DSP Room Correction

I’m making this quite a broad request, as I’ve seen quite a bit of coverage over at AS. Some examples include incorporating branded plug-ins (DIRAC, REW, Audiolense to name a few) into Mosaic.
Or even a dCS software in itself (a big ask)? If the FPGA can handle this sort of processing, I’d love to experiment with this feature.

1 Like

This is an intriguing notion. I would throw Linn’s Space Optimization into this mix as well. I do room correction right now with a Wavelet in front of my Legacy Audio Aeris speakers. And the improvement it achieves is not subtle.


i think you may be asking dCS to take Mocaic where they do not intend to go. They have stated they have no intention of developing a product to compete or work similar to Roon (or other such programs.)

I agree dCS don’t want to take Mosaic the Roon-lite way, but room correction DSP is a feature that audiophiles are turning towards as solutions to improve any setup. One can implement it, but it involves another workstation to fulfil the duties, and given the potential of Mosaic as a platform, I have confidence in it being able to handle such duties. Some may want more than a test signal, phase tuning and LR balance. As for asking for an in-house design, again dCS are experts in the digital field, so even a small team can design an elegant solution that can better a lot than what’s around.


I was going to post here to ask if it was possible to upload a FIR filter into Mosaic (like one can with Roon) to perform room EQ and found this old thread buried in the sediment. Is it fair to assume that nothing has happened here in the last four years?

But just for the avoidance of misunderstanding, is it correct that there is currently no way to incorporate Room EQ into an all-dCS system? If so does anyone know if dCS is philosophically opposed to this or have they just not got to it yet?

No, you can’t do this. As I wrote elsewhere yesterday, Mosaic was never intended to be an all singing all dancing app but something to provide all of the basic needs for a music listener. What it provides probably meets most needs. dCS have ensured that their units are certified Roon endpoints for those who want something further.

Again, as I have also already said, one day there might be Mosaic v.2. Who knows what that may contain?

Many thanks Pete, I was hoping you would chip in!

I would contend that room correction is a basic need for a music listener (at least for those listening via speakers, and in a slightly different way even for headphone listeners). In fact one could even argue it is one of the most basic. And to be clear, what I am asking for here is simply a bit of modularity so FIR filters can be inserted upstream of DAC, I am not asking dCS to offer digital room correction (DRC) functionality themselves (although why not given the potential to raise the real world performance and enjoyment of their equipment?). This could theoretically be done via an extension to the Mosaic platform, and/or even added as a capability to dCS’s various upsamplers.

I don’t really see this as falling under the heading of “all singing, all dancing”, I see it as an absolutely basic necessity! I mean look at the rooms on the “Showing (off) your dCS setup…” thread. Mouthwatering systems of the very best components money can buy, in esoteric racks, cable looms costing tens of thousands of £/€/$ (or more :see_no_evil:) and yet most of them seemingly in rooms with no acoustic treatment whatsoever. Some hold this to be the biggest upgrade one can make to a system and yet judging by the photos at least only a small minority of forum members here has done so (there also seems to be a distinct lack of soft furnishings in many rooms although of course one never knows what is out of shot).

This is certainly understandable in the context of systems in multi-purpose domestic living spaces and I will put my hand up directly as one of the guilty parties**, but it is precisely this constraint that makes DRC so relevant and essential. Even if acoustic treatments are not an option for reasons of domestic harmony DRC can take you some (a lot?) of the way. Of course to those of you fortunate enough to enjoy dedicated and properly treated listening rooms all I can say is hat off to you :tophat:, colour me jealous!

I have no doubt at all that there will be a Mosaic V2 at some point. What I am interested to understand is if dCS have stated a position on DRC. Is it something they believe in and might embrace within their own platform (yes, I know Roon provides for this and that is the direction I am going) at some point in the future or do they have some ideological objection to it? Is anyone from dCS here on the board authorised to comment?

I believe (hope) they will, and when they do so I predict that when we all hear the magnitude of improvement this makes to the sound of our systems the only thing we will be left wondering is why they didn’t do it earlier.

