What sample rate are you using?

[Lambda]

Most of the newer DACs internally use sample rates over 44,1khz and up/over sample material to there native sample rate.
Do you send the native source samplerate (for example 44.1k) to your DAC and relay on its internal up/over sample algorithms and filters
Or do you up/over sample in software (Player or operating system or drivers) and send higher sample rate (for example 192k) from your PC to the DAC.


You can make 2 votes, so pleases let us know what rate you upsampe to if you do.
And leaf a comment on what algorithms,filter, dithering you use.

Thanks.

 

[Vini darko]

I'm using a old android phone for streaming via usb. Amazon hd reports device as 24bit/192khz capable and says it's playing appropriate rates. However I have no idea if android is messing with it before usb output. Also can't initialize the dac so it just sits at 24bit/192khz for everthing I stream. For native files using UPP I use the upsampling mode as "bit perfect" mode creates a lot of noise for some reason. So I'm playing mp3's at 32bit/384khz according to the dac .

 

[Lambda]

However I have no idea if android is messing with it before usb output.

I don't have android audio knowledge.

But i have some Linux audio knowledge.
Its looks like the default config its not developed with best quality in mind but with high compatibility.
ALSA and Pulseaudio can do resampling, mixing and volume control
https://linuxreviews.org/PulseAudio

I'm sure Android is also doing resampling by default.

So I'm playing mp3's at 32bit/384khz according to the dac

Thanks for the info.

i try to up sample everything to to 96K if its not alredy 96K and from there i go "bit perfect" to the DAC using Jack and ALSA

 

[AnalogSteph]

I am using driver-level upsampling here - system sample rate 44.1 kHz, hardware sample rate 96 kHz. The hardware (Asus Xonar D1) is known to lose a few dB of dynamic range at 44.1 kHz. This way, the OS-level resampler is not required for typical CD quality material, YouTube etc. It's all a bit 2010 era but then so is this entire computer.

 

[krabapple]

You'd have to add a new category for me --

I feed my Denon AVR whatever the native sample rate of the file is. This ranges from 44 to 192 kHz. (this is one of your options)

But because I use Audyssey, everything is resampled to 48 kHz . This can mean upsampling (also an option) OR downsampling (not an option)

 

[threni]

Why isn't 44.1kHz an option? I'm playing .flac files at 44.1kHz to an E30 which is displaying 44.1kHz. I thought that was sort of normal, and that the other rates were silly. If the DAC wants the upsample internally that's fine, but why make my Pi do it?

Ideally this mpd.conf is correct for "leave it as it is and just shove it down the USB port":

audio_output {
     type            "alsa"
     name            "Direct"
     device          "hw:1,0"
     auto_resample   "no"
     auto_format     "no"
}

 

[sejarzo]

I now run at native sample rate 99% of the time. For a couple of months, I was using only 96k impulse files for headphone FR correction, so I resampled all non-96k files using SRC in Foobar to 96k ahead of the convolver plug-in. I now have impulse wav's at 44.1/48/96k so I swap the wav rather than resample. A bit of a PITA, admittedly.

 

[sejarzo]

Why isn't 44.1kHz an option? I'm playing .flac files at 44.1kHz to an E30 which is displaying 44.1kHz. I thought that was sort of normal, and that the other rates were silly. ...

I believe the OP is assuming that most folks are starting with CD-quality files so sending 44.1k to the DAC is neither up- or re- sampling.

Less than 5% of my library is at 88.2k or 96k. Tidal master quality runs at either 88.2 or 96k.

 

[Lambda]

Why isn't 44.1kHz an option? I'm playing .flac files at 44.1kHz to an E30 which is displaying 44.1kHz.

in this case "No re-sampling native samplerate to the DAC"

. If the DAC wants the upsample internally that's fine, but why make my Pi do it?

Because the Pi has more computation power and you have the freedom to chose what algorithms,filter, dithering you use.

I believe the OP is assuming that most folks are starting with CD-quality files so sending 44.1k to the DAC is neither up- or re- sampling.

