Discussion in 'Digital: DACs, USB converters, decrapifiers' started by Bill-P, Oct 7, 2015.
Speaking of modifications. In the old diy days installing a 'better' clock in CD-players was considered a must have improvement. The internal clock for the A2 has a frequency stability of 50 ppm (which is pretty bad compared to other clocks); femto clocks go down to 10 ppb. Has anyone tried installing a better clock in his/her Gungnir Multibit?
As a drunk idiot on the forum, I have to ask..... Where do you get the frequency stability metric for Gungnir Multibit from? And what makes it bad? 50 ppm.....parts per million? compared to 10 ppm?
.0050 vs .001 ??
I think it's 50 ppm = 50/1000000 = 0.00005 vs 10 ppb = 10/1000000000 = 0.00000001
The question is, does that make an audible difference? guessing no
The 50 ppm @ 12Mhz of the gungnir clock (ATS120B) translate to 4,17 ns jitter. (When my calculations are correct) With the graph I found on the internet the influence of a factor 5000 (50ppm to 10 ppb) seems to add up to -28 db s/n. When you start out with -80 db this seems to be (at least) close to the audible range. Maybe someone can elaborate on the impact of this amount of jitter on sound.
I'm really looking forward to the phase noise measurements at the word clock with PPM vs PPB oscillators. Bonus points for multiple different manufacturers' products, correlation to sonics, and discussion of applicability with respect to internal jitter-reduction mechanisms as used in many delta-sigma DACs.
I am a fan of a more analog sound. That is why I like my Gungnir. My experiences are that perceived audio quality is largely depending on power supply quality and 'machines' downstream can't fix the problems caused upstream. Hence my interest in the influence of clocks on sound. A TI paper states (p16): the DAC output noise is a simple summation of clock noise and DAC intrinsic noise.
Maybe fellow audio freaks active on forums that are more measurement oriented can help answering the comparison question between ppm and ppb clocks. Or it can be an interesting topic for a future Schiitr event comparing two dacs with different quality clocks. I think it could even be a business wise interesting upgrade path for the future.
Better femto, atto, and zepto clocks are passe. It was almost a prerequisite back in the day in DIY / modding to buy $732 clock modules. WIth respect to audiophile uses, I didn't care for the turboclock upgrade on the Off Ramp 5 or having the Mutec MC-3 smart clock in the chain. They made the sound from the DACs different rather than better, and to my tastes, worse: too tight, too focused, unnatural. Heck, remember LHLabs, I heard one of their super-upgraded Geek Pulse DACs with femto-whatever clocks and it sounded like ass (as if it had a really crappy clock, loose, fuzzy, smeared, etc.)
I can guarantee they won't. Mike Moffat over at Schiit was a pioneer with the "jitter" thing back in the Theta days. I'm pretty sure the guys at Schiit done what they needed to do to solve the problem and see no more need to advance with better clocks. I heard that the latest jitter optimization was with their USB interface via their Unison solution. It's not 1994, nor 2010 anymore. Pretty sure their R&D is focused on the D to A side. This is where there will be a real difference.
P.S. Multi-bit DACs are less sensitive to clock jitter. Think about it. DS DACs run the MHz.
Clock jitter noise isn't white noise in behavior. Best to pick a clock that doesn't pollute the most critical parts of the audible spectrum. A clock with great close-in phase noise but random crap spewed broadband noise could sound worse than a clock with decent close-in phase noise but less broadband noise.
Minor tangent but while I've always had gearhead tendencies it wasn't until I built a desktop PC for the first time early last year and really looked into the hows of various processor architectures (e.g. why clock speed isn't the end-all-be-all metric to rule all and why the Ryzen 5800X3D kicks the vanilla 5800X in the nethers for gaming due to stacked cache but takes a dive in Cinebench) that I realised common wisdom here of "more expensive/bigger number != betterer" applies to a lot more in the tech space (and jewellery as well, but that's another matter entirely).
Call me a dumbass but don't think it's necessarily common knowledge that multibit DACs are less reliant on clock accuracy and that the differences between the approaches have system-wide ramifications with how it seems like people are still advertising rubidium atomic clocks with their high end DACs.
