Roon RAAT versus Devialet AIR

Over time, I have often compared the sound quality of RAAT versus Devialet AIR. I used various network configurations, various Expert Pro amplifiers, various amplifier settings / configurations, etc.
Each time, the result was the same: there is a small difference, but easily noticeable, especially in the bass area.

Devialet AIR seems to reproduce bass more accurately, the bass being tighter and better controlled. With RAAT there seems to be a kind of bass dissipation and less punch.
I know that this difference, theoretically, should not exist, but I have heard it every time.

Has anyone noticed this difference between RAAT versus Devialet AIR?

Hi Daniel

I did some comparisons of RAAT vs AIR. This was some time ago. Iā€™ve since sold my 250 Pro.
In the beginning I preferred AIR over RAAT (just after the CI upgrade) but I guess Devialet (or Roon) did something so after a firmware update or two on the 250 I ended up preferring RAAT. Roonā€™s AIR might still have an edge in the mid-range, but - I donā€™t have the 250 anymore so I cannot say for sure. As for bass quality I used both SAM, parametric EQ and RC filters from Audiolense. Any differences between RAAT and AIR in bass would probably be lost :slight_smile:

Iā€™ve been listening to AIR for a few weeks now and just updated to Dos 2.2.6 and lost RAAT, so once I git it working again I did a little listening test, with Roon itā€™s real easy as you use the volume setting to set Transfer Zone, picking AIR while I was listening to RAAT. then use the Zones icon and you can quickly click between AIR and RAAT.

After a while I felt RAAT had just a bit more Bass, and then noticed using the Signal Path, that I had the Bass adjusted +2 on RAAT, so now Iā€™m stumped I really canā€™t hear a difference.

Next I jumped over to JRiver and tried my USB connection, which I have a Jitterbug and Uptone Audio USB Regen in the path with ASIO configured. It sounded great except didnā€™t seemed to be about -5dB quieter, then I noticed I had the JRiver volume not set at 100%.

So now all sound Great. Iā€™ll keep listening for some minor detail different but if I can find one I am not to sure it would matter as everything sounds GREAT. I do like RAAT because of the control I get where I can startup up my Devialet just by Queuing up a song in Roon.

Iā€™m a Dev 220 out to KEF LS50s with SAM configured along with Transparent Supers and a JL E112 in the background.

I did all my tests without any DSP (neither in Roon, nor in Devialet), so as not to influence the result.
However, what intrigues me is that there is a difference, although it is the same system, both protocols are bit-perfect, both use the same input to Devialet (Ethernet).
I understand the differences between USB and Ethernet inputs, but I donā€™t realize where the differences in the same input (Ethernet) come from.
Moreover, the differences always appear, in multiple conditions (with switch, with router, with direct cable, etc.) and always in a similar way. The only thing that differs is the protocol used (Roon RAAT versus Devialet AIR), but both should be bit-perfect, so they should have had the same result when using the same input to the amplifier. And yet the result differsā€¦

1 Like

Psycho-acoustically perhaps. Iā€™ve also done many tests, sometimes I prefer one, other times the other. Iā€™m really not convinced there is a tangible difference. RAAT has advantages in Roon, so I stick with it.

Is it about displaying additional information in Signal Path and grouping it into zones, or about other advantages?

There is a small difference. AIR is UDP while RAAT is TCP. This may or may not represent a difference.
I take it you use Roonā€™s AIR, not the Devialet virtual sound card?

Yes, Roonā€™s Air.

UPDATE: Apparently, using AudioLinux, instead of Roon OS, diminishes the sound difference between Roon RAAT and Devialet AIR.
This leads me to think that the cause of the difference may be related to latency (this is the area on which AudioLinux focuses).

I just found this quote pulled from: https://help.devialet.com/hc/en-us/articles/360000193445-What-is-Devialet-Air-

  • Iā€™ve been told that Roon onboarded Devialet AirĀ® - what does it mean?

Thanks to our partners Roon, they worked to get directly inside Roon the protocol Devialet Air, which is allowing you to directly stream to your Expert (when available on the network).

As a customer, it also means that you donā€™t have to setup Devialet AirĀ® if you want to play with Roon into your Expert, you will only have to setup the option (options are available into ā€œSettingsā€/ā€œAudioā€ menu - all DSP options are also available from there).

