SBAF DAC Talk II

@baldr has said that the "warm up time" is for the DAC chips to error correct, the more bits the longer this takes. For example, Yggdrasil has 21 bits, so it takes longer than Gungnir Multibit with 19 bits, which takes longer than Bifrost Multibit with 16 bits.

 

Interesting, what does this mean exactly?

EDIT: Thanks for the info @LetMeBeFrank

 

These are quotes from Mike on Head-Fi

 

I know what INL is, and I don't see why the corrections can't be completed in more or less real time, if the DAC does function as a DA conversion in real time? You take a whole bunch of measurements and then adjust as necessary. When you're sampling at hundreds of points a second, presumably hitting arbitrary precision on a 20-degree-of-freedom grid is a trivial matter? *

Also, this guy's position on DS DAC's is laughable. I'd happily put my Nagra Classic DAC, with its built-in upsampler, against any R2R design in existence. You can build absolutely fantastic DAC's on either design, and the DS one will happily kick ass on 12W.


* I am a dynamic optimizations guy, not a DAC guy, so this might be a "engineering problem is trivial" moment.

 

I wish I could answer that but I have absolutely no idea. I am merely relaying information. I have seen this topic raised many times and I figure if anyone has the answer to "Why does *insert Schiit Multibit DAC here* take so long to warm up / sound good?" it would be the creator of said DACs. These are the best quotes I could find by him on the topic but you are welcome to search through head-fi and this forum for better ones. Maybe he will chime in here since I tagged him.

 

Huh. Didn't notice that the quotes are from Baldr himself. I would like a technical explanation if he sees this post.

It just seems that he's going for the "you won't get it even if I explained it to you, this is a feature of our design that's massively superior to alternatives, and if you want to debate about this, go to another subforum" track in those posts, and that alway smells fishy to me.

 

Wasn't Moffat instrumental in the development of some of these technologies? I'd say that earns him quite a bit of legitimatacy when it comes to opinions on certain DAC types. I'm all for open discourse, but I always groan when randos start throwing their weight around, demanding vetted professionals continually validate their (extremely informed) opinions.

If there's a better way to implement schitts multibit tech, I'm sure they would've figured it out by now. Waiting a few days for a DAC to warm up to sound optimal is a small price to pay for the ridiculous sound quality.

 

Just to Second @Zhanming057 questioning about the “warmup” explanation: The technical explanations given are as yet not convincing to me either. Nonetheless, despite my doubts of the putative cause, the effect of long settling time seems to be audiuble to me and thus “real” (in my reality).

Blader’s statement about DS being inherently bad is inherently bullshit. The same could have been said about CCD vs CMOS imaging chips many years ago. CMOS were noisy, had low areal coverage, and sucked in almost every way but price. But R&D drove CMOS to be incredibly good and they now outperform CCD in an amazing number of performance metrics. Technology religion leads to stupid religious wars. Remain open minded and let your ears be the judge on a case-by-case basis. I point to Convert-2, by way of example.

 

Regardless of how integral Mike was to developing the technology in question, I am not a fan of technical people casually blowing off technical questions. If the answer is too long for a forum post, he could always put up the technical documentations on a blog somewhere. If he's not sure about what's going on, he should say that straight up. "Go to Sound Sciences" is the technical person's equivalent of "I know you are but what am I".

"If there's a better way to do it the pros would have done it" and "groaning about demanding vetted professionals to validate their opinion" is an appeal to authority. Pros make big errors in judgement all the time, and Schiit, while having some fantastic products, is far from the last word on resistor ladder DAC design.

 

After being asked for the bazillianth time? I'm surprised he's so casual about it.

 

The dac chips are what? .. milliwatts of dissipation. They will be largely influenced by psu and local regs heat generation.
It will take the whole thing days to creep up the log scale in internal temperature.
More annoyingly the external conditions are going to have an effect as the internal power generation is relatively tiny.
Why do 6..8 digit multimeters have oven for vref? 19 bits of 'accuracy' is in this playground.

Class A poweramp output stage is literally an oven. It doesn't take it long to push the whole enclosure to equilibrium. Usually no more than hours. And hold it there.

 

That's appeal to authority again. If this is indeed a frequent question, and if this was me, and I have gotten repeated questions about one particular technical aspect of my work, I would take an hour, write a few hundred words about it, and paste the link every time I got the question. Even better, put a link to it somewhere in the DAC's product page. No need for "if you don't believe in anecdotal evidence then GTFO" type responses.

At the very least, it's unprofessional.

If the DAC chip temperature is affected by heat from the PSU and the output stage, won't the system reach thermal neutral faster? If the chassis is hot, I don't see why the DAC chip won't be exactly the same temp as the rest of the machine.

 

I like coming here to learn and hear people's opinions, but I don't think the technical and business minds behind a company are in any way required to chime in on certain details when the question comes from the general public. We can try to make this into a direct link to those technical minds and expect them to respond to our every desire, but it's unlikely to work out well for the majority of requests. I'd love to see Schiit designers or anyone else spend hours and days of their own time pandering to my need to "know" but you have to be reasonable when they are unwilling, unable, or simply don't care to entertain your requests. Try not to take offense to their position because it probably comes from a standpoint of being swamped with work, not due to arrogance or lack of interest.