** Based on recommendations from friends I have ordered a DRC filter for my living room from Mitch at Accurate Sound and will of course report back on progress

Love where this is going.

Happy Mitch customer (for my SR1a). Really transforms them.

Very, very happy with the start of my room treatment journey, too — adding three subs has meant I no longer have the irritating bass modes I had earlier. Tight, articulate bass…and without a single bass trap. (Bass traps would have cramped our style with art way too much.)

Next step is installing a range of diffusion and absorption products that are either in the garage or with a picture framer who’s finalising the way I wanted things mounted. Will be fun to hear what that does to things.

After that, some DSP would be good. Looking forward to seeing if dCS can help here. Failing that I have Roon :+1:

1 Like

You can add me to your request for adding the ability to Mosaic to add an FIR filter or filters. Although I don’t feel any immediate need regarding my room acoustics ( it may help though) I may like to add one or more of the filters available for some of my headphones. Incidentally I have no idea what this might entail.

I have never heard anything to this effect. However as you are new to the forum it may be worthwhile restating something that long-time readers should be aware of in relation to feature requests. We are given these changes for free. However they may cost dCS a great deal of money ( multi tens of thousands) to develop. At any given time they have a list of potential projects. There is a budget and these projects are competing for a share, if any. It is not just a technical question but a business decision also has to be taken.

Looking forward to your Accurate Sound report in due course.

Thanks Pete, in fairness I think it was @Kelvin_L 's request but the more the merrier!

Appreciate the reminder that dCS is a commercial venture with shareholders to answer to and that they must prioritise their finite R&D resources. However in the context of the potential to improve the sound of what are seriously high-end (not to mention high cost!) systems I think this merits consideration. The majority of dCS’s R&D budget is after all aimed at delivering sound quality improvements (via existing or new products) so this is not exactly inconsistent with their company mission.

Depending on how it is architected, adapting the Mosaic Processor to accept uploaded FIRs might not require a monumental R&D investment. Integration into the Upsamplers would probably be more complex, albeit that could be offered as a charged upgrade a la Apex. The development of a complete DRC solution would no doubt require a substantial investment. But such an offering would likely be packaged as a new product and as such constitute an incremental revenue stream.

As a guy who loves the room correction in my speaker system (Legacy Valor/Bohmer Wavelet), and has also enjoyed the convolutions supplied by Accurate Sound for my RAQL/Requisites, I will play devil’s advocate here. I don’t see this anywhere near compelling for Mosaic. One can say that room correction is a necessity—and I get that point of view (I spent a lot of money on it)—but there are so many approaches to the subject and so many available pathways, that I don’t see any ROI for dCS here to try and pick the “overall best/most popular” method and implement it. The market is already doing this, in ways that likely prevent dCS from realizing any incremental revenue, let alone worthwhile amounts.


Agreed. However the original request was not that Qobuz should implement DSP room correction themselves. It was more simply that provision should be made to enable the installation of third party produced plug-ins. This would encompass FIR filters such as those produced to match specific headphones. However the priority of this is not known and perhaps dCS may see it as competitive with Expanse (although I don’t believe they really tackle the same issues).

I’m with Greg on this.

The ability to host 3rd party filters makes sense for platforms that are wholly based on general purpose compute, but it does not make sense for embedded systems such as dCS DACs. The Mosaic Processor board (with its ARM CPU) is dedicated to streaming duties, burdening it with convolution filtering would be a big no no. And I surely hope no one is suggesting that dCS support 3rd party filters onto it’s FPGAs :laughing:

I’d say keep it off-board and flexible. If you must have it for your streaming sources, get a dCS Network Bridge, Room-correct the AES output to your hearts desire, and feed the output to your favourite dCS DAC. I can recommend the Trinnov ST2 Pro for those duties in a fully dCS clock synchronised setup.


Thanks Anup. As I said earlier :

So your points are much appreciated.


