schiit bifrost upsampling 24bit?

brad44

Active Member
Looking at schiit bifrost multibit. Im likely misunderstanding this....

This devices upsamples 24/96 via usb?

if the schit dac does this then i
assume all dacs then are upsampling to 24 bit?

Im confused by the technobabble... please advise.
 
The Bifrost MB will intake up to 24/192 via USB. Bifrost MB is an R2R DAC meaning that it plays back what you send it - it does not "upsample" anything. If you read far into the Schiit literature and philosophy you will learn that there are not really any 24 bit recordings (not in actuality). According to Schiit (and this is all beyond my technical capability) the "so-called" 32 bit D/S DACS actually have 19.5 bits (if you do the math-according to Schiit). Read Schiit's FAQ's on the Yggadrisil DAC for a bit more information. For what it is worth I owned the Bifrost MB and it is a very fine DAC. I recently upgraded to the Gungnir MB DAC from Schiit, and it is even better.
 
It does upsample. There is a DSP chip in front of the DAC that will convert all input signals to 176 or 192 kHz sampling rate and 24 bits. The actual DAC chip is 16 bits so the 24 bits is dithered down to 16 bits in the end. 176 or 192 kHz is not upsampled and it gets passed through. Most DACs these days employ oversampling filters in one way or another. There are NOS (non-over-sampling) DACs too, there is a discussion about them here as well (the R2R thread). Btw upsampling generally refers to increasing the sampling rate of the signal and not necessarily the bit-depth.
 
I guess i am confused about the comment that 99.9 percent of recordings are not 24 bit.
 
Last edited:
I guess i am confused about the comment that 99.9 percent of recordings are not 24 bit.

CDs are 16 bit and CD-like quality is probably the bulk of material out there. You can buy and download 24-bit tracks these days from high-resolution audio shops not sure about the ratio but definitely not every release can be had in hi-res. Technically you need a DAC capable of 24-bit resolution or more to get the best possible SQ playing 24 bit material. In reality it is all very subjective and some may not be able to hear any differences between say a 16 or higher bit DACs playing true 24-bit format, and also much depends on how resolving the rest of the system is.

It sounds like it depends on the generation of the Schiit DAC.

The original question was about Bitfrost Multibit which uses Schiit's proprietary upsampling filter in front of the AD5547 16-bit DAC chip.
 
I suppose the OP needs to state what he is trying to achieve and his setup details so that we can provide info in the appropriate context.
 
Looking at schiit bifrost multibit. Im likely misunderstanding this....

This devices upsamples 24/96 via usb?

if the schit dac does this then i
assume all dacs then are upsampling to 24 bit?

Im confused by the technobabble... please advise.
I would suggest emailing Schiit Audio directly with your questions. Getting information second and third hand.

I love the comment above, "CD-like quality is probably the bulk of material out there".

:rflmao:
 
Im currently using a Jolida tube dac III.... and just today am learning about R2R and NOS etc dacs....

I cant tell from the specs, does the Jolida upsample to 24 bits like Bifrost Multibit?

Im assuming dac's that are capable of 24 bits without upsampling are big bucks?
 
Last edited:
I cant tell from the specs, does the Jolida upsample to 24 bits like Bifrost Multibit?

Looks like the Jolida uses PCM1798 which is an oversampling Delta-Sigma, so yes. Unlike the Bitfrost MB the DAC chip is spec'd to allow full 24 bit resolution, even if doesn't realize the full 24 in real life it should in theory be better than 16 in Bitfrost. Which one "sounds better" to you ears is a different question, you can't answer that just by looking at specs.

Im assuming dac's that are capable of 24 bits without upsampling are big bucks?

Yes, but upsampling isn't necessarily bad, in fact it helps to improve DAC specs and if done right it sounds good. Why is this important to you?
 
Last edited:
Looks like the Jolida uses PCM1798 which is an oversampling Delta-Sigma, so yes.



Yes, but upsampling isn't necessarily bad, in fact it helps to improve DAC specs and if done right it sounds good. Why is this important to you?

To be honest, Im not sure why. Blind faith perhaps that I am missing out on 8 bits or more of sonic bliss!!! :confused:
 
To be honest, Im not sure why? Blind faith perhaps that I am missing out on 8 bits or more of sonic bliss!!! :confused:

If you're playing actual 24-bit material and your DAC has a 24-bit or higher rated DAC chip you're not missing much. The Bifrost Multibit is limited to 16-bit at the actual DA stage, so yes you will be missing those 8 bits or at least you will think you do. It almost feels that this is another Multibit vs. Delta-Sigma topic. You need to cough up much more for a "true" Multibit 24-bit DAC than for a comparable Delta-Sigma based solution. There is a prevailing opinion that Multibit sounds better than Delta-Sigma. Multibit DACs can be oversampling, non-oversampling, or both (user selectable). Delta-Sigma is oversampling always.
 
There's no such thing as 24-bit or 32-bit material anyway as it's impossible to get a noise floor low enough to record much more than 20 bits of resolution. Schiit DACs take the approach of maximizing redbook playback since that's 95% or more of most people's music collections.
 
