[Joe Bloggs]

so are you trying to say Mike Moffat and crew are damn lucky, or outstanding mathematicians?

No, I'm saying they're either not actually preserving the original samples in the upsampled bitstream (as it does not make sense to do so) or compromising the quality of the upsampling by insisting on doing so.

 

[captblaze]

No, I'm saying they're either not actually preserving the original samples in the upsampled bitstream (as it does not make sense to do so) or compromising the quality of the upsampling by insisting on doing so.

 
theoretically speaking wouldn't any resampled stream introduce new artifacts that in turn would need an algorithm to restore the original sample? if there is a flaw in the original, I don't see how (depending on the flaw) resampling would make for a more accurate representation. unless the goal is to interpret the data and make corrections based on an individuals preference and not to pass along the original content?
 
I would guess that is why some makers choose a method they believe will provide the result they desire?

 

[oldson]

just bought the mojo today.
 
using solely as a dac in my desktop system.
 
its got me thinking about how good a 2qute must sound!

 

[Joe Bloggs]

theoretically speaking wouldn't any resampled stream introduce new artifacts that in turn would need an algorithm to restore the original sample? if there is a flaw in the original, I don't see how (depending on the flaw) resampling would make for a more accurate representation. unless the goal is to interpret the data and make corrections based on an individuals preference and not to pass along the original content?

I would guess that is why some makers choose a method they believe will provide the result they desire?

It sounds like a good idea to preserve the original samples, but mathematically impossible (short of computing using an infinite length filter) unless you deliberately bend the interpolated sample points to fit the original sample points--this bending further compromises a non-ideal resampling rather than making it ideal. That's what I'm trying to say.

The mathematics behind this are admittedly beyond me to fully explain (as if strewing the contents of a page like this on head-fi would make any sense anyway), but the Adobe Audition example showed that Adobe knows how to implement a "preserves original samples" filter but chooses to only do so for real-time visualization rather than actual audio application. We may deduce that Adobe also regards "preserves original samples" would do more harm than good for actual listening.

 

[captblaze]

It sounds like a good idea to preserve the original samples, but mathematically impossible (short of computing using an infinite length filter) unless you deliberately bend the interpolated sample points to fit the original sample points--this bending further compromises a non-ideal resampling rather than making it ideal. That's what I'm trying to say.

The mathematics behind this are admittedly beyond me to fully explain (as if strewing the contents of a page like this on head-fi would make any sense anyway), but the Adobe Audition example showed that Adobe knows how to implement a "preserves original samples" filter but chooses to only do so for real-time visualization rather than actual audio application. We may deduce that Adobe also regards "preserves original samples" would do more harm than good for actual listening.

 
advanced mathematics is what chased me from a degree in computer engineering. so thank you for allowing my brain to remain in tact. I am interested in how digital audio is reproduced though. it seems that Chord has their method, as well as the other heavy hitters in the industry.
 
I do also understand the concept  you put forth regarding an infinite length filter to allow the original sample to remain original. I just have a hard time getting my head around how resampling would provide a more accurate representation of the source.
 
it is definitely something I would like to further research and if you can direct me to some material regarding the science I would appreciate that

 

[pr0b3r]

Hi guys! I'm kinda new to this hobby. I mainly use my mojo with my x3ii as my go to gear recently. I happen to own a Q1 also and tried to add it to the stack. Just a question, does this render the mojo useless or less useful? I'm thinking it's better not to add the Q1 here but I'm not entirely sure what to make of this setup. Thanks for any suggestions you might have.

 

[warrior1975]

You will get all sorts of different answers, but only you can answer that question for yourself. Did you like what you heard by adding the Q1? If so, then do it and be happy. If not, than just use the Mojo. Most people here would probably tell you not to add the Q1 as it will add distortion, but if you prefer that sound, go for it.

 

[pr0b3r]

You will get all sorts of different answers, but only you can answer that question for yourself. Did you like what you heard by adding the Q1? If so, then do it and be happy. If not, than just use the Mojo. Most people here would probably tell you not to add the Q1 as it will add distortion, but if you prefer that sound, go for it.

Thanks man. I was just wondering if anyone has an elaborate explanation abut the technical differences between the two setups. Anyway, I know you're right. It still depends on my preference whether I liked a more than b. But I guess there's no harm in asking. Thanks again.

 

[rkt31]

@Joe Bloggs, thanks for the efforts. now I understand what you want to say. it is not possible to interpolate to fit original data points without using infinite samples right . if you use finite sample then interpolated wave won't go through the original points obviously no dac can have that much processing power . so how schiit is doing that is a question which can be answered by the designer of yggdrasil . on the other hand Rob has always said the a bandwidth limited signal can be re created perfectly by using infinite tap length but he never said that the original signals will be preserved as it is not possible with the processing power on hand. however by using a algorithm (wta) and high tap length through fpgas the wav form can be match much closer to the original . right ?

 

[Mython]

LOL - I don't claim to grasp all the technicalities, but it seems, to my limited understanding, that the original ADC isn't accurate, anyway, so unless one improves both the recording/mastering ADC and the playback DAC, even an 'infinite'-tap or 'infinite'-interpolation DAC will only address half of the problem of reconstructing an analogue waveform, from a digital recording, in a truly accurate manner.
 
