[Currawong]

DACs essentially cannot reproduce the equivalent of more than 21 bits of resolution (and rarely even 20-bit equivalent) due to the physical limitation of electronic components.*

What is a "bit" in audio terms anyway?

Digital numbers are basically base 2 values. So "1, 2, 3, 4, 5, 6, 7, 8, 9, 10" becomes "1, 10, 11, 100, 101, 110, 111, 1000, 1001, 1010". Each digit is a "bit", so 16 bits equals up to 16 digits, from "0000000000000000" to "1111111111111111". Knowing that each added bit doubles the numbers (a factor of 2, whereas in regular base 10 maths, each bit/numeral multiples the available numbers by a factor of 10) 16 bits gives us 65,536 values. 24 bits gives us 16,777,216 values. These values are the maximum number of levels of sound available to record digital audio in a 16-bit and 24-bit file respectively.

Volume is measured in dB, and in a file, in -dB. That is, starting at -0db (maximum volume) each bit describes how many levels, in 6dB increments, below maximum volume (-0dB) we have to work with. That means -96dB for a 16-bit file, and -144dB for a 24-bit file. Below about -120dB you begin to hit the physical limitations of electronics, so about 20-21 bits of resolution are the physical maximum.

Dithering allows 16-bits to reproduce audio below its -96dB limit, but that's another discussion.

The Modi Multibit uses a 16-bit DAC. Measurements show that it starts to lose the ability to resolve small signals at -90dB, so anything higher res than that, eg: a 24-bit file, is simply a complete waste of time to use with it. In my own listening, through various high-end amps, you can make out the distortion from it quite easily. With cheap amps, up to about the Lyr 3, it doesn't have any very audible detrimental effect. With a Vali 2 or CTH, you just wont notice. It actually sounds nice. With the new Lyr 3 it's pretty good too.

The DACs in the Schiit multi-bit line are industrial DACs, as has already been explained. I'm guessing that the specific purpose of these DACs is for applications such as fine motor control, where output voltages are constant (and wont run into the low-signal-level issues that occur in audio applications) and the output has to be exact.

Something I might try is seeing if I still can use software that will output a pre-dithered signal to the Mimby and see if that improves things. IIRC Mike doesn't use dithering on any of his multi-bit DACs (otherwise he wouldn't be able to say that they are processing the original digital values) but it might improve the low-level linearity.

*Measuring equipment gets around this in various ways, but that's another thing altogether.

 

[BrotherKathos]

DACs essentially cannot reproduce the equivalent of more than 21 bits of resolution (and rarely even 20-bit equivalent) due to the physical limitation of electronic components.*

What is a "bit" in audio terms anyway?

Digital numbers are basically base 2 values. So "1, 2, 3, 4, 5, 6, 7, 8, 9, 10" becomes "1, 10, 11, 100, 101, 110, 111, 1000, 1001, 1010". Each digit is a "bit", so 16 bits equals up to 16 digits, from "0000000000000000" to "1111111111111111". Knowing that each added bit doubles the numbers (a factor of 2, whereas in regular base 10 maths, each bit/numeral multiples the available numbers by a factor of 10) 16 bits gives us 65,536 values. 24 bits gives us 16,777,216 values. These values are the maximum number of levels of sound available to record digital audio in a 16-bit and 24-bit file respectively.

Volume is measured in dB, and in a file, in -dB. That is, starting at -0db (maximum volume) each bit describes how many levels, in 6dB increments, below maximum volume (-0dB) we have to work with. That means -96dB for a 16-bit file, and -144dB for a 24-bit file. Below about -120dB you begin to hit the physical limitations of electronics, so about 20-21 bits of resolution are the physical maximum.

Dithering allows 16-bits to reproduce audio below its -96dB limit, but that's another discussion.

The Modi Multibit uses a 16-bit DAC. Measurements show that it starts to lose the ability to resolve small signals at -90dB, so anything higher res than that, eg: a 24-bit file, is simply a complete waste of time to use with it. In my own listening, through various high-end amps, you can make out the distortion from it quite easily. With cheap amps, up to about the Lyr 3, it doesn't have any very audible detrimental effect. With a Vali 2 or CTH, you just wont notice. It actually sounds nice. With the new Lyr 3 it's pretty good too.

The DACs in the Schiit multi-bit line are industrial DACs, as has already been explained. I'm guessing that the specific purpose of these DACs is for applications such as fine motor control, where output voltages are constant (and wont run into the low-signal-level issues that occur in audio applications) and the output has to be exact.

Something I might try is seeing if I still can use software that will output a pre-dithered signal to the Mimby and see if that improves things. IIRC Mike doesn't use dithering on any of his multi-bit DACs (otherwise he wouldn't be able to say that they are processing the original digital values) but it might improve the low-level linearity.

*Measuring equipment gets around this in various ways, but that's another thing altogether.

