Upsampling

[Mike Rubin]

I have been purchasing hi-res tracks from sites on the web and, of course, have seen that, at some of the sites, the price goes up as the resolution increases.  I always have assumed that the pricing is intended to exploit the consumer's expectation that higher resolution equates with better sound, but I suppose it's possible that there's more work in bringing the higher res files to market.

 

I recently realized that my dbPowerAmp Converter can upsample all the way up to 384khz just by my pressing a button.  

 

So that begs the question of whether the hi-res sellers are doing something substantial to create these variously-priced versions of the same titles or just running the same sources through upsampling software like I can do at my home computer.  I am not so cynical as to assume the worst, but am genuinely curious about how commercial hi-res files are created.  What do the sellers do that we can't do at home with a CD quality source and upsampling software?

[firedog]

1 hour ago, Mike Rubin said:

I have been purchasing hi-res tracks from sites on the web and, of course, have seen that, at some of the sites, the price goes up as the resolution increases.  I always have assumed that the pricing is intended to exploit the consumer's expectation that higher resolution equates with better sound, but I suppose it's possible that there's more work in bringing the higher res files to market.

 

I recently realized that my dbPowerAmp Converter can upsample all the way up to 384khz just by my pressing a button.  

 

So that begs the question of whether the hi-res sellers are doing something substantial to create these variously-priced versions of the same titles or just running the same sources through upsampling software like I can do at my home computer.  I am not so cynical as to assume the worst, but am genuinely curious about how commercial hi-res files are created.  What do the sellers do that we can't do at home with a CD quality source and upsampling software?

 

It depends. For instance, on some releases they have a 24/192 master and downsample it to make a 24/96  release from that - and may or may not also sell it in 24/192. On others (not very frequently) there are 2 completely  separate masters.  On others the 24/192  is a mix frequency, but the master for release is created in 24/96-there's no actual 24/192 master created or available for sale.  And then sometimes they create a "sale" version in 24/192 just by upsampling a 24/96 master. There are lots of possible variations of these scenarios. We can then argue all day if 24/96 and 24/192 from the same master sound different.

 

It may also be system dependent, which makes the "argument"  a total waste of time. 
 

Yes, you can upsample  at home to whatever rate you like. Lots of people are doing that with programs like Roon or HQP because they think it sounds better on playback. Some people think the upsamping/filtering in HQP is superior sounding to that available elsewhere. Some people think that if your DAC receives a high sample rate, the result sounds better than if you have the DAC upsample on the fly. 

Upsampling at home is essentially no different than what the labels are doing if/when they create a higher res version just by upsampling. Whether/if that sounds different/better is a matter of debate. 
 

Bottom line: I wouldn't worry about it too much. The quality of the master is way more important than the rate of the file. Buy what fits your budget, or "on principle" buy the sample rate of the master (if it's available and if it's known- some labels or album booklets tell you the original recording rate), or just decide to always buy the highest rate available for sale. It's up to you. Do you think upsampling to higher rates improves the sound? If so, then do it, or buy the higher available rates. 

[Mike Rubin]

Thank you very much.  That's exactly what I was seeking. 

 

I couldn't agree more that the original source is the primary determinant of sound quality, but I do hear differences between the same files when resolution increases.  Often, however, that seems most affected by volume level differences between the two versions under comparison.  Are those differences improvements?  I can't tell without doing more analytical listening, really, and it is too much work to try to live on the differences that seem apparent on casual listening. 

 

It probably is a topic for another day, but I have had consistently good sound from DSD files.  Part of that is self-fulfilling, because I tend to buy in DSD format stuff that I already own that has decent sound, but newer native DSD files routinely have sounded great.  However, they tend to present pedestrian music by unknown artists and I wouldn't want too steady a diet of that stuff for listening rather than system evaluation. 

 

 

[firedog]

51 minutes ago, Mike Rubin said:

 

 

It probably is a topic for another day, but I have had consistently good sound from DSD files.  Part of that is self-fulfilling, because I tend to buy in DSD format stuff that I already own that has decent sound, but newer native DSD files routinely have sounded great.  However, they tend to present pedestrian music by unknown artists and I wouldn't want too steady a diet of that stuff for listening rather than system evaluation. 

 

 

 The people who record in DSD are almost certainly going to be more oriented towards making a good recording than most others. 

[Peter Hyatt]

With Chord Dac, I was advised to not use software up sampling; let the DAC do its work as the original file is the important element. 

[peter73]

On 5/13/2019 at 6:57 PM, Mike Rubin said:

... I couldn't agree more that the original source is the primary determinant of sound quality, but I do hear differences between the same files when resolution increases.  Often, however, that seems most affected by volume level differences between the two versions under comparison...

 

Volume level doesn't change in simple bit depth or sample rate conversion. If you are doing the conversion yourself, then you are doing something wrong. If there is a difference in volume level between two versions of the same album that are supposed to differ in sample rate/bit-depth only, then someone is lying.