I have been using Melda MConvolutionEZ with pretty decent success with UPnP after @Antoine recommended it. Streaming Tidal to a WiiM Pro using a MacBook M1 with 8gb RAM.
I have one question - bear in mind that I am new to convolution/DSP and still learning the ins and outs.
When I play a track with a certain sample rate, say 48khz, and I use a 48khz convolution file, everything is great. However when I play a track with a higher sample rate, it often results in distortion (e.g. Dream Brother by Jeff Buckley @192Khz). Loading up a convolution filter with the same sample rate as the track seems to fix this, but there is a chance that I will forget to do so at some point (its not exactly the most user friendly way around). This may or may not be related to the piercing, trouser soiling static noise that occasionally plagues listening sessions (as described in other threads). I will make a separate thread for that though because it seriously ruins the experience to an extent that should spark .
Has anyone else experienced and found a workaround for this via MConvolutionEZ or another convolver?
I tried upsampling lower sample rates and downsampling higher ones to a middle-of-the road figure (88khz) and loading an 88khz file, but that results in super low volume with lower sample rate tracks.
What is the behavior when you connect directly from your computer to the DAC via USB?..
You have 8GB of System RAM to work with… How many filters are you employing in the convolution host?.. Also, you must manage your gain structure through the entire processing chain so to avoid overload distortion (digital clipping). Adding a single plug-in alone adds approximately 2db of gain.
What is the behavior when you connect directly from your computer to the DAC via USB?..
No difference. Problem persists and the fix is the same.
You have 8GB of System RAM to work with… How many filters are you employing in the convolution host?..
Just one stereo filter. RAM isn’t the problem here as far as I can tell. Maybe the piercing static noise is a result of RAM, but the distortion can’t be.
Also, you must manage your gain structure through the entire processing chain so to avoid overload distortion (digital clipping). Adding a single plug-in alone adds approximately 2db of gain
I can’t find a way to do this with MConvolutionEZ, maybe there is a way that I haven’t found. HLC makes it possible. Where would you tweak this in the signal path? Just lowering the volume doesn’t do anything and limiting the volume control also doesn’t stop the clipping.
These are one in the same… digital does not distort in the way analog distorts… digital clipping is sharp transient related… Digital “hash” is related to loss of synchronization of the DAC with the signal sample-rate… If the system is running out of processing RAM these anomalies can be produced, relative to all System + Audirvana related operations… Obviously you have file sample-rate mis-matches, that produce the “hash” …
The Hang Loose Convolver has a feature that allows you to have multiple FIR filter wav files that are created at the multiple different sample rates and will upsample or downsample the FIR filter files on-the-fly to match your music sample rates… BTY the HLC uses an embedded version of the r8brain resampler…
For instance I use 44kHz and 48kHz FIR filter files (power of two for the various sample rates) for room correction and let the Hang Loose Convolver upsample those FIR filters on-the-fly to match any music I might stream locally or remotely… If you were a perfectionist you could create a FIR filter file for each and every sample rate you could possibly run across and the HLC would use the matching sample rates on-the-fly…
I don’t upsample using the Audirvana r8brain upsampler as my DAC does a good job of that internally… I also don’t use or convert to DSD… I am not a fan of DSD… And use other methods to achieve some of that special sauce…
HLC also has level matching and level management that you should read about… No trouser soiling here…
Oh wow - that’s amazing! Mitch continues to impress. I have kind of made filters for all sample rates from 44 - 192 (WiiM limit) but switching manually is just a PITA, especially seeing as the fix for UPnP effects seems to be turning real time off… Might have to get back into HLC. Thanks for the response - this helps.
