There are a whole lot of bad routes and as soon as a particular possible new route exceeds any constraint the routing algo goes to the next. In a simple example case, as routes get longer they get slower and soon signal level changes wouldn’t get to the destination in time. So tho there are many more resources in a bigger FPGA you can only reach about the same number from any given source. The longer ones aren’t even considered (this is a gross simplification, but you get the idea.) Similarly at the placement level: if, say, a higher precision multiply might want multiple DSP blocks, if those blocks are right next to each other there are all manner of direct block to block connections. If the blocks aren’t right next to each other the routing needs to use generalized resources outside of the DSP blocks for connections. That adds time and congestion, so the tools can rule out having those blocks too far apart.
As far as generated noise is concerned that’s due to how many signals switch at the same time near a particular power pin. As the chips get bigger they have more power pins… But as tech moves on they also generate less noise per signal transition.
So in general the fewer resources used per pin, etc. the better the sound so bigger chips have a good chance of sounding better. Also simpler routing has a good chance of sounding better. I really try to keep congestion down since it adds a lot more kinks in the path from here to there.
I’ve talked to them. They like the problem and it’s actually closely related to some power saving features they are working on (perhaps did work on now.) Getting rid of signal changes that won’t be paid attention to later saves both power and noise. They have more hardware features to support suppression of unneeded transitions in newer FPGAs and the newer tools support more algorithms that look for opportunities to save power.
I guess you’d have a lot of options to make your knowledge available to other industries where it has more financial impact than in audio, that’s why I think it’s luck for us you’re a music guy, too.
What are other fields of application for FPGA’s? Any where microelectronics are in the game? Does any other field of application react as sensible as audio?
There’s a continuum of customizable logic from just a few gates up to chips with stupendous amounts of resources (e.g. a single chip with 21 Tera Multiply and adds / second, 128 transceivers that can run at 56Gb/s, 8Gbytes and 460Gbytes / second bandwidth to it… integrated 100G Ethernet…) You can program them in C++/MATLAB/Simulink… You can get multiple ARM cores built in…
When are they the most cost effective solution? If your only building a few of something, if you keep changing the things you need to do, if you need ridiculous connectivity, amounts of DSP, …
I suspect that using them for audio is pretty rare, we probably know most of the companies that do that to one degree or another.
I do not own a bridge so I quickly switched back to v1 out of curiosity. To be honest, I was hoping for a difference but my ears told me another story.
Thanks Ted for your patience and the explanations.
There is no logic in audiophiles’ minds as we are not just serious listeners, but testers too. =\
Interesting comment Ted, do you have any idea what it is that is triggering your tinnitus. I too have tinnitus and notice changes but haven’t identified a pattern to what is causing it. Of course, music is just one thing that plays into that cause/effect scenario.
I think maybe I’m going to start any comments hereon with: “It could be my imagination, but”, I just backdated my DSD to Snowmass V1 and I do think it sounds better than V2.
. . . And the remaining 21% voted that they think Snowmass v1 and v2 sound the same (i.e., no one has yet voted that he prefers Snowmass v2 to v1).
Of course, the polling is taking place in a thread entitled, “Snowmass V1 sounds better than Snowmass V2,” so the thread may not be getting traffic of random unbiased potential responders!
You should have yourself checked by your physician. It could be any number of things. I had mine checked several years ago and found to be unknown. A few years ago when I found I am fibromyalgiac I discovered it was but one of many symptoms of this genetic syndrome. So best you go and undergo tests.
No, I’ve had my tinnitus essentially every hour of every day for so many decades I don’t remember not having it. On the other hand it doesn’t bother me much at all, except when I want to listen to details in quest settings. I’m glad it doesn’t bother me like, for example, my ex-wife’s father: it was a continual distracting painful annoyance that seemed to be the bane of his existence. He was pretty sure his was from the loud noises of battle in a submarine.
Mine gets worse with stress (and I can use it to measure my stress level). It gets worse if I’m on pain relivers very long. I can make it go away for a day or so by listening to loud music for an album or two, but that’s clearly not a real solution in that that would be pretty bad for my hearing. There are a lot of possible causes of tinnitus and a few things that help some. For some it’s the thinking about tinnitus that makes it worse and some sort of distraction, perhaps white noise, etc. can relieve the symptoms. For me it’s not a problem except that it masks details that sometimes I would like to hear so those kind of treatments aren’t very useful.
more interesting stuff… Yes - the psycho-acoustic definitely comes into play … but I never switched because there was no real need and things just sounded splendid on V1. I also am not voting in your poll since I didn’t even compare V2… Maybe there is something to it… but I couldn’t really say… I don’t know the rhyme or reason why we have some of the appearance of inconsistency either in loading … or sound. I understand systems differ - and what might sound good on one may not on another… But, it seems that the loader of the software seems to create different results… Although I have never experienced this - only see people post …‘I reloaded and everything sounded great’…
Added Edit: I get that people are another variable here… I am pretty anal about this - new sd card every time… not pushing software eject makes me want to take a shower… this sort of thing… and the general population as a whole may not be like that. My system is as simple as it gets as well - no bridge… But - if we accept that its not ‘the people’ - and we accept that people here differences when they re-load - well that begs the question - Whats up with the loader… Ted & Co. may all ready know some of the niggles… but I can only blab…
I had it checked about a year ago by an ENT specialist. They did an MRI and found nothing. I’m sure it’s from noise exposure and there is no cure … yet
I was misdiagnosed with tinnitus for years. It turned out to be palatal tremor (aka palatal myoclonus), which is a rapid spasming of the palatal (roof of the mouth) muscles, which results in clicking or popping in the ears. There is a treatment, but it is not fun: Botox injections into the roof of the mouth to temporarily paralyze the palatal muscles, which greatly reduces the clicking/popping. The injections last about three to six months. Eventually, I decided the treatment is worse than the disease, and now I simply live with it.
You may have misunderstood my question, I was referring to your comment specifically about changes in the compiles of the FPGA having different impacts on the tinnitus. I’m thinking high frequency noise or some such thing is triggering more “ringing”.
Sorry. It’s more that, for me, jitter adds stress and when my tinnitus is on the edge changes in stress cause disproportionate changes in the tinnitus. When we were trying a lot of CD treatments (green pen on the edges, edge bevellers, weight mats on the surface, that sort of thing.) at the local audio club I could hear a distinct and repeatable difference in tinnitus with almost every treatment. I decided to vote base purely on my tinnitus response and found I was voting almost always with the members of the club that really liked vinyl and the opposite of the club members that really liked CDs That experience convinced me that jitter was something to be taken seriously and I used my tinnitus to judge jitter levels for years. Lowering jitter was one of the goals of the DS development.
I’ve made sure that the release choices that PS Audio choose don’t make my tinnitus worse. During development if I find my tinnitus getting worse I try to understand what I might be doing wrong. But in general the changes in the DS code don’t have much of an affect on my tinnitus.
Right, i’ve finally got a version of Snowmass I am (almost) happy with - I’ve had to rake back my Martin Logan’s a few turns to reduce treble brightness.
There must be a problem somewhere as it’s taken about five attempts to get snowmass that sounds acceptable to me, I had gone back to Red Cloud.
Ted can you confirm that when the flash ram is flashed that a crc check is made as well as checksum AFTER programming? If so then there must be a timing type issue in some people’s units where the data consistently fails to copy properly to the fpga.
By the way I’m in the V1 camp. V2 I thought sounded a bit clinical. (Display always switched off)
That experience convinced me that jitter was something to be taken seriously and I used my tinnitus to judge jitter levels for years. Lowering jitter was one of the goals of the DS development.