Playback of vinyl records including realtime click removal

Dear Volumio Team,
I am a big vinyl fan and have a lot of records collected over the years. I thought about the possiblity to play back the records (some of them are old had have a lot of scratches and clicks…) via volumio and the software could remove the clicks and scratches in realtime while using the power of the rasberry.
Volumio could be the first audiophile product to feature real-time, non-destructive, click & pop removal on any vinyl LP record while using already available sophisticated algorithms to be able to detect, isolate and remove only unwanted noise separate from the music.
I tried it with the mac software “vinylstudio” and it worked perfect - but offline… - old scratched LP´s appear as nearly brand new…
I am sure, that this feature would help Volumio to find additional fans around the vinyl lovers worldwide!!
I would like to feed in my analog signal into any supported DAC which offers this feature.
Thank you for your great work
Greatings from Frankfurt Germany
Peter

Hi!
I guess that a “realtime” click/plop removal is not possible.
The algorythms depend on a whole file. The Algorythm takes the complete song and then searches for some pattern, like click’s (extreme short peaks) and plops (short peaks with flat tops). As long as this is live, the algorythm will not be able to “predict” if the signal will alter in a few ms or not.

But I have an Idea for this topic:
What if the ADC input of the DAC would be used to:

  1. record the input (vinyl playback)
  2. perform a “click/crack” removal with a tool like Audacity?
    -> This could be done with a “plugin-like”-method…

Anyway:
I will stay at my manual recording and manual editing of the recordings.
In my opinion the “algorythmic” “autoremoval” makes more damage to the file…
@volumio:
Anyway: A feature to make Wave/Flac recordings from the ADC would be nice!

Hi Machine2501,
many thanks for your comments and advise!
I agree - the “easiest” way (and available as well) of removing clicks/plops is to process the digitized audio file afterwards and for me, it works without too much degradation of the sound. But it´s time consuming and I do not want to store my vinyl music on disk. That is not my usecase…
I want to listen to vinyl and on the fly remove the “defects” during the listening session - how cool is that :slight_smile:
and there are systems out, which offer this - but unfortunately not in my budget range…


So it seems like the “problem” was already solved and of course, there must be a “lag” between the needle and the speaker - as the analog signal needs to be digitized (looseless) and processed - but I would not care about this. The result counts - Clear Audio - no matter how bad the condition of the vinyl record (of course there is also a limit).
Maybe the Raspberry PI is not powerful enough for this (nice) use-case - I can´t tell you.
And maybe the effort to integrate this functionality into Volumio is too high.
But I am still positive that this would attract an additional user group which is growing…

and you can hear the result in this video: https://www.youtube.com/watch?v=kTpsjlvlJ4I&t=132s

You would need to take into consideration the processing power required for this, as it wouldn’t be something a Raspberry Pi will handle in real-time, especially if the algorithm is set to a greater sensitivity and for a higher quality rendering. To get an idea, you mentioned that you processed a file offline and then played it back. I’m willing to bet it took longer to process than the file’s actual playback time. I know it certainly is the case for my PC, running Izotope RX7 with an i9 processor and that’s only marginally slower than realtime but still too slow to smoothly render during playback.

De-clicking needs to read ahead and process a fair chunk of data in order to be able to keep up with the audio it ends up playing out. Taking that SweetVinyl example you’ve put forward, that money is going toward both the R&D cost of the algorithm but also the FPGA that’s likely to be at the heart of its processing capabilities. It will have some serious grunt compared to anything CPU based, be it x86, AMD64 or ARM.

Hi Metalio,
many thanks for your comments, information and feedback.
I think you lost your bet :wink:

My “offline” click removal was done with the VinylStudio Software and an old late 2009 Imac. https://www.alpinesoft.co.uk/

After I recorded to flac the process of “de-clicking” took a very short time - 2 minutes max. and the result was very good - the clicks disappeared by 90%.
Maybe it started with analytics in parallel to the recording - I am not sure.
So there is no “additional” time needed after the recording session which is fine. This software does not have a realtime click removal function :frowning:

But have a look at this project:
https://www.youtube.com/watch?v=V5vygs60Txo

I already spent some time to get the software ClickRepairRT installed and running on a Raspberry PI4. It´s an old piece of software and it needs java…
I am still struggling with the configuration of Pulseaudio and receive error messages but I will not give up on this (I am a linux beginner and it´s realy hard with Linux and Audio).
Again - this is looking promising and let´s find out, if the raspberry can cope with this task.

At this time I am not sure if the processing power will be enough. In parallel I will test the java software on my Mac.

Hello,
Just to better understand…
I can understand users who love vinyl for the object, for the turntable and the sound of vinyl…
But if you digitised it, remove defaults (clic, pop), what is the interest against a (probably better) true digital track from master?
Can you explain? Or I missed something?

1 Like

Hi Balbuze,
You would be surprised at how many people will go on about playing analog purely and are running their signal through a receiver made since the 80s that have DSP and are already listening to something digitized and then converted back and never checked to see if their preamp/amp played things through a stritctly analog path. You would also be surprised at how many people failed to hear the difference when an a/d/a convertor was thrown in-line and they didnt know it. You would be surprised at how many records have a digital preview in line when cutting and have been converted once already…

I would estimate at 20% of my collection hasnt been converted to a digital format by record labels… I would also say that collecting cheap albums for me is fun, and having them sound good for a small anount of money by the computer is part of my fun. I would also say that in my own tests of listening to a record and then listening to the digital playback of the same record showed that when conducted blind, myself and others couldnt discern which was which (this is just straight digital, not repaired) with my equipment.

But I get it… you dont see the point, but want to let us know you don’t see the point.

success
RASPBERRY PI 4 witch cheap USB card and ClickRepairRT removes clicks and pops in REAL-TIME
It’s working flawless and as expected

What is important is that you like what you ear… so no problem for me… there is already so many talk about this…
Can you give a link for ClickRepairRT for RPI ? TY!
edit : it is not free and for win9x/10

The software is not easy to find and same with the registration. I want to buy it but can’t find a way… strange.
I got it here :wink:
Go to YouTube link above - look in comments- Jorge Pablo Alonso - find link internet wayback machine
Good luck!

If someone wants to sell his old Clickrepair key or found a way to register the software - please let me know

It’s a Java software and therefore it runs under win, linux, Mac

And how do you use it in Volumio?

This has nothing to do with Volumio… this was just my input regarding a feature request (and it looks like I am pretty much alone with this…)

I did a proof of concept and tested if a raspberry pi4 is powerfull enough to click remove in realtime. Its working! and I tried it with ClickRepairRT running on Raspberry Pi OS.
I am not a Linux expert as well so I will not offer any technical help here. Just follow the documentation which is provided by the youtube video - link above.