Thanks for that. It was very interesting and informative.

 

[rutter]

DACs essentially cannot reproduce the equivalent of more than 21 bits of resolution (and rarely even 20-bit equivalent) due to the physical limitation of electronic components.*

What is a "bit" in audio terms anyway?

Digital numbers are basically base 2 values. So "1, 2, 3, 4, 5, 6, 7, 8, 9, 10" becomes "1, 10, 11, 100, 101, 110, 111, 1000, 1001, 1010". Each digit is a "bit", so 16 bits equals up to 16 digits, from "0000000000000000" to "1111111111111111". Knowing that each added bit doubles the numbers (a factor of 2, whereas in regular base 10 maths, each bit/numeral multiples the available numbers by a factor of 10) 16 bits gives us 65,536 values. 24 bits gives us 16,777,216 values. These values are the maximum number of levels of sound available to record digital audio in a 16-bit and 24-bit file respectively.

Volume is measured in dB, and in a file, in -dB. That is, starting at -0db (maximum volume) each bit describes how many levels, in 6dB increments, below maximum volume (-0dB) we have to work with. That means -96dB for a 16-bit file, and -144dB for a 24-bit file. Below about -120dB you begin to hit the physical limitations of electronics, so about 20-21 bits of resolution are the physical maximum.

Dithering allows 16-bits to reproduce audio below its -96dB limit, but that's another discussion.

The Modi Multibit uses a 16-bit DAC. Measurements show that it starts to lose the ability to resolve small signals at -90dB, so anything higher res than that, eg: a 24-bit file, is simply a complete waste of time to use with it. In my own listening, through various high-end amps, you can make out the distortion from it quite easily. With cheap amps, up to about the Lyr 3, it doesn't have any very audible detrimental effect. With a Vali 2 or CTH, you just wont notice. It actually sounds nice. With the new Lyr 3 it's pretty good too.

The DACs in the Schiit multi-bit line are industrial DACs, as has already been explained. I'm guessing that the specific purpose of these DACs is for applications such as fine motor control, where output voltages are constant (and wont run into the low-signal-level issues that occur in audio applications) and the output has to be exact.

Something I might try is seeing if I still can use software that will output a pre-dithered signal to the Mimby and see if that improves things. IIRC Mike doesn't use dithering on any of his multi-bit DACs (otherwise he wouldn't be able to say that they are processing the original digital values) but it might improve the low-level linearity.

*Measuring equipment gets around this in various ways, but that's another thing altogether.

So the Yggy handles hi-res music as well as anything out there or is there misleading by omission going on?

You mention distortion with the Mimby. Is this the kind of distortion that is detrimental to hearing or?

 

[cobrabucket]

Wow! Thanks for all the insight! I have been using 16-bit files with the Modi Multibit and Hi-Res and DSD with the SU-8. To my ears, at least, this seems to sound the best with the equipment I currently own. BTW, been thinking about getting a tube headphone amp and am currently leaning towards a Project Ember. Any thoughts?

 

[Currawong]

So the Yggy handles hi-res music as well as anything out there or is there misleading by omission going on?

You mention distortion with the Mimby. Is this the kind of distortion that is detrimental to hearing or?

Nothing misleading. I recall the original Yggy truncating the last 4 bits, so I guess the Analog 2 does the same, since the DAC is only 20 bit. Mike has posted about it.

As for the Mimby, if I A/B it with the Yggy, for example, female vocals sound a bit wobbly from it in a way that reminds me a bit of old MP3 files. Just to emphasis, this is in comparison on a high-end amp. It's not something I notice so readily otherwise, especially not through a Magni 3 or Vali 2.

 

[rutter]

Does the Yggy having more "real" bits of resolution than most other DACs mean it is as capable or more with hi-res music, even though other DACs can be advertised as having up to 32 bits? Or is there more to the story that Schiit didn't deem relevant? Come on now, look at what is written on the product page rather than refer to something in a post somewhere.

How can this possibly be better than, say, the Arglebargle $15,000 DAC when this is so much less expensive?
Because the Arglebargle was most likely designed to the expectations of today, using commonly available parts in a super-fancy case, while we started with a clean sheet of paper.

But the Arglebargle has like twelve 32-bit DACs in it! Yours only has 21 bits! Hell, that’s not a full 24 bits even! What about my 24-bit recordings?
If your 24 bit recordings actually have 24 bits of resolution, we’ll eat a hat. And those "32-bit" DACs? Well, they have this measurement known as “equivalent number of bits.” This means, in English, how many bits of resolution they really have. And that number, for most of them, is about 19.5.

But it’s only 21 bits! I can’t get over that!
We can’t get over the fact that delta-sigma DACs are actually 2- to 5-bit designs. Different strokes for different folks.