Am I reading this right, it sounds like if your using ā€œDevialet AIRā€ from Roon your really talking from Roon to Devialet via the AIR protocol, and RAAT is the RAAT protocol, which would mean that Devialet would have to have built code into Devialet to communicate with Roon via both those options. I donā€™t know the differences in the actual protocols but they do sound like they are different and thus similar to comparing any of the other available options like USB, you may or may not experience a difference.

Correct. Roon is the only ā€˜third partyā€™ that has been given the AIR (Asynchronous Intelligent Route) protocol specification to stream natively to Devialet Expert amps. It will work all the way back to the D-Premier amp fitted with the wireless board. Quality wise it is on level with Roonā€™s own RAAT but cannot be grouped with other Roon zones. As I said further up AIR uses UDP while RAAT is TCP.

EDIT: Roonā€™s AIR streaming was added in Roon 1.3 b242 (July -17). RAAT was also using UDP in the beginning, but was changed to TCP in 1.3 b234.

1 Like

According to Devialet, AIR uses a buffer (you can adjust it in AIR settings to various sizes) to isolate the DAC from the incoming stream. Iā€™ve not been able to determine whether the RAAT implementation uses that same buffer. If not, might that explain some of the difference?

Maybeā€¦
Roon RAAT seems to use buffer in DAC, but it does not appear explicitly written (https://help.roonlabs.com/portal/en/kb/articles/raat):
ā€œWe might decide to change the buffer size requirements on the device to increase stability.ā€

Right, RAAT requires some kind of DAC-side buffer, but Iā€™ve never found any clear documentation that on Devialet it uses the same buffer as AIR does (and Iā€™ve never found any RAAT setting that varies the buffer size, as in AIR settings).

I suppose I could unplug the Ethernet cable while the Devialet is playing and see how long it continuesā€¦ and then see whether making changes to the AIR buffer size setting makes any change to the time it continues playing on RAAT.

But it would still be rather hard to ā€œproveā€ that differing buffer implementations account for differences in sound. FWIW, I also prefer the sound of AIR over RAAT (this on an Expert Pro 1000 fed over Ethernet).

This topic of the sound difference between RAAT and AIR protocols has an interesting side effect. In the two compared situations we have:

  • the same audio chain: the same hardware components, the same software, the same connectivity, etc.
  • the only difference is the transport protocol, but both protocols are bit-perfect.

The fact that in these conditions there is a difference in sound, I think is a pretty strong proof that in the audio field the expression ā€œbits has bitsā€ does not apply.

If anything, itā€™s only a demonstration that ā€œbits are bitsā€ is a meaningless over-simplification that means different things to different people depending on the axe they want to grind.

Itā€™s not ā€œproofā€ of anything (in the scientific sense) unless the effect is repeatable under conditions that eliminate the possibility of bias.

1 Like

Many users perceive the sound differences between RAAT and AIR. This is a fact.
ā€œBiasā€ and ā€œimaginationā€ counter-arguments are very easy and convenient to use at any time and for almost any topic of discussion, but they are not constructive (nor can they be proven in this case).

There is something interesting here: for a long time, there has been a bug in the Roon - Devialet communication via RAAT, not solved at present. Sometimes it even causes a temporary interruption of sound. Isnā€™t the total absence of sound a difference in sound quality?.. Or is it still ā€œbiasā€ and ā€œimaginationā€?..
Is there anyone here who can say with certainty that this bug does not affect the RAAT sound quality (when sound it is not completely interrupted) and that no bit loss occurs? Maybe this is exactly the cause of the sound differences heard by usersā€¦

Thatā€™s an interesting take on the subject. I myself have heard no difference between RAAT and AIR but I do hear an improvement when using USB so I guess Iā€™m not a ā€˜bits are bitsā€™ believer, even though I ā€˜believeā€™ I am.

I have however had something happen to me a few years back that might give some weight to your theory. I was listening and wasnā€™t enjoying the sound of my system at all. it was harsh and annoying. After a while the music started to drop out. I found out when there was no sound that the fan of my Mac mini was running. I had never heard the fan run before. It turned out that I had made some changes to my rack and put the Mac on the same shelf as my amps. The amps had over heated the Mac and I guess the processors werenā€™t working at their optimum level. When it got too bad, the drop outs started.

This all ended when I took my Mac away from my amps. Lesson learned; there can be a change in sound quality even when the music doesnā€™t drop out. Whether that is because bits are missing or too much data needs to be re-sent, I can only guess.