Yes

 

[levimax]

I now run at native sample rate 99% of the time. For a couple of months, I was using only 96k impulse files for headphone FR correction, so I resampled all non-96k files using SRC in Foobar to 96k ahead of the convolver plug-in. I now have impulse wav's at 44.1/48/96k so I swap the wav rather than resample. A bit of a PITA, admittedly.

I currently resample everything in foobar to 96K just as you describe and for the same reason i.e. one FIR room correction file for all sources. Why did you stop doing it that way and go the more complicated route of switching correction files?

 

[Sal1950]

No idea why anyone would upsample? You can't make more musical information with magic dust, only possibly lose it. All my music in on my harddrive in it's native rate and sent to the DAC in bit perfect manner.. I've got hundreds of PCM/DSD 4.0, 5.1, Atmos, and Auro files which are all at better then 16/44, along with thousands of ripped Redbook. The only manipulations I get involved in is upsampling 2 channel sources to multichannel playback. Upsampling for ease of doing DSP is a different case, my AVR handles the chores out of my control.
YMMV

 

[Jimbob54]

I'm using a old android phone for streaming via usb. Amazon hd reports device as 24bit/192khz capable and says it's playing appropriate rates. However I have no idea if android is messing with it before usb output. Also can't initialize the dac so it just sits at 24bit/192khz for everthing I stream. For native files using UPP I use the upsampling mode as "bit perfect" mode creates a lot of noise for some reason. So I'm playing mp3's at 32bit/384khz according to the dac .

Highly likely android is resampling everything up or down to 48. Only a few apps can bypass this and send unfettered to a USB dac. USB audio player Pro is one that can. I don't believe amazon app is.

 

[levimax]

No idea why anyone would upsample? You can't make more musical information with magic dust, only possibly lose it. All my music in on my harddrive in it's native rate and sent to the DAC in bit perfect manner.. I've got hundreds of PCM/DSD 4.0, 5.1, Atmos, and Auro files which are all at better then 16/44, along with thousands of ripped Redbook. The only manipulations I get involved in is upsampling 2 channel sources to multichannel playback.
YMMV

I up sample / down sample to match multiple sources to one FIR room correction file to make things easier... nothing to do with sound quality.

 

[Thalis]

Since I am using Peace/EQ APO I don't upsample anything......... I set the Windows output to 24/44.1 on 2 DACs and 16/44.1 on one since it's max is only 16 bit. Sources are Tidal, Spotify and Youtube.

 

[Aerith Gainsborough]

System samplerate is set to 48KHz due to the piano VSTi I use and because I am too lazy to change it around.

Don't care either way for music, can't hear any difference between native, upsampled, downsampled whatever.

 

[Sal1950]

I up sample / down sample to match multiple sources to one FIR room correction file to make things easier... nothing to do with sound quality.

Yes, I edited my post for such situations. I was thinking of the folks just upsampling everything to higher rates or PCM to DSD in the belief they get better sound. I've had huge debates in other places on the subject.
It does sound exactly like they believe it should.

 

[Lambda]

I set the Windows output to 24/44.1 on 2 DACs and 16/44.1 on one since it's max is only 16 bit. Sources are Tidal, Spotify and Youtube.

but do you have windowes and Tidal, Spotify and Youtube volume at 100% all the time?

 

[Thalis]

but do you have windowes and Tidal, Spotify and Youtube volume at 100% all the time?

yes

 

[Lambda]

No idea why anyone would upsample?

Beaus the 98% of the DACs will upsample it anyways without telling you.
Some DACs/driver tell you about this and allow you to chose and algorithms,filter, dithering... method.

So up sampling is happening (in most cases) and you have the freedom to chose what algorithms,filter, dithering you use if you do it ins software and not in the DACs hardware DSP.

yes

Sadly im to lazy to do so and most of the time i don't do it

 

[sejarzo]

I currently resample everything in foobar to 96K just as you describe and for the same reason i.e. one FIR room correction file for all sources. Why did you stop doing it that way and go the more complicated route of switching correction files?

It's not for room correction, it's for headphone correction, and I switch headphones more often than sample rates anyway.

I am redoing my speaker rig to use minidsp/Dirac, but taking care of elderly parents for the past 6 months with no end in sight has me wondering when I will get that done.