[kstuart]

  OK so if I understood correctly all these little things add up to provide a superior subjective listening experience and I assume better measured sonic performance even though the listed specs for M2U and A2 are almost exactly the same within reasonable margin of error?

Generally, "no feedback" means worse measured performance and better subjective listening experience.  It was the first big schism between "objective measurement" and "subjective listening".   IIRC, in the 1980s, Yamaha had a big program of getting the very best specs possible, and used lots of feedback to accomplish this, which resulted in a line of really clean and dull sounding amplifiers.  This led to many "no feedback" designs in reaction.

 

[Vtsailor]

I would like to second, third?, the vote for a preamp or amp with remote. My spouse is in a wheelchair, it would sure be nice to be able to turn down the speakers without having to truck over to the stereo when you need to adjust the volume.

 

[atomicbob]

  Atomic Bob
  First off I must say I a very envious that you have all three MultiBit Schiit Dacs and also the fact that you are paid to have some awesome equipment at your disposal. Your Honest Opinions of each DAC and not making claims of which one is endgame, end all be all...etc is probably one of the Most Insightful,Concise and Unbiased comparisons I have read. You managed to not only answer some of the questions I had about The differences between the SE Outputs of the Gumby vs. the Bimby but also present a good case for each Dac to help those who are considering a multibit figure out which way they should go. The FAQ question about Schiit Multi Bit Dacs should just be redirected to your last post.

Thank-you, I do feel very privileged to work in the field of my passion. However, all the DACs I've been evaluating have been bought personally, for my personal acoustic laboratory. Some might have a boat, or fix up vintage cars or some other money consuming hobby; acoustic research is both my profession and hobby. From my evaluations I make recommendations to my day job for my acoustic lab on the corporate campus. Over the years I've received some unexpected stock awards that have helped pay for it all. Corporate is smart in that respect, they know exactly what I will do and how they benefit.
 
I appreciate that there is no single perfect component or system that is "the holy grail" of sound reproduction, just as on the recording side, there is no single microphone that does it all, above all others. They each have their respective strengths and weaknesses. I am not a fan of the sports ranking of components, that one is a winner and one a loser. There are so many variables to the components, in feature set and sound qualities that I prefer to understand how a given component might fit into a particular system.
 
I am also somewhat infatuated with attempting to find the lowest cost system that delivers sound befitting of being considered a reference system. If such a reference system was created and available at the many meets around the globe, it could serve as a known sound. This could provide better comparison comments with other components / systems at the same meet. Two components already figure prominently in such a system, Bifrost MB and HD650. I still haven't settled on the amp. For a tube hybrid I am rather fond of the Project Sunrise III with a 12BH7 tube and Talema linear power supply. It has been suggested that I spend some time with an Asgard 2. Both are non-feedback class A designs with near DC to very high frequency responses. So an Asgard evaluation will be in my future. In any case, we are all indebted to Schiit for creating such an extraordinary line of multibit source components.

 

[atomicbob]

  Generally, "no feedback" means worse measured performance and better subjective listening experience.  It was the first big schism between "objective measurement" and "subjective listening".   IIRC, in the 1980s, Yamaha had a big program of getting the very best specs possible, and used lots of feedback to accomplish this, which resulted in a line of really clean and dull sounding amplifiers.  This led to many "no feedback" designs in reaction.

Years ago I would have been one to challenge the no-feedback concept. Then someone challenged me that I get off my butt and do some listening. I have a new appreciation for the NFB concept and own several for my personal sound lab. Do you have any citations to others attempting to measure objectively and evaluate subjectively why the NFB provides a better subjective listening experience? I'd like to get my head around why I too appear to be enjoying NFB designs. Is it expectation bias on my part or real? (this last question is rhetorical, only I can answer that for my own experiences.)

 

[FrivolsListener]

  Thank-you, I do feel very privileged to work in the field of my passion.

 
So, all your Schiit is tax-deductible!
 
(Runz away fast)

 

[kstuart]

So, I wanted to clear up my understanding on one point.  I dug out the following post from (!) four years ago about the original Bifrost:
 

  Hey all,
 
To answer the above question, we do *not* do asynchronous sample rate conversion, or "upsampling." We're philosophically opposed to it, since it destroys all the original samples in the recording. Of course, so do successive-approximation D/A converters (aka delta-sigma--pretty much every major converter on the market today--Sabre, Wolfson, AKM.) Although we cannot do anything about the converter, we choose to preserve the original samples as far down the chain as possible. 
 