My concern with distortion is an association with hearing and headphone damage. I would expect the Mimby to be worse in terms of sound quality.

 

[Currawong]

Hearing damage is connected with excessive exposure to high sound pressure levels. Whatever "Schiit didn't deem relevant" is something best asked to Jason or Mike directly. Everything else is too off-topic for this thread and gets very complex very fast, but ultimately comes down to technical accuracy versus a more euphoric sound.

 

[joeexp]

A 24 bit sample can have 2^24=16,777,216 discrete levels, Modi can only realize 2^16=65,536 discrete levels. Some information will inevitably be lost. Do you care? Do you hear? It is for you to decide.

The dynamic range of a 24-bit ADC is 144dB!
Cleary shows, that you haven't got a clue.

 

[gvl2016]

The dynamic range of a 24-bit ADC is 144dB!
Cleary shows, that you haven't got a clue.

Enlighten me, please.

 

[Guidostrunk]

Agreed. I don't even get into the debates anymore. I've been through some blind testing as well. It really woke me up to reality. I'm 43, and been in the construction field for the last 25 years, and my ears are probably as good as yours.
I failed miserably trying to distinguish between FLAC vs 24/192 vs 16/44. Same with Tidal vs Spotify. Paid $20 a month thinking I had the best streaming fidelity. Blind test cured that. Lol.

All this highres vs redbook, multibit vs DS talk has led me to buy a bunch of DACs of both types, and try some blind testing. With my 45 year-old metalhead ears, none of this makes that much difference. I spent a bunch of time comparing a DX7s vs Jotunheim internal 4490 vs Gumby on both headphones like DT880, GS2000e, LCD2 and on my magnepan speakers. I failed every ABX test trying to distinguish the DACs once level matching was done, and also couldn't tell 192/24 songs from the same source files downsampled to 44/16.
I think people can fool themselves into thinking one thing or another sounds better or worse in sighted listening tests either because they want to like multibit, or because they saw some AP measurements. For me, differences between different transducers and amps are order of magnitude higher than the DAC or high res file choice.

 

[BrotherKathos]

Agreed. I don't even get into the debates anymore. I've been through some blind testing as well. It really woke me up to reality. I'm 43, and been in the construction field for the last 25 years, and my ears are probably as good as yours.
I failed miserably trying to distinguish between FLAC vs 24/192 vs 16/44. Same with Tidal vs Spotify. Paid $20 a month thinking I had the best streaming fidelity. Blind test cured that. Lol.

You might have some hearing loss from the construction job, or the files you listened to in high res vs standard were not mastered well. I can absolutely hear a difference between some high res and regular files. One in particular is Werewolves of London by Warren Zevon. I have both CD and High res flac, and there is an immediately noticeable change in the bass line of the track between the two. The high res file has way more separation in the bass line where the reg bleeds together. I had a friend test be by playing both files randomly while i just had my eyes closed. He would play the reg file a couple of times in a row then switch to high res and then back again. I could pick out the high res one every time since I was focused on that one specific part of the bass line. Very easy. High res makes a difference if you have the ears and equipment to benefit from it.

 

[jimmers]

... I can absolutely hear a difference between some high res and regular files. One in particular is Werewolves of London by Warren Zevon. I have both CD and High res flac, and there is an immediately noticeable change in the bass line of the track between the two. The high res file has way more separation in the bass line where the reg bleeds together. I had a friend test be by playing both files randomly while i just had my eyes closed. He would play the reg file a couple of times in a row then switch to high res and then back again. I could pick out the high res one every time since I was focused on that one specific part of the bass line. Very easy. High res makes a difference if you have the ears and equipment to benefit from it.

Did you try downsampling the high res version (Rhino 192kHz-24bit?) and comparing it to the high res so you could be sure they used the same master?

 

[BrotherKathos]

Did you try downsampling the high res version (Rhino 192kHz-24bit?) and comparing it to the high res so you could be sure they used the same master?

No, so as you suspect it could be a difference in the master and not the resolution that I'm hearing. I just wanted to point out I do hear a difference in some high res vs standard files, for whatever reason that may be.

 

[438555]

No, so as you suspect it could be a difference in the master and not the resolution that I'm hearing. I just wanted to point out I do hear a difference in some high res vs standard files, for whatever reason that may be.

Yes, that’s been my experience. It’s easy to tell my old CDs apart from hires purchases, but they are different masters. When I downsample the hires tracks, I can’t tell them apart

 

[BrotherKathos]

Yes, that’s been my experience. It’s easy to tell my old CDs apart from hires purchases, but they are different masters. When I downsample the hires tracks, I can’t tell them apart

I think the lower res files should in theory only have less dynamic range, making them sound more like you have the night mode on your stereo vs actual content perceived. The best way to test is to listen to them at very low volumes and try to see if parts of the song drop out of audible range with the high res vs low res at the same volume setting.