Is upsampling a good thing or does it introduce artifacts?

Hello and good Sunday morning to you. Lately, I have disabled all upsampling of music. It seems that I either get a slight lateral move, no change at all or a degredation, mixed in with unwanted artifacts whenever I try to upsample. Over the course of the last few months, I do not recall ever hearing any "Wow!" moments or even a, "Hey that does sound better!" moment.

I had always thought that upsampling was a good thing and offers you more to the music. My experience, albeit very limited at this point, has told me otherwise.

Now, I am not talking about video here. Strictly 2 channel audio. Again, my experience might as well be considered a rookie with this, as I have never really futzed around with it much, until recently.

That said, is there a certain way to do it? If I upsample, is/are there other parameters that would need to be changed in order to bring out the best of the upsample? Am I trying to upsample too little or do you go from whatever the current feed is to 192?

Like I said, I am a rookie at this kind of stuff and reading the Wikipedia page is basically a bunch of math equations that do not speak to me in layman's terms.

This thread is for a discussion about this specific subject. One that I'd like to learn about, if it actually can increase the end result as to what hits these ears.



The above images were taken from the Lumin app. I turned back on the resampling to show you all what the options are, took the screenshots and turned the upsampling (or re-sampling, as they call it) back off.

Please allow me to pick your brains and experience on this one. I'd like to improve, if possible, but it seems as if my experimentation thus far hasn't been a benefit.

One final question on this. At the very bottom, it has an option for Output Bit Depth. When experimenting, I usually set this to 24bit. My theory was that you would want the output to be maximized when upsampling. Is this where I messed up? Please advise and thanks.

~ In search of accurate reproduction of music. Real sound is my reference and while perfection may not be attainable? If I chase it, I might just catch excellence. ~


  • skipshot12
    skipshot12 Posts: 1,091
    Good question, can’t wait to read the consensus.

    Personal observation, from an Esoteric player years ago, I never heard a difference when upsampling.
  • SCompRacer
    SCompRacer Posts: 8,408
    edited April 2023
    You will find folks will speak in authoritative absolutes either way about oversampling/filtering. You have camps for and against. Best suggestion is try it for yourself, hear what happens and pick your camp. :)

    Results also depend on your DAC architecture. For example, if your DAC converts DSD to PCM, sending it Native DSD may not be the best choice. You can also try it and hear what happens. My DAC has an R-2R ladder and a 1bit DSD. If I select DSD direct, native DSD goes to the 1bit DSD unmolested. If I turn DSD direct off, the DAC converts DSD to PCM, then I have a choice of PCM direct to R-2R ladder or direct OFF and put it through the DAC's filters. Better, pricey DAC's have better filters so sometimes best to have a better device doing the upsampling/filters and just let the inexpensive DAC play it.

    I use Roon DSP to upsample files to Native DSD512 with 7th Order CLANS (Closed Loop Analysis of Noise Shapers) and smooth linear phase to my DAC. This conversion is CPU intensive. Roon shows the Processing Speed that alerts you to how hard the CPU is working. Anything above 1.3 is good, below that and you get dropouts, stuttering or no sound at all. It's best to work up slowly, DSD64, DSD128, etc. There are also camps that say Roon DSP sucks and HQPlayer (which will work with Roon) is the only way to go.

    The other option is upsample with PCM. Many folks cannot tell the difference between DSD512 and PCM 768kHz. So, if your DAC converts DSD to PCM, just send it PCM. PCM upsampling is also easier on the CPU. (Obviously, you'll need to know your DAC specs for max DSD/PCM rates and max rates your device will upsample to).

    Currently my server is a Signature Sonore i7 running Roon Rock (Linux) on NVME SSD with 6TB spinning music drive. Server and USB card have separate quiet linear power supplies. SOtM USB out to DDC (has USB isolation and temp controlled clocks) I2S out to DAC over HDMI.