However, the AKM D/A does perform oversampling. Oversampling is done by the digital filter preceeding the delta-sigma modulator, and the amount of oversampling varies by sample rate.  
 
All the best,
Jason

So, the oversampling that was done by the AKM 4399 D/A in the original Bifrost, is now done by the "megaburrito filter" in the Multibit Bifrost, and we can hear the improvement.
 
But, can Jason - or anyone - shed some light on the  asynchronous sample rate conversion, or "upsampling"  that Jason says is not done in Schiit DACs.   What is the purpose of that upsampling if the D/A unit is going to oversample later anyway?
 
 

 

[madwolfa]

  But, can Jason - or anyone - shed some light on the  asynchronous sample rate conversion, or "upsampling"  that Jason says is not done in Schiit DACs.   What is the purpose of that upsampling if the D/A unit is going to oversample later anyway?

 
I don't think they're against "upsampling" per se, but specifically against losing the original samples and approximating the output in process. Having that problem solved with megaburrito filter, they can happily implement it in Schiit DACs, using all the advantages without breaking the principle.
 
EDIT: I think I see your point, there's a confusion in terminology here, "upsampling" vs "oversampling"... I think "upsampling" is a type of asynchronous sample rate conversion, since it can also be used to downsample or whatever. It's usually done with a separate ASRC IC in front of the DAC itself. So the DAC never-ever gets to see the original samples; they're all decimated by ASRC. And then I guess there's no way to get rid of internal "oversampling" in D/S DACs as it's used "by design" to shift the HF noise out of the audible band. No such need in multibit DACs (and so many of them are labelled NOS - non-oversampling), but there's still benefit from using a quality oversampling filter like megaburrito.

 

[jfoxvol]

   
I don't think they're against "upsampling" per se, but specifically against losing the original samples and approximating the output in process. Having that problem solved with megaburrito filter, they can happily implement it in Schiit DACs, using all the advantages without breaking the principle.
 
EDIT: I think I see your point, there's a confusion in terminology here, "upsampling" vs "oversampling"... I think "upsampling" is a type of asynchronous sample rate conversion, since it can also be used to downsample or whatever. It's usually done with a separate ASRC IC in front of the DAC itself. So the DAC never-ever gets to see the original samples; they're all decimated by ASRC. And then I guess there's no way to get rid of internal "oversampling" in D/S DACs as it's used "by design" to shift the HF noise out of the audible band. No such need in multibit DACs (and so many of them are labelled NOS - non-oversampling), but there's still benefit from using a quality oversampling filter like megaburrito.

 
Here's a good breakdown in the terminology.  I just broke out my undergrad linear system theory book to confirm but below gives better illustrations. Oversampling and Upsampling are similar, but differ in their location in the signal chain.  They can be synchronous or asynchronous.  Both synch and asynch are interpolaters but the latter absolutely does away with original samples.  The former (synchronous) may or may not depending on implementation and the type of filter.
 
http://www.audioholics.com/audio-technologies/upsampling-vs-oversampling-for-digital-audio

 

[kstuart]

 

   
I don't think they're against "upsampling" per se, but specifically against losing the original samples and approximating the output in process. Having that problem solved with megaburrito filter, they can happily implement it in Schiit DACs, using all the advantages without breaking the principle.
 
EDIT: I think I see your point, there's a confusion in terminology here, "upsampling" vs "oversampling"... I think "upsampling" is a type of asynchronous sample rate conversion, since it can also be used to downsample or whatever. It's usually done with a separate ASRC IC in front of the DAC itself. So the DAC never-ever gets to see the original samples; they're all decimated by ASRC. And then I guess there's no way to get rid of internal "oversampling" in D/S DACs as it's used "by design" to shift the HF noise out of the audible band. No such need in multibit DACs (and so many of them are labelled NOS - non-oversampling), but there's still benefit from using a quality oversampling filter like megaburrito.

 
Here's a good breakdown in the terminology.  I just broke out my undergrad linear system theory book to confirm but below gives better illustrations. Oversampling and Upsampling are similar, but differ in their location in the signal chain.  They can be synchronous or asynchronous.  Both synch and asynch are interpolaters but the latter absolutely does away with original samples.  The former (synchronous) may or may not depending on implementation and the type of filter.
 
http://www.audioholics.com/audio-technologies/upsampling-vs-oversampling-for-digital-audio

Excellent, that's a good explanation, thanks.