Am I missing something?
 
Incidentally, this is veering into Sound Science Forum territory...

 

[rkt31]

one more thing do all expensive r2r DACs are nos ? if not then what algorithm they use to updample ? does that mean the these expensive r2r DACs could use a better upsample algorithm of higher tap length and sound still better or there is a technical problem in r2r dac and they need to have original data point format unlike DS DACs which work by converting the data into a pulse ?

 

[Joe Bloggs]

@Joe Bloggs, thanks for the efforts. now I understand what you want to say. it is not possible to interpolate to fit original data points without using infinite samples right . if you use finite sample then interpolated wave won't go through the original points obviously no dac can have that much processing power . so how schiit is doing that is a question which can be answered by the designer of yggdrasil . on the other hand Rob has always said the a bandwidth limited signal can be re created perfectly by using infinite tap length but he never said that the original signals will be preserved as it is not possible with the processing power on hand. however by using a algorithm (wta) and high tap length through fpgas the wav form can be match much closer to the original . right ?

The longer your reconstruction filter the closer the upsampled bitstream can be to matching the original data points (where provided). But it will always be "close, but no cigar", and I have strong reason to suspect that any attempt to postprocess the calculated output to match the original data points (where provided) would then be doomed to further distort the signal instead of making it better.

LOL - I don't claim to grasp all the technicalities, but it seems, to my limited understanding, that the original ADC isn't accurate, anyway, so unless one improves both the recording/mastering ADC and the playback DAC, even an 'infinite'-tap or 'infinite'-interpolation DAC will only address half of the problem of reconstructing an analogue waveform, from a digital recording, in a truly accurate manner.

Am I missing something?

Indeed. And never forget that we listen to music with our ears, not our eyes. An upsampled bitstream that passes through all the original points may LOOK better on a graph, but would introduce distortion in terms of aberrant frequencies and/or frequencies beyond the Nyquist of the original signal (unless you had infinite computer power; see above)

On the other hand, if we look at the goal of Redbook upsampling as to try to guess at the ultrasonics that could not be recorded by the original CD, then the technology to look towards should be blind spectral band replication, not some magic way of connecting the dots:
https://www.google.com/search?q=blind+spectral+band+replication

Incidentally, this is veering into Sound Science Forum territory...

You can see how the conversation got started in here. But I'm fine with a mod forking these discussions over to the Science forum.

 

[Joe Bloggs]

one more thing do all expensive r2r DACs are nos ? if not then what algorithm they use to updample ? does that mean the these expensive r2r DACs could use a better upsample algorithm of higher tap length and sound still better or there is a technical problem in r2r dac and they need to have original data point format unlike DS DACs which work by converting the data into a pulse ?

There's nothing preventing an R2R DAC from being coupled with any kind of upsampling algorithm. They wouldn't throw a fit if the upsampled bitstream doesn't quite pass through the original dots.

 

[x RELIC x]

Actually I think I may have fallen into a trap here.

Spoiler

Preserving the original sample points when upsampling, in a non-cheating workflow, would imply having a perfect brickwall filter of infinite length as the interpolation filter. Because otherwise there's simply no way the resampled audio can naturally pass through the original sample points.

IF (and that's a big IF) the megaburrito filter actually does what it claims, short of pulling an infinite computing miracle, it can only do this by bending the interpolated points toward the original sample points in a non-optimal manner.

This can be demonstrated using Audacity and Adobe Audition:

1. Here, a Dirac impulse drawn on silence sampled at 44.1kHz, shown in Audacity which does not perform any interpolation in its display (or, strictly speaking, "linear interpolation"):


As you may know, a Dirac impulse contains all frequencies theoretically encodable by the format, so this is a real test of resampling.

2. Here, the same impulse as interpolated by Audition in display. It has to show the original sample points where they are, so there's your megaburrito filter! h34r:

See how the interpolated curve passes through all the original points (i.e. all on the middle line except for the impulse in the middle)? Is this really the ideal resampling we're all looking for?

3. I perform offline resampling of the track using the following settings which I believe to be optimal for the present application. Feel free to correct me:


4. The result. Note that the original sample points are NOT on the middle line anymore. (they're close, really quite close, but no cigar.) If keeping the original samples were desirable, why would Audition do it for the real-time visualization but discard it for (supposedly highest quality) offline rendering?


Note also that the pre / post impulse waveforms undulate for further before and after the impulse. You may know this to be a sign of a resampling filter that's closer to ideal, as a mathematically ideal filter would cause the impulse to pre-ring and post-ring for infinite time. (not that this is a bad thing, because proper music samples ought to have all frequencies so close to the transition band filtered out during recording analog-digital conversion.)

To summarize, the fact that
1. Adobe Audition uses a "megaburrito" filter for real time visualization but not high quality offline conversion, and
2. The offline conversion shows characteristics of a higher quality conversion despite not looking like a "megaburrito" filter

Speaks volumes about the actual desirability (or not) of preserving the original samples in integer-multiple sample rate conversion.

Then again, AFAIK no one has actually been able to tap the internal upsampled digital stream of a Schiit DAC for analysis, so who knows if it actually preserves the original samples? :rolleyes:

A Sound Science trap!

 

[warrior1975]

Interesting info, from what very little I understand.