I can relate to your comments about Nagra DACs. I'm not sure what is in it, but my local dealer is big on Nagra and the new HD series. I've had an insatiable appetite for MSB DACs ever since I heard my first one about 7 years ago. We recently did an A-B test with the Nagra HD and the MSB Premier DAC. I am beyond upset that the Nagra HD sounded better. It was more real, better tonality, and presented a more realistic sound stage for this 5 piece jazz band we were listening to. The MSB took the same recording and altered the tonality of the instruments such that they didn't sound like real instruments. It also created a sound stage that felt artificial and "all over the place." The difference was immense and not subtle. They leave both DACs on 24/7. The MSB was not at all bad, but the difference was so startling that I am still having a hard time coping with what it has done to my personal ideologies.

 

I guess my point boils down to the following:

- If the question demands a lengthy response, then be upfront about that and say that if you have the time, you'll write a formal explanation.
- If you get the question a dozen extra times, maybe go ahead and write that explanation because it's apparently important to your customers. If you have the technical details somewhere, publishing an excerpt would go a long way.

I'm not saying that he's obligated to respond, but the stuff he has said doesn't pass the sniff test, and I do feel that when people are spending $2,000 on a DAC, it's not unreasonable that they ask some tough questions. I build big computer simulations involving hundreds of CPU's and dozens of algorithm bits assembled over more than a decade. If one of my (very much) non technical clients ask about a technical detail, regardless of how trivial it is, the only two responses I'll ever give is 1. the concise argument about why we're doing this and 2. I'm not sure and I will find out ASAP.

Your comment about the MSB is interesting, I've been dying to try a MSB Reference or Select DAC on my setup and now I wonder if I should just stick to the Nagra and be happy about it. The new DCS stuff has a ton of merits but ultimately I don't like the way they nickel and dime you for the upsampler and external clock, and the extra cost wasn't worth it over the Nagra.

 

FWIW regarding the warm up, I haven't heard an explanation yet that makes sense to me either, but I accidentally left my Modi MB on last night and I honestly felt it sounded a little different (I usually only let it warm up for 10min to a few hours max). Thats what prompted me to post the question. So like @skem I figure something is going on, I just don't know what it is or understand it (yet).

On a side note, sweet baby Jesus, what speakers/headphones are you guys using with $20,000 DACs? There are a lot of $20,000+ speakers I'd love to own but Ive never even contemplated a $20,000 DAC.

 

It was EgglestonWorks "The Viginti" speaker with Audia Flight Strumento electronics.

 

Do Nagra still have that handy remortgaging payment plan?

Point taken on the whole thing around Multibit being the “one true DAC architecture to rule them all”. It’s a given that there are other DACs out there that can probably smoke Yggdrasil. But at what price?

Convert-2 seems to show that there are D/S implementations that are starting to play in the same ballpark as Yggdrasil for similar cash. And to me that’s great because variety and competition is a good thing. If only there were other consumer oriented products doing this in the price range that Schiit play in.

On the warmup thing, it’s a minor inconvenience, but you can definitely hear it when it’s cold, very clearly. Don’t really care why considering the sound quality achieved as a result of waiting. It’s like decanting certain wines before drinking them - a bit of patience goes a long way for those with discerning tastes.

 

Do take in mind what kind of music Mike uses for DAC evaluation. He's mostly a classical and recorded music listener which the Yggdrasil and Gungnir and multibits do excel at. Delta sigmas commit sins of omission at best and bludgeon your ears with high freq grain at worst according to Mike, which I do agree with but I don't think it completely ruins the music experience and can have some benefits like blackground and lack of warmup,etc... Yes, just as with the CMOS/CCD thing, there are still film fanatics that claim CMOS is too "clean" looking. It very much depends on what music you listen to and what presentation you're expecting from that music whether you're gonna like a DAC. Read @Cellist88's impression of the Convert-2 to see what I mean.

 

The vast majority of gear buyers don't seem to be interested in the minutea of DAC chip function. So I don't believe Schiit cares to put it on the product pages, and I don't blame them. Their primary concern as a retailer seem to be sales, and they seem to be doing a bang up job of it so far.

And in the era of Google, letting the search function do its job seems to be enough for people who really wanna know what Mike thinks. The information is already in cyberspace for the curious.

I mean, now we're getting into "well, if it were me" territory, which becomes akin to arguing about somebody's favorite color. What's the point.

 

I'd rather ignore price in discussing dacs beyond a certain price level (say, 2k usd). They're all cost-no-object things imo.

I can't think of any specific products that exceed yggy2 in every major criterion (information amount, accurate-ish presentation, and timber) beside personal sonic tastes, no matter how much $. Some are rarely in the same league -- would put convert-2 and dave (maybe a little behind) into this league. But nothing to date assured me to believe the existence of performance-wise "higher" leagues. Secretly have some fantasy on bricasti (trying to find some chance to hear it) though.

Can't say much about msb or nagra because of unfamiliar configurations when I auditioned a while ago. But clearly not motivated to hear again. Acceptable at best.