Format Conversion / RAAT

Just did a new test.
I picked the song: thanks to you from Boz Scaggs for comparison
Compared WAV with FLAC compression 5
Difference I hear between the 2 is
The trumpets are more relaxing.
drum stick on metal is less harsh
Soundstage is a bit wider

all in all more relaxing, more analog ?

My system consists out of :
Intel NUC I5 for roon core
Optical link for galvanic isolation
Raspberry PI 4 with VitOS
Iso Regen from uptone
Denafrips Ares 2 DAC
Schiit Saga Preamp
Schiit Aegir amp

What was the original source from to create the new digital copies?

1 Like

my source was a copy of a CD FLAC(5)
The flac file was converted to WAV using a program called pro audio converter

Could you post a screenshot of the Roon signal path for both files?

Hmm. Roon Bridge on the Pi should be seeing identical bits from the RAAT stream. I have no idea what might be causing an audible difference. And I hear the Ares is pretty immune to usb effects anyway. Does the ISO Regen actually make any difference?

To be honest I had my doubt about the impact it could make when I bought it online for some reviews were pretty negative.
Reason for buying it is that I noticed that my DAC’s sounded different sound depending on which USB port I used on my Raspberry Pi 4b
Adding the IOS regen between the Pi and the DAC made a huge difference.
The difference between the USB ports is gone but most important, the soundstage is also much wider.
I still do not understand why this is happening.
At the moment I’m going into the direction that it has to do with impedance mismatch between the PI and my DAC’s ? leading to reflections, creating smearing. a bit the same impact you get as with jitter
So yes it’s for been a pretty good investment.

Nice system, but I think I’ve spotted the problem: you have no loudspeakers!

2 Likes

:slight_smile: good spotting
I had left them out for they are non brand specific
In the Netherlands a few companies are designing high end speaker systems that you then can buy as a component package.

https://www.speakerland.nl/luidsprekermodellen/vloerstaand-vanaf-800-/vifa-vivace-2012/

1 Like

While I’m not going to dispute that USB output from a Raspberry Pi (even a model 4) might be problematic I’ll point out for curious readers that an I2S HAT would likely solve any issue at between 1/10 - 1/5 of the price of the regen. It’s also a single box solution that works unless the device you’re looking to feed is USB input only (there are no USB HATs to my knowledge).

sounds like an open door to make money…

Or just have a DAC that has a decent USB interface so it makes no difference. Both my SMSL and RME don’t change one iota with any different ports on the pi , cables or otherwise.

1 Like

Sad to hear that all the DAC’s I bought are not Decent.
Seems that Auralic, Schiit and Denafrips are not up to your standard

@Simon_Arnold3 makes a fair point though, many a Hi-Fi add on seems to be designed with the intention of fixing an problem that’s really an engineering issue elsewhere. There’s a logical argument here. If the add on is capable of “cleaning” the USB signal fed to a DAC, why didn’t the DAC manufacturer take the same precautions when they designed and built it? Noise rejection at the input stage is an engineering fundamental (garbage in garbage out?) and shouldn’t be left to add-ons. As an example I’m all for the the use of linear PSUs for analogue amplification, but I wouldn’t buy a pre-amp/active speaker that didn’t incorporate one already.

Interesting that without any questioning the DAC is the root cause and therefore considered to be poorly designed
Why is it the DAC ? Is the Pi a perfect device ?
If it is caused by impedance mismatch/ reflections then there are 2 sides.

Now this is valid in all things but audiophile.
If a car manufacturer could make a higher quality part that won’t have the failure rate the current part has for only $1.00 more, why don’t they?

Because people won’t pay for it and the car manufacturers complain they can’t make money.

  • if they sold 1 million cars they save 1 million dollars by not improving the part.
  • If they sold 1 million cars, they have 1 million more products that could fail that could give them service and repair income.

So it’s a win for car manufacturers.

Now with audiophiles it’s different… if it cost them $1.00 to make a better usb input… They’ll just charge you $899.00 more for the audiophile grade low noise, low jitter, galvanically isolated usb port.

And audiophiles will buy it.

So it’s a good point that @Simon_Arnold3 brings up.

I do agree with you up to a point, and the Pi is a general-purpose computing device, not specialist audio kit. But the DAC is an audio device which is designed to generate an analogue signal that’s susceptible to noise. If they’re not eliminating noise at input I’ll immediately worry about where else they might have overlooked the issue. The Pi’s USB output doesn’t produce fundamentally poor USB data, i.e. all manner of devices like external discs work, so the signal is fine, the DAC just needs to reject the noise.

I don’t think it’s noise It might have the same impact on the sound but I believe that it the root cause is impedance mismatch.
I noticed that when changing my usb cord between the Pi and DAC to the brand Audio quest the sound stage also changed. Now with the IOS regen in-between I can use any usb cord and it doesn’t have any impact. That to me looks like impedance mismatch

This may be the issue, and you could also resolve the “blame game” here. The USB 2 and USB 3 specifications are pretty clear that a 90 Ohm impedance between data negative and positive is required. If there’s a problem either the Pi or the DAC are out of specification, possibly both :wink: Knowing which might be informative…

Well, actually the case is probably the cable. A lot of “high end” cables are actually out of specification on purpose. With an in-specification cable, there should be no difference.

Also, most DAC manufacturer’s use “off the shelf” USB components, aka XMOS, and their engineers probably have little experience in integrating USB. I mean there is a reason Schiit spent the time and money to develop their own USB implementation.

I find it interesting that most XMOS USB firmware has been updated in the last couple of years to add MQA. How come people aren’t complaining that the additional overhead of processing MQA on the USB chip isn’t causing audio issues.

2 Likes