrange/upper bass, and that tilted the overall sense - I called it voicing - in a way that is different from earlier versions and undermines the appeal of the greater top to bottom clarity of this version. And since I heard the same thing that many others did between 3.0 and 3.0.5, I don’t imagine that I (or we) are alone on this.
It seems like those who preferred 3.0.0 still do, and those who preferred 3.0.5 often now prefer 3.0.6.
3 Likes