 

[RCBinTN]

  Thank-you, I do feel very privileged to work in the field of my passion. However, all the DACs I've been evaluating have been bought personally, for my personal acoustic laboratory. Some might have a boat, or fix up vintage cars or some other money consuming hobby; acoustic research is both my profession and hobby. From my evaluations I make recommendations to my day job for my acoustic lab on the corporate campus. Over the years I've received some unexpected stock awards that have helped pay for it all. Corporate is smart in that respect, they know exactly what I will do and how they benefit.
 
I appreciate that there is no single perfect component or system that is "the holy grail" of sound reproduction, just as on the recording side, there is no single microphone that does it all, above all others. They each have their respective strengths and weaknesses. I am not a fan of the sports ranking of components, that one is a winner and one a loser. There are so many variables to the components, in feature set and sound qualities that I prefer to understand how a given component might fit into a particular system.
 
I am also somewhat infatuated with attempting to find the lowest cost system that delivers sound befitting of being considered a reference system. If such a reference system was created and available at the many meets around the globe, it could serve as a known sound. This could provide better comparison comments with other components / systems at the same meet. Two components already figure prominently in such a system, Bifrost MB and HD650. I still haven't settled on the amp. For a tube hybrid I am rather fond of the Project Sunrise III with a 12BH7 tube and Talema linear power supply. It has been suggested that I spend some time with an Asgard 2. Both are non-feedback class A designs with near DC to very high frequency responses. So an Asgard evaluation will be in my future. In any case, we are all indebted to Schiit for creating such an extraordinary line of multibit source components.

 
That would be great.  I'm close to recommending the BiMB - A2 - HD650 as a really decent mid-fi rig to a Head-Fi friend.  Please send your impressions, if you have time.  
 
Cheers -
RCB

 

[Odin412]

  That would be great.  I'm close to recommending the BiMB - A2 - HD650 as a really decent mid-fi rig to a Head-Fi friend.  Please send your impressions, if you have time.  

I​ use the Bifrost Multibit - Valhalla - HD650 combo and for just sitting back, relaxing and listening to music for enjoyment I have yet to hear a better combo. I have an Asgard 2 as well and it works great with the HD650 - however, I prefer the Valhalla. I use the Asgard 2 with my Audeze LCD-2Fs since they are too low impedance for the Valhalla to drive.

 

[RCBinTN]

 
I​ use the Bifrost Multibit - Valhalla - HD650 combo and for just sitting back, relaxing and listening to music for enjoyment I have yet to hear a better combo. I have an Asgard 2 as well and it works great with the HD650 - however, I prefer the Valhalla. I use the Asgard 2 with my Audeze LCD-2Fs since they are too low impedance for the Valhalla to drive.

 
Thank you - great recommendation.  I'd forgotten about the LCD-2F for my friend...maybe a better choice than the HD650's.

 

[hodgjy]

   
Thank you - great recommendation.  I'd forgotten about the LCD-2F for my friend...maybe a better choice than the HD650's.

 
That depends. Once upon a time, I thought the LCD-2F was the clear cut winner over the HD650. However, once I upgraded to a Gumby, it's not such an easy choice. Both shine, but for different reasons. The HD650 has been getting more head time lately.

 

[tonykaz]

AtomicBob and the "lowest cost reference system" .
 
Wonderful concept, a base line, a "Standard" of comparison.   
 
Schiit & Sennheiser stuff would be ideal as every meet would have folks that own these devices. I'd be thrilled to hear that Garage1217 Project Sunrise vs the "Standard".  Geez, this would give all of us a baseline for our conversations. 
 
Meet attendance would double or more.
 
I wonder if Stoddard would object?  Might be a little scary, always being under the gun over the field performance of their gear, globally. phew
 
A universal performance Standard available for comparison would sell one hell of a lot of gear, it'd be good for the headphone industry.
 
Tony in Michigan

 

[Deftone]

  Generally, "no feedback" means worse measured performance and better subjective listening experience.  It was the first big schism between "objective measurement" and "subjective listening".   IIRC, in the 1980s, Yamaha had a big program of getting the very best specs possible, and used lots of feedback to accomplish this, which resulted in a line of really clean and dull sounding amplifiers.  This led to many "no feedback" designs in reaction.

sounds like the O2 amp.. its was clean, cold and boring, also i might add horrible clipping at high volumes.