My conclusion is that I hear a difference between a level 5 FLAC and level 0.
Wav 24/96, Flac 24/96 level 0 and Flac 192/96 level 0 all sound identical.
The lower the bitrate the more level 5 and level 0 sound identical
At 16 Bit 44.1 KHz (I’ve used a CD version for this) there’s no hearable difference between FLAC level 0,5 and WAV.
I, for now, have concluded that Roon is not transparant when decompressing High res FLAC.
Makes sense or… where do I go wrong?
Open to hear your thoughts.
My Setup is:
QNAP NAS for file storage.
PC intel I5 with 8GB of RAM + SSD that’s only running Roon.
Wired transport with optical link to block electric network noise entering the DAC.
Roon endpoint : Auralic Mini with BOTW power supply
DAC : Schiit Modi Multibit DAC
Preamp : Schiit Saga preamp
Amp : Schiit Aeger Amp
No blind testing , you are correct I wasn’t 100% objective.
My mind was set to proof that there could be no difference, that I was making things up.
Therefore In reality the sound difference will likely be bigger than I’m now willing to admit.
Still hoping that Ive done something wrong.
If you are concerned about the processing footprint of a Roon Core affecting sound then you can use a network connection to a small footprint Roon Ready or Roon Bridge device to isolate the Core from your sound system. See this KB article.
I use a microRendu as an HQ Player NAA. It can also operate as a Roon Ready endpoint.
The way I read your suggestion is to add a roon bridge into the config.
The Roon core will communicate with the bridge and the bridge will become the communication point to my Auralic endpoint
Although it’s a good distributed architecture I don’t see how it will resolve what I’m experiencing.
What I read from the documentation is that there’s no separation in functions when adding a bridge.
That means the Roon core is still doing the Flac decompression. Correct ?
Might be interesting to try the same test with this configuration (auralic rendering via network connection to core) as you have the hardware. I prefer a network renderer rather than listening to the core via the amps usb input.
Correct, the Mini is a Roon ready end point.
I’m connected via ethernet 1Gb and added a optical link to reduce the electrical noise out of the network entering my Auralic Aries Mini.
This is not the first time I’m hearing a difference between compressed flac and uncompressed flac (WAV)
8 years ago I streamed my music using the logitech squeezebox touch.
I then learned that the processor in my squeezebox was just not up to the job.
I have enough storage and converting flac 5 to flac 0 is a matter of letting my PC work work on it for a few days and I’m done.
Perhaps some of you might think that what I’m going to say is incorrect and something in-between my ears. I would say give it a try, do the conversion yourself and judge your hearing instead of your brain logic.
I’ve converted all my Flac files from the compression they where on (most of them 5 and higher) to Flac 0. This makes them uncompressed and they are about the same size as a WAV file
Yes it did took 300GB of extra space on my harddrive + my computer was bussy for a few days to do the conversion but it’s good value for money.
From now on, every FLAC song I put in my collection will be Flac 0.
You would need to used “Uncompressed” FLAC for the file to be uncompressed. Also, the compression level chosen really only effects encoding time. Decoding compressed FLAC files at any compression level takes pretty much the same amount of CPU time.
Also, if you are using a separate Roon Core talking to some Roon Endpoint over Ethernet, the Endpoint receives the exact same PCM regardless of format or compression level. So it HAS to sound the same.
part 1 of your statement that the Roon endpoint is receiving the same data that I agree with.
That therefore it also has to sound the identical that and 3 others with me who helped me with comparison disagree.
Don’t ask me why for technically I don’t understand but it does sound different. Flac uncompressed just sounds better than Flac compressed.
But give it a try !
Repectively, that’s not the way blind tests work. Your explanation is only one step above the old standard response of “even my wife, who doesn’t care about audio, came from the other room and asked me what I’d done to the system because it sounded so much better.”
edit: But if if makes you sleep better at night to have uncompressed FLAC, there’s nothing wrong with that. At least it’s better than WAV files (because of tagging and built in CRC checking).
But it doesn’t when you are using a a separate Roon Core and Roon Endpoint. The Endpoint gets the exact same bits in the exact same way regardless of the file format used to store lossless files. The Endpoint has no idea what format the files are stored in. All it gets is uncompressed PCM data. So there work the Endpoint does in all cases is EXACTLY the same. The endpoint is what sends the data to the DAC. Not the Core.
I have tried WAV and uncompressed FLAC. All those uncompressed formats do is take up more space on my hard drive. I wish they sounded better because I am always looking for a free way to get better sound quality. I wanted the uncompressed formats to sound better but they did not.
It is not necessarily the case that the more compressed a FLAC is the longer it takes to uncompress. It depends on the file, but sometimes a more compressed file can be quicker to uncompress. Depends on the file. The real cost of highly compressed files is the time taken to compress them - the FLAC algorithm tries more strategies in order to find the best. Mostly FLAC is trivial to decompress. This is easy enough to prove. Compress a large boxed set to different levels and the calculate the checksum (ie decompress) each. Won’t be much difference. When you factor in the lower I/o of the compressed file it is quite possible that a compressed file presents less load on the system.
Correct, the end points are receiving uncompressed PCM and that has nothing to do with Flac being compression or not.
That’s why I’ve started this topic on this community for I believe it’s something in the Roon software.
I run Roon on a dedicated pc with a I5 processor and 8G of Ram. that should not create a bottleneck i presume ? Years ago I owned a logitech touch and here I also heard a difference between WAV and FLAC. A Few years later this problem was identified as a result of limited processing power in the logitech touch.If was just not powerful enough to decompress FLAC in a good way.
I agree with you that if you’re not hearing the difference don’t waste harddisk space, that doesn’t make sense. Like I said. to make sure that I was not fooling myself 3 other persons did the same comparison and confirmed my observation.
Really? It happens to be the case that, by default at least, the Logitech Media Server converts WAV to FLAC before sending it to the Touch, to save network bandwidth. WAV and FLAC are both sent to the Touch as FLAC.