    I just built a PC that will handle the toughest filters that HQPlayer has to offer. I haven't tried that yet. Considering installing a SOtM ethernet card to send music via ethernet cable to DAC.

    EDIT: Oh, regarding bit depth. 24 bit provides a wider dynamic range and lower noise floor over 16 bit.

    Post edited by SCompRacer on
    Salk SoundScape 8's * Audio Research Reference 3 * Bottlehead Eros Phono * Park's Audio Budgie SUT * Krell KSA-250 * Harmonic Technology Pro 9+ * Signature Series Sonore Music Server w/Deux PS * Roon * Gustard R26 DAC / Singxer SU-6 DDC * Heavy Plinth Lenco L75 Idler Drive * AA MG-1 Linear Air Bearing Arm * AT33PTG/II & Denon 103R * Richard Gray 600S * NHT B-12d subs * GIK Acoustic Treatments * Sennheiser HD650 *
  • Jazzhead
    Jazzhead Posts: 533
    I'm not streaming, but my Cambridge 840C is an upsampling player that also allows one to select sampling frequency, word length, and dither "on" or "off" etc. Doug Schroeder's review of it explains upsampling and such digital processing in layman's terms (much of it still goes over my head, lol). In the end, I agree with Schroeder that this player, using balanced interconnects, sounds best (realism) using the "pass through" mode (no upsampling):
  • heiney9
    heiney9 Posts: 25,110
    I don't have a lot of hands on experience with choosing up sampling or not. But I have done some reading in the past and it seems it can be recording dependent. Some sound good up sampled and some don't.

    I have a personal issue with paralysis of analysis so I need to keep my settings and choices to a minimum or I go crazy analyzing everything to the nth degree. That's why I was thinking (just thinking) of getting rid of tubes. The combinations based on the tubes I own are almost infinite. I have been able to "set it and forget it" with far.


    "Appreciation of audio is a completely subjective human experience. Measurements can provide a measure of insight, but are no substitute for human judgment. Why are we looking to reduce a subjective experience to objective criteria anyway? The subtleties of music and audio reproduction are for those who appreciate it. Differentiation by numbers is for those who do not".--Nelson Pass Pass Labs XA25 | EE Avant Pre | EE Mini Max Supreme DAC | MIT Shotgun S1 | Pangea AC14SE MKII | Legend L600 | BlueSound Node 3 - Tubes add soul!
  • treitz3
    treitz3 Posts: 18,685
    So, no further thoughts or observations with changing the sample rates?

    ~ In search of accurate reproduction of music. Real sound is my reference and while perfection may not be attainable? If I chase it, I might just catch excellence. ~
  • F1nut
    F1nut Posts: 50,098
    You over think things.
    Political Correctness'.........defined

    "A doctrine fostered by a delusional, illogical minority and rabidly promoted by an unscrupulous mainstream media, which holds forth the proposition that it is entirely possible to pick up a t-u-r-d by the clean end."

    President of Club Polk

  • Emlyn
    Emlyn Posts: 4,430
    I listen to digital upsampled at DSD128 most of the time because that's the default on my gear. That does involve upsampling and upsampling does produce artifacts, from everything I've read, but those are moved so high above hearing level through computing power and then filtered so that they have no negative effect on the sound that can be heard.

    The result is a smoother more natural tone to my ears on tracks that are native 16 bit PCM audio. Makes no difference with high resolution PCM or FLAC that I can hear because the bandwidth with 24 bit audio is very high. I can tell the difference by switching the upsampling feature on and off and ending up with a harsher and grainier sound on 16 bit tracks. It's similar but better than what I used to hear with HDCD on a high quality player.
  • VR3
    VR3 Posts: 28,320
    Keep in mind the lumin will only up sample to what your dac can play.

    The lampizator doesn't sample anything over 24 bit/96khz
    - Not Tom ::::::: Any system can play Diana Krall. Only the best can play Limp Bizkit.