Not to derail this thread, but I’ve been intrigued by the Trinnov line for some time. Their current all-in-one optimizer/DAC/preamp called the Amethyst is a nice piece of kit and not crazy-priced, though it has some limitations. The ability to keep all this optimization work inside the digital domain before the DAC with no ADA round trip is significant. Right now, in order to have both room correction and my Vivaldi stack in the same system, I have just such an ADA round trip. But it’s not a requirement of the Bohmer Wavelet system; it is quite a nice self-contained DAC/Preamp/Optimizer at about half the price of the Amethyst (though nowhere near as nice looking). To my ears, the Wavelet II is an excellent package, that also has some functional limitations similar to the Amethyst (and not all of the latter’s bells & whistles), and doesn’t sound as organically engaging as when the Vivaldi is added back into my system. I’ve gone back & forth between a Wavelet-only and a Wavelet+Vivaldi system (and during my Apex upgrade cycle, the former all I listened to for several weeks when no traveling), and I consistently prefer the inclusion of the Vivaldi. That may be the Vivaldi’s euphonics, perhaps influenced by my wallet’s long memory of what the Vivaldi stack plus external clocking has cost me. :wink:

[I think the ST2 Pro was marketed toward the studio segment, and is now discontinued in favor of newer models, but there appear to be plenty out there on the used market. I sure like that ability to clock it from the Vivaldi.]

1 Like

Greg, quite right, the ST2 Pro was squarely targeted at the Studio market (hence the Clock input). There was a consumer version, the ST2 HiFi (which IIRC, lacks a separate Clock input, and has RCA instead of AES inputs).

My interest in it was piqued because of; (1) the digital-in digital-out, (2) the External Clock support, and (3) a dedicated mic with precise phase measurement capability. It’s a pity they’ve discontinued it in favour of the Nova (which unfortunately is not ideal for a dCS system).

As I was messaging another user here who asked, my system chain with the ST2 Pro and the Vivaldi stack was a little tricky; I used my dCS Network Bridge as the digital source, feeding 24/96 to the ST2 Pro (AES). The Network Bridge was dual-clocked as normal, but had it’s SDIF-2 Clock Output feeding the single Clock input on the ST2 Pro (that SDIF-2 Clock varies depending on the source rate, similar to the Word Clock Output on the Vivaldi Upsampler).

The ST2 Pro then applies room-correction and feeds a digital output to the Vivaldi Upsampler AES input, from which point everything on the dCS stack works as normal. So, room correction is applied exclusively in the digital domain to the source before it’s fed to the Vivaldi stack.

1 Like

That’s intriguing. Are you still using it? Did/do you like it? I might just get interested enough to try that. I’d still need the Wavelet II for crossover purposes, but then its function would be analog only.

So, at that point, obviously the stream is no longer bit perfect. Have you done any analysis on the stream to observe its changed characteristics?

I originally intended the ST2 Pro for my home-office headphone rig, but gave it a full spin on my main Vivaldi system to really hear what it could do.

Technically, it does exactly what it says it does in correcting both the room frequency response (amplitude) and the phase response (group delays).

However, sonically (subjectively), it did some things well and some things not so well. To Trinnov’s credit though, the system is so configurable - including things like the FIR filter tap length! - that one could probably tweak it to minimise (or maybe eliminate) any of the negatives.

The biggest issue for me though was that it wasn’t transparent in my system even though it was fully clock synchronised and supposedly in full bypass mode, as in not doing any digital signal processing and just passing the PCM bitstream through. I didn’t investigate further beyond that point.

I think in a severely constrained studio, the ST2 Pro is probably invaluable, but in a properly treated room, maybe not so much. While I haven’t actually done any direct head-to-head sonic comparison, using Roon DSP is far far simpler than all the connectivity gymnastics required with the ST2 Pro :grin:

1 Like

Damn. That’s the kiss of death. :wink:

Here’s the thing though, when the ST2 Pro is plugged into the system, the difference between Bypass and Optimise (DSP active) is day and night, and not in a subtle way.

Some aspects of how my system sounded I’d never heard before - fixing room phase issues is quite an eye-opener. But it didn’t take long to notice that bypass in comparison to streaming directly to the VUP sounded sub-par; the main symptom was the missing “air” in delicate tracks.

Ultimately people will make their own decisions on what compromises they want to live with. Like everything, it’s another YMMV situation :laughing:

1 Like