Hi,
Some dac & dac hats using Rpi as a renderer and/or music library server have a non over sampling design.
Are there RaspBerry Pi distros with up/oversampling setup options ?
I assume most of the Pi music distros pass through the materials with their original samplings rate : Redbook or higher when existing : 88.4/96 K Hz and so on.
I would like to test up/oversampling on the fly with a Rpi distro as I don't use a pc to stream and 99% of my material is Redbook (44.1/48 KHz samplinng rate).
Many thanks
Some dac & dac hats using Rpi as a renderer and/or music library server have a non over sampling design.
Are there RaspBerry Pi distros with up/oversampling setup options ?
I assume most of the Pi music distros pass through the materials with their original samplings rate : Redbook or higher when existing : 88.4/96 K Hz and so on.
I would like to test up/oversampling on the fly with a Rpi distro as I don't use a pc to stream and 99% of my material is Redbook (44.1/48 KHz samplinng rate).
Many thanks
Sorry diyiggy I didn't know Ario myself. 
I meant Aroio.
ABACUS electronics - AroioOS - Streaming - Produkte
A German website but in the upper right corner is a Google Translation add-on button.
I meant Aroio.
ABACUS electronics - AroioOS - Streaming - Produkte
A German website but in the upper right corner is a Google Translation add-on button.
Why?
Hi CharlieLaub,
Well, 99% of my ripped library is 44.1 K Hz.
I have a NOS dac (with no low pass output filter) with a 45/49 MHz clock digital frontend and a dac board which is 192 K Hz speed proof and perhaps able of more. So I'd like to test upsampling playback. My clock can go to x8, the dac chip as well. My purpose is to be able to switch easily on the fly from a rate to an other to listen to wheter I like it or not.
My basic understanding is Audirvana and Sygnalyst can do that with upgraded and priority digital filter but I have just a Rpi to stream.
Is it a bad idea and why ?
Hi CharlieLaub,
Well, 99% of my ripped library is 44.1 K Hz.
I have a NOS dac (with no low pass output filter) with a 45/49 MHz clock digital frontend and a dac board which is 192 K Hz speed proof and perhaps able of more. So I'd like to test upsampling playback. My clock can go to x8, the dac chip as well. My purpose is to be able to switch easily on the fly from a rate to an other to listen to wheter I like it or not.
My basic understanding is Audirvana and Sygnalyst can do that with upgraded and priority digital filter but I have just a Rpi to stream.
Is it a bad idea and why ?
Last edited:
Hmmm, I see.. you (diyiggy) are really looking for software that does on the fly variable up-sampling. Not sure exactly how fast your changes need to be, but Gstreamer can do resampling to arbitrary rates and is available for the Pi. You would need to write some Gstreamer pipelines to do the resampling, as well as something (a script?) to kill and restart a new pipeline each time you want to change rates. That might take 0.5 seconds.
Hi.
I'm running synchronous resampling to 352k8 and 384k to bypass the
mediocre internal filters on my DAC nowadays.
I basically figured: Resampling is not ALWAYS a bad thing.
And it's not as trivial as it might look!
In a client/server audio environment you can usually choose between server
based or client based resampling.
I'm not sure if that also applies to streaming services like
Tidal/Quobuz etc. while being routed over LMS.
You might be forced to use client-based resampling.
When looking at a LMS/squeezelite setup, you can use either LMS
or squeezelite to execute a highest quality resampling task.
On LMS you'll find the 3-CPO plugin, on squeezelite it's an embedded feature.
Both resamplers are based on sox/libsoxr .
libsoxr is one of the best resampling libraries around. It of course
requires well chosen settings for best results.
The next question is where to execute the job.
On the server or on the client!?!?
I do it on a LMS (i5-NUC) server to keep the load from the RPI.
The RPI is extremely inefficient in running high performance DSP tasks.
(I did some benchmarking and wrote some articles)
Hint: The piCorePlayer project did improve the performance of the resamplers
not long ago.
Make sure you run the latest pCP or LMS on any platform!!!
And keep in mind. A low muscle NAS of course wouldn't do much
better than a RPI. Make sure the NAS offers a most recent LMS version!
Now. Once you've taken the decision what setup you prefer,
the next big challenge will be to figure out what parameters
to run your resampler with.
The parameters can have quite an impact on the result you'll be listening to. But that's another story.
Good luck.
I'm running synchronous resampling to 352k8 and 384k to bypass the
mediocre internal filters on my DAC nowadays.
I basically figured: Resampling is not ALWAYS a bad thing.
And it's not as trivial as it might look!
In a client/server audio environment you can usually choose between server
based or client based resampling.
I'm not sure if that also applies to streaming services like
Tidal/Quobuz etc. while being routed over LMS.
You might be forced to use client-based resampling.
When looking at a LMS/squeezelite setup, you can use either LMS
or squeezelite to execute a highest quality resampling task.
On LMS you'll find the 3-CPO plugin, on squeezelite it's an embedded feature.
Both resamplers are based on sox/libsoxr .
libsoxr is one of the best resampling libraries around. It of course
requires well chosen settings for best results.
The next question is where to execute the job.
On the server or on the client!?!?
I do it on a LMS (i5-NUC) server to keep the load from the RPI.
The RPI is extremely inefficient in running high performance DSP tasks.
(I did some benchmarking and wrote some articles)
Hint: The piCorePlayer project did improve the performance of the resamplers
not long ago.
Make sure you run the latest pCP or LMS on any platform!!!
And keep in mind. A low muscle NAS of course wouldn't do much
better than a RPI. Make sure the NAS offers a most recent LMS version!
Now. Once you've taken the decision what setup you prefer,
the next big challenge will be to figure out what parameters
to run your resampler with.
The parameters can have quite an impact on the result you'll be listening to. But that's another story.
Good luck.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.