I haven’t really “tested” that - but that is my subjective experience. Any time I have put it in, I want to take it out, even if it is a better level. Wierd.
Edit: Sorry, hadn’t read the above response, Ted…
Can you expand on what that means, Ted? Sorry if you have before. I suppose the only reason I ask is that you have said the volume control is (nearly) lossless due to the surfeit of bits. Does this factor not matter particularly?
When you upgrade from one release to another two things happen that might cause an effect like burn-in.
Just turning the DS off and then back on causes a small temporary change in sound until the DS reaches steady state again.
When doing a software update the transformers have a small amount of DC on them. It can’t hurt the transformers, but it does take a little time for the effects to wear off when you start playing music. Some don’t hear any difference after doing an update and it’s extremely difficult to A/B something when it takes minutes to change one way and hours to go the other way. Still some like the sound better a day or two after an update than right after it.
Indeed each release MAY use FPGA resources that haven’t been used (or used much) in previous releases. But in those cases 1) every system will be different depending on how long it was running each previous release, 2) the length of time for a new section of the FPGA to become like a used section of that FPGA isn’t well understood, nor do we have accurate empirical data. In any case Snowmass will use fewer resources of most types than previous releases and I’m guessing that it probably won’t show this effect as much as some previous updates.
All of these are small compared to the differences that hundreds of hours bring to a new DS.
6 Likes
Thanks. Great insight.
1 Like
Sorry guys, we got really close today then found a last minute bug - but it’s one you’ll be happy we found. I gotta go play grandpa this evening in Denver but will hopefully release to the beta testers tomorrow.
9 Likes
Perhaps it’s not clear but since the volume setting in the UI is in dB and the actual volume change is done with a multiply, there has to be some math to convert the UI value to a multiplier. The multiplier is something like 2^20*10^((100-{vol})/2/20). But the first number (2^20) is really something slightly different: the actual number was empirically derived to set the output volume to match the spec. There were other multiplies that would have been in the signal path for various reasons so folding them all into one multiply saves resources. (One of them is really weird, there’s a previous post with it somewhere.)
The only reason I mentioned specific numbers in my last post is that many people are used to volume control multiplies being skipped when the volume set at unity gain or multiplying by one when set at unity gain. Then when some other volume setting is used the bits are actually changed (and in the old days with only 16, 18 or 20 bits that change might be noticeable, especially if multiple precision wasn’t used and then dithered.)
None of these matter in the DS, since all bits are kept after the multiply things like bits falling off the bottom or noise introduced with dither aren’t an issue.
In the DS multiplying by one gain vs another gain doesn’t save any processing (nothing is skipped) doesn’t affect the final precision (all bits are kept) and doesn’t save electrically generated noise with lots of 0’s or 1’s because for a unity output (a volume of 100) the actual multiplier is something really weird, not something a 1 with lots of zeros.
1 Like
I found the other post with the weird numbers:
3 Likes
Thanks, Ted!
To avoid clutter, I don’t use the remote of my DS. I just use Roon running on a tablet. So does my wife.
May I request a “display off” selection from the LCD touch panel for the next firmware release?
Is there any sound quality disadvantage running volume to the DS Sr or Jr through Roon v the DAC (slider, knob or remote).
I am under the impression better to avoid running vol through Roon to get best sound… 
Roon up-samples to 64 bits and applies dithering when you engage the software volume option. It may or may not make a discernible difference to the sound. It will require more processing power so depending on the capabilities of your Core it might cause some change to the sound. Personally if I had to make a choice, I’d try both and see if you can hear any degradation using either of them at higher levels of attenuation.
I think the level of up-sampling in the DAC is much higher so I’d assume that using that option would cause less degradation. Ted has explained at length how the volume control works in the DD and DSJ. It is all way above my understanding. The only advantage that the software volume control offers me in my setup is the ability to use it like a fast mute function if I need to drop the volume of the music suddenly.
1 Like
Have fun with your grandchildren 
1 Like
I’ve run 100 without attenuator for months. Last night I tried 99 then 95. 95 is going to be the set point for the foreseeable future. DS feeds BHK preamp, better synergy between them in my setup at 95 definitely.
2 Likes
Yes Brett. If one is so inclined to listen carefully with ones ears… changes in setup can bring much pleasure and costs nothing to try.
And yes…without attenuator in use.
Thanks for suggesting it!
He was going to release it if no one asked! 
2 Likes
That’s hilarious, @brett66. Reminds me of the famous story of Miles Davis performing at the Blackhawk in San Francisco. Someone in the audience shouted out a request – “play ‘Stella by Starlight,’ Miles.” Miles became enraged and demanded that the patron be ejected. ‘Get that guy out of here. No one tells me what to play,’ Miles exclaimed.
After the patron had been removed and the dust had settled, Miles turned to the band and announced the next number . . . “Stella by Starlight.”
4 Likes
I hope so. Matt and Tom are working on the bug now. It was quite a revelation and confirms what many of you have discovered independent of our findings. That the DS sounds better with the display off.
We added a sliding dimmer function to the GUI and neither of us paid any attention to it when we started testing. When Darren and I connected the PIC to FPGA 19 we were horrified, the life just got sucked out of the music and our jaws nearly fell off. Ugh. We know the PIC makes a difference, but haven’t heard it make this much a a change. Ready to roll our sleeves up to start at the beginning now with the updated PIC, our Director walks in and says “Have you tried the dimmer?” We hadn’t and Matt opened the screen to find the default (thankfully) had placed it at 50% brightness. As we listened he ramped it up to 100% and the damned life returned to the music!!!
Hope crap! My first thought was, “nuke the dimmer” and let’s ship! But then reason prevailed as Matt said there’s no difference between the finer gradation control of the dimmer and the older control of just low/medium/high brightness. So, a light bulb went off. Perhaps this explains why some people liked the upgrades while others did not. Perhaps this explains why PIC code impacts the FPGA so.
The problem is the display uses PWM to control its brightness. The PWM switches at 1.6kHz, right in the middle of the audio band. It draws a ton of current and as the pulse width narrows to amount of RF increases dramatically and impacts sound quality.
The simple answer is to slow the PWM speed to something reasonable like a couple hundred Hertz. That’;s what they’re working on now but have run into technical problems. Hopefully we get it sorted out today so the brightness of the screen doesn’t muck up the sound.
A great discovery.
9 Likes
Nice story. I thought the Black Hawk was in San Francisco.
Very bright recordings though at least on the CBS box set I have - I get earbleed everytime i listen to it but one of the few instances of Miles with Hank Mobley.
You are correct, thanks. (I must have been thinking subconsciously of the Chicago Blackhawks (the NHL team)).
The “Blackhawk” is one word, though.