Upsampling is a gimic to me - I have heard both upsampling DACs and non and prefer the non-upsampling DAC. Also, if you run bit perfect - what you are haring is what is in the file or disc.

Unless you are using a pure binary, R2R (or equivalent), 44.1kSa/s, 16-bit DAC, you are probably hearing some upsampling.

Upsampling isn't a gimmick. It's a technique that allows a designer to use a more gentle reconstruction filter, rather than a brick-wall filter. This has benefits for amplitude and phase integrity, and signal-to-noise ratio of the reconstructed signal; that sounds like a good idea, right...?

https://en.wikipedia.org/wiki/Reconstruction_filter

An upsampling, digital interpolating filter is used to 'join the dots' between the samples coming off the digital source. This moves the alias image up in frequency, and spreads the frequency of the output quantisation noise, reducing the noise power spectral density (the same noise power is spread over a wider bandwidth, because the output steps occur more frequently (e.g. 2x, 4x, 8x Fs etc.). This means your reconstruction filter bandwidth can be pushed out from the input Nyquist frequency, and, since the input samples contain no information above the input Nyquist frequency, the reconstruction filter can be designed to roll off gently from the input Nyquist frequency, as the alias image is now much further from the wanted signal. Result: more linear amplitude and phase performance of the reconstructed signal, and reduced quantisation noise.

Fairly good coverage of the noise sources in digital signal conversion here (page 44 onwards)

https://www.intersil.com/content/da...ulators/noise-in-the-signal-chain-webinar.pdf

This gives a good explanation of the benefits of upsampling filtering:

https://www.embedded.com/design/rea...cks-Reducing-A-D-Converter-Quantization-Noise

Or one from Analog Devices:

https://pdfs.semanticscholar.org/3716/a58310339cb34efd19234c132990f8f8ef78.pdf
 
Yep - upsampling allows for the use of a slower roll off filter, which in turn can reduce the amount of pre-ringing in the decoded signal.Thanks for the clear explanation.

Upsampling is not a gimmick.Chord and other DAC manufacturers use upsampling and custom filters that dramatically reduce the amount of pre-ringing and other filter artifacts.

What about 24 bits, anyway?
24 bits is a dynamic range of roughly 144 dB!
A well treated listening room will have a noise floor of 40 dBa. Even given the fact that you can hear below the broadband noise floor, that's still more dynamic range than you can possibly handle in a home environment.

Also the very quietest electronics are running -125 dB snr and that is extremely hard to do. Most are much poorer performing.

The practical limit is more like 20 bits - still 120 dB dynamic range! Most electronics can't handle this due to their inherent thermal noise.

This is what Schiit maintains and why they spec their best DAC at 20 bits.

Tom
 
There's no such thing as 24-bit or 32-bit material anyway as it's impossible to get a noise floor low enough to record much more than 20 bits of resolution. Schiit DACs take the approach of maximizing redbook playback since that's 95% or more of most people's music collections.

Still, having 4 extra bits for playing 24-bit content (20 bits in reality), can't hurt, right? Even for redbook more than 16 bits should help as interpolating filters produce 24 bits (see cpt's comment above).
 
Not sure the extra bits matter because other DAC specs like linearity will limit performance anyway. And besides there are more important things happening in a DAC - like how well the analog circuitry is designed, how well the reconstruction filter is designed, power supply quality, etc.
 
Lots of good info here. I will only add/clarify one point: Almost every DAC increases the bit-depth of incoming signals to 24 bits if the signal is not already 24 bits (or higher). This is not upsampling per se - it's just padding the 16-bit digital samples with extra zeros to make them 24-bit for subsequent processing by the DAC.

In fact, most - and maybe all, not sure - audiophile computer playback apps will output redbook/CD quality files in 24-bit format instead of 16-bit. Even in "bit perfect" mode they will do this, because most DACs "prefer" to be fed 24-bit input rather than 16-bit. Of course the bit-perfect playback apps retain the original 44.1kHz sample rate of CD/redbook sources. It's just the bit depth that they pad.
 
Still, having 4 extra bits for playing 24-bit content (20 bits in reality), can't hurt, right? Even for redbook more than 16 bits should help as interpolating filters produce 24 bits (see cpt's comment above).

The more I read about R2R DACs and their architectures (ladder and string) and their limitations, the more I'm conviced a few extra bits won't really matter because of a host of other limitations. When you start considering the limitations of the recording chain, the ADC and the rest of the playback chain it becomes even less important.
 
The more I read about R2R DACs and their architectures (ladder and string) and their limitations, the more I'm conviced a few extra bits won't really matter because of a host of other limitations. When you start considering the limitations of the recording chain, the ADC and the rest of the playback chain it becomes even less important.

Some newer offerings utilize FPGA to split the input word into 2 ladders, so 24 bits is split apart and 12 bits are sent to 2 16 bit DACs avoiding the linearity issues at LSB, the 2 output analog signals are then summed allowing for full 24 bit dynamic range. That's the claim anyway. This doesn't address the recording chain and other issues but it is at least nice to know that your DAC isn't the bottleneck.
 
Back
Top Bottom