Jump to content
Sign in to follow this  
kevin1969

How Does Up-Sampling Work?

Recommended Posts

I have a long history in telecom so I understand the concept of sampling an analog signal and turning it into a digital stream. Perfect example is the early 64K channels in T1's for voice (sample 2x highest frequency and create 8 bit word or each sample (8000Hz * 8 bits = 64K channel).

 

And obviously a digital bit-stream must be sampled at a certain rate to accurately turn it back into analog. But I am confused as to what and how up-sampling works and how it could even create a better quality analog signal from the original digital signal.

Share this post


Link to post
Share on other sites
But I am confused as to what and how up-sampling works and how it could even create a better quality analog signal from the original digital signal.

Upsampling of course can't produce a better quality file , because the quality is not getting better if you take everything times 2 or 4 and divide it later again by 2 or 4. But it definitely can help , providing a file with a higher sampling frequency, which leaves more space between the highest parts of the sampled signal and the critical point where - due to the nature of decoding - the limit of usable area ends. So the decoding process can be managed with not so steeply rising filters, which diminishes the phase distortion of the signal. In other words you don't gain quality at producing the higher sampled file but in decoding a file in a softer manner.

Share this post


Link to post
Share on other sites

The main benefit of up-sampling is that it simplifies the D/A process. Let's start at the output. The DAC will always have to convert a stream of digital data into an analog signal at the end of the conversion chain. Put simply, the more bits that analogue output stage (i.e. filter) has to work with, the easier it is to have that happen, and hence to produce a high-quality result. With your background you will easily see that reconstructing an analogue signal via a filter from a 44.1/16 data stream is more challenging to design and implement than doing the same from a 192/24 stream, for example. Given that it's much easier and cheaper to do things in the digital domain (especially inside a single IC at the design stage), it naturally follows that DAC designers will want to up-sample on the input in order ease the life of the analogue filter output stage. In short, those output samples will be "filled-in" anyway by the analogue filter, and these days helping this process by adding more samples in the digital domain is cheaper and easier than in the analogue one; part counts are lower, filter design is easier, and the impacts of the filter on the signal's phase and frequency response easier to manage.

 

Now, there are of course many ways to achieve that up-sampling. The simplest is just to double the input rate and stick the average of two successive input sample in between the source ones to create the intermediate value. However, that's also not a very good solution, because it takes no account of the rate of change of the signal pre- or post-sample. Consequently, a range of other digital techniques have evolved, all striving to do a better job of curve-fitting, if you will, the digital samples to what the original analogue input signal must have looked like before being digitized.

 

Hope this helps?

Share this post


Link to post
Share on other sites

Ah ok. This makes sense. So based on the DAC I have (Bifrost UBER analog 24/192) is there any advantage to setting my Audirvana software on my MacMini to upsample?

Share this post


Link to post
Share on other sites

I'm not familiar with that DAC so don't know to what degree it's chip set over-samples. Try it both ways and see which you prefer, because I'm guessing that whether you over-sample in Audirvana or allow the DAC to do it all, will change the sound signature. Personally, I use HQPlayer and up-sample/convert from Red Book to 2xDSD because I prefer the resulting sound in my system to just feeding the DAC 44.1/16 PCM. Software players these days have some very good up-sampling algorithms, and doing it there, all other things being equal, can produce great results.

Share this post


Link to post
Share on other sites
Ah ok. This makes sense. So based on the DAC I have (Bifrost UBER analog 24/192) is there any advantage to setting my Audirvana software on my MacMini to upsample?

 

The bottom line is what sounds best to you. It is usually an easy thing to try (i.e. upsample) with the various software players like A+.

 

To add to what has already been said, a DAC will generally have a sweet spot. Internally, most DAC's are doing some type of upsampling, etc. So why not have your music server do the work instead given it (music server) has much more compute resources than the DAC? So basically you are feeding the DAC what it wants (or fairly close) so that it has to do the least amount of processing. In other words, you are making the DAC work less.

 

Years ago, the holy grail was being a "bit perfect" data feeding the DAC until it was realized that the DAC was manipulating the data. As such, you don't really see that much on "bit perfect" anymore. I use HQ Player to upsample/convert everything to 2x DSD.


Eric


Ubuntu Studio Linux box (i7-9700, 8 cores, 16GB RAM, Intel X520-DA1 NIC, Roon, HQP) > fiber optic > MikroTik CRS305-1G-4S+ > fiber optic > opticalRendu (HQP NAA) > Holo Cyan (DSD version) > Nord One UP NC500MB mono blocks > Klipsch La Scala — digital volume control with HQP via Roon client

Share this post


Link to post
Share on other sites

1. Here how work oversampling How Convert Sample Rate. Oversampling

 

2. Oversampling don't improve audio quality as itself.

 

3. Any DAC contains analog filter for restoring analog signal from digital.

 

Analog filter should suppress artefacts of digital signal (frequency range [0 ... sample rate/2] mirrored to [sample rate / 2 ... sample rate] range).

 

More sample rate - analog filter possibly provide more suppression (at higher frequencies).

With higher sample rates these better suppressing area become available.

 

Analog filter has low growth of supression comparing digital.

 

In DAC (or into PC software or both) can be applied oversampling and additional digital filtration.

 

It allow decrease artefacts level in frequency range of low supression of analog filter.


AuI ConverteR 48x44 - HD audio converter/optimizer for DAC of high resolution files

ISO, DSF, DFF (1-bit/D64/128/256/512/1024), wav, flac, aiff, alac,  safe CD ripper to PCM/DSF,

Seamless Album Conversion, AIFF, WAV, FLAC, DSF metadata editor, Mac & Windows
Offline conversion save energy and nature

Share this post


Link to post
Share on other sites
Ah ok. This makes sense. So based on the DAC I have (Bifrost UBER analog 24/192) is there any advantage to setting my Audirvana software on my MacMini to upsample?

 

Yes. I've owned the Bifrost and heard the difference. Miska has also mentioned the chip in the Bifrost benefits nicely from software oversampling.

 

Go ahead and try it. If you don't like it you can just stop. :)


One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> wi-fi to router -> EtherREGEN -> microRendu -> USPCB -> ISO Regen (powered by LPS-1) -> USPCB -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Share this post


Link to post
Share on other sites

As rule, better to maximum sample rate of your DAC.

 

In general, need found "best sounding" sample rate.


AuI ConverteR 48x44 - HD audio converter/optimizer for DAC of high resolution files

ISO, DSF, DFF (1-bit/D64/128/256/512/1024), wav, flac, aiff, alac,  safe CD ripper to PCM/DSF,

Seamless Album Conversion, AIFF, WAV, FLAC, DSF metadata editor, Mac & Windows
Offline conversion save energy and nature

Share this post


Link to post
Share on other sites
As rule, better to maximum sample rate of your DAC.

 

In general, need found "best sounding" sample rate.

 

Do you mean to just set the Audirvana software to the maximum rate of the DAC? For Bifrost that would be 24/192.

Share this post


Link to post
Share on other sites
Do you mean to just set the Audirvana software to the maximum rate of the DAC? For Bifrost that would be 24/192.

 

A+ has radio button selections for max rate oversampling and powers of 2 oversampling. (For a CD rip at 44.1KHz, the latter would oversample to 176.4KHz, the former to 192KHz.) It's very easy to switch back and forth, so just try each for a while and see if there's one you like better than the other.


One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> wi-fi to router -> EtherREGEN -> microRendu -> USPCB -> ISO Regen (powered by LPS-1) -> USPCB -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  



×
×
  • Create New...