Won't do much at all to engage the "all DACs are the same" crowd but could be fun to see how DS and MB might have respective strengths and weaknesses when paired with suboptimal upstreams like really cruddy USB hubs out of a laptop.
Did clocks make more of a difference back in the earlier days of usb audio when it was isochronous/synchronous/whatever the heck it was called back then? I vaguely remember it being the "new" thing back then when everybody was bragging that their interfaces were asynchronous vs whatever the old transfer method was.
edit: let's just ignore the whole "using a master clock to sync multiple devices in a mastering studio" blah blah
For async USB, it made a difference between the data was reclocked by a dedicated master clock in the DAC.
Discuss of clocks in DACs for home audiophile purposes gives me PTSD.
I am playing with using spidif with my Gungnir MB it has Unison USB. I bought a used around 10 year old Stella U2 usb to spidif convertor and after about 20minutes of play it becomes very garbled? The seller swears it worked perfectly for him so could it be the Gungnir? The cable Im using is a cheap (supposedly 75ohm) one from amazon(china). Was suggested its a bad cap and the guy says it worked without problem for him. Any thoughts or suggestions from those with experience with all of this much appreciated.
I had Gungnir Multibit fed from a TV for a while and experienced a similar problem. The TV has since fed another dac without problems. And the Gungnir Multibit has been fed by other sources (CCA, Pi2AES, HiFiBerry..) without problems.
Wishlist for Gungnir Multibit upgrade (inspired by recent bifrost upgrades )
No return to factory - firmware upgrade by user with addition of SD card socket
Someone correct me if I am wrong but I believe Schiit DSP on Gungnir and Yggdrasil is an 8x upsample that tops out at 176.4/192khz so I think that's your speed limit unless that changes, which I am not sure there is a material need for. I'm sure Rob Watts disagrees.
8x 44.1k = 352.8k
8x 48.0k = 384.0k
The 176.4/192kHz rate you're thinking of is likely a SPDIF limit. That's why upsampling to the higher rates has to be via USB.
I don't think the upsample starting point is 44.1/48 if I'm reading correctly.
This is combined with a frequency domain optimization which does not otherwise affect the phase optimization. The 0.968 of Nyquist also gives us a small advantage that none of the off-the shelf FIR filters (0.907) provide: frequency response out to 21.344KHz, 42.688KHz, 85.3776KHz, and 170.5772KHz bandwidth for native 1,2,4, and 8x 44.1KHz SR multiple recordings - the 48KHz table is 23.232, 46.464, 92.868, and 185.856KHz respectively for 1,2,4, and 8x. This was the portion of the filter that had the divide by zero problem which John Lediaev worked out, to combine with #4 above AND retain the original samples.
I'm kinda guessing that the latest revision of the Bifrost *is* the "Gungnir upgrade" that's been so long coming. But this is purely speculation of course.
Per @purr1n's DAC chart in another thread, it looks like both the A1 and A2 have the same relative greyground. The greyness of the A1 was my only complaint with the Gungnir, loved the liquidity and tonality, but having never heard the A2, any Yggdrasil and the Bifrost 2/64 for only a few songs, would it be advisable to go for a used A1 for ~$750 or is the A2's step-up in technicalities enough "betterer" to warrant the colder/different tonality?
Or should I be looking at saving up for the LiM/Yggdrasil OG? From what I heard of the BF2/64, I wasn't quite as impressed as I was with the A1 Gungnir, so I'd like to leave that DAC off the table even though it had great blackground (or unless someone can convince me it's a better DAC for the money). I'm sure with speakers though, that thing kicks.
(For context my system is going to look like this: Pi2AES/Mercury Streamer > BNC/AES > Gungnir/Yggdrasil > Liquid Gold X > RD-LCD-X)
You have to hear things in your chain and know what you're looking for.
If you like how Gungnir A1 sounds then Yggdrasil A2 is along those general lines with better technicalities. Gungnir A1 was my DAC for years and it's a great DAC but I think Yggdrasil A2 is superior in basically all respects.
I haven't heard Gungnir A2 but it's supposed to be close to Yggdrasil A1/GS with the more neutral signature.
If you like how Bifrost 2 sounds then LIM is like that. Not sure how close 2/64 gets to LIM.
Separate names with a comma.