Expand KEF LS50WII + KC-62 system to include multichannel with DL room correction

[thmeky]

Dear all,

I am very happy with my current system, but I miss a proper center channel and surrounds for movies, LFE downmix doesn't always work properly and would like to use DL room correction.

It would be awesome if KEF adds Dirac live to their software platform and cooks up a smart RJ45 switch to expand to multichannel, but I don't think they will.

So the idea is to use a Mac Mini or small PC as a decoder for Dirac (VST in Jriver or stand-alone) as well as a playback/streaming device for movies and music. The computer would be attached via USB to a digital audio interface: https://www.rme-audio.de/digiface-usb.html. Am I correct to assume that all the four toslink outputs can be configured as two channel SPDIF? One of those optical outputs would be fed to a simple, two channel, high quality DAC without volume control. The analog outputs would be connected to a power amp that feeds the center channel (would a NC250 be enough for a LS50 Meta crossed over at 100 Hz?) and the subwoofer respectively. Two different SPDIF outputs from the Digiface will be connected to the Toslink inputs of both the LS50WII fronts and the LSX surrounds.

Do you think that everything can work as intended? Will the Digiface be visible as a 7.1 audio device to Dirac, assuming its configured correctly? I understand ADAT and SPDIF are different protocols, but RME clearly states that the I/O is Adat and SPDIF. And would Dirac be able to tackle the timing/phase issues that will arise when everything is combined?

Thanks in advance!

 

[Kal Rubinson]

The issue will be synching the various outputs. It is possible that DL will take care of that at the outset but it is questionable whether the un-named stereo DAC will drift differently from the KEFs over time or if even the two KEF pairs will drift differently. Such issues are usually dealt with by using clock links but that's not possible in your proposed configuration.

 

[thmeky]

The issue will be synching the various outputs. It is possible that DL will take care of that at the outset but it is questionable whether the un-named stereo DAC will drift differently from the KEFs over time or if even the two KEF pairs will drift differently. Such issues are usually dealt with by using clock links but that's not possible in your proposed configuration.

Won't the Digiface take care of clock sync issues? Or does this happen in the digital to analog domain?

 

[mdsimon2]

I think this should work well. SPDIF contains a clock signal so that should handle clocking issues as long as your DAC is not something with a FIFO buffer. I've done testing using a miniDSP nanodigi which has 4x SPDIF outputs with different DACs and it is easy to get them to sync up by setting delays and they don't drift.

Michael

 

[Kal Rubinson]

Won't the Digiface take care of clock sync issues?

Initially, yes.

I think this should work well. SPDIF contains a clock signal so that should handle clocking issues as long as your DAC is not something with a FIFO buffer. I've done testing using a miniDSP nanodigi which has 4x SPDIF outputs with different DACs and it is easy to get them to sync up by setting delays and they don't drift.

"Should" is operative word here. Glad to hear of your success but the KEFs insert significant DSP in the signal path before routing to the dual DACs per speaker and that may be a complication.

 

[mdsimon2]

Initially, yes.

"Should" is operative word here. Glad to hear of your success but the KEFs insert significant DSP in the signal path before routing to the dual DACs per speaker and that may be a complication.

It would seem odd that a DSP speaker would have variable latency, if only because it would make subwoofer integration difficult / impossible. But I guess anything is possible, just wouldn't be my base assumption. Probably best to ask KEF directly.

Given that the OP only needs two stereo digital outputs and one stereo analog output it might be a good idea to get an audio interface with the appropriate outputs but also analog/mic inputs. For example a RME UCX II or a MOTU Ultralite Mk5 would provide all the I/O the OP needs without an additional DAC. This would allow for acoustic phase measurements to ensure that nothing weird was going on with the KEFs.

Michael

 

[thmeky]

Both of you thanks for the input. The MOTU Ultralite Mk5 could be an interesting altenative.

 

[phoenixdogfan]

Won't the Digiface take care of clock sync issues? Or does this happen in the digital to analog domain?

Confused why your need the Digiface. From what I see it has only 4 digitial i/o's and you should have at least 6. Why not just buy a mulltichannel DAC like the ones from Octo, Motu, or Topping--all of which Amir has reviewed. Since all channels of these DACs run off the same clock, no syncing problems, and they will accept mullti channel USB input. That's what I do. I have a Mele Quieter 3, running JRiver with the multi channel Dirac Live plugin outputting via USB to an Octo DAC 8 Pro. L-R amp is a Purifi Eval 1 to LS 50 Metas, the center goe to an Aiyima A7 which powers a Kef HTC 3001SE, the sub channels go to dual SVS SB 2000s, and the surrounds go to a Behringer A500 which powers OG LS 50's. JRiver handles all the routing over USB just fine, and does bass management as well. Even if you use powered speakers, a multichannel DAC like the Octo can drive them AND the amplifiers for any passive speakers. I believe all the LS 50 Wireless IIs have analog inputs as well.

Again I don't see the need for the Digiface.

 

[thmeky]

Confused why your need the Digiface. From what I see it has only 4 digitial i/o's and you should have at least 6. Why not just buy a mulltichannel DAC like the ones from Octo, Motu, or Topping--all of which Amir has reviewed. Since all channels of these DACs run off the same clock, no syncing problems, and they will accept mullti channel USB input. That's what I do. I have a Mele Quieter 3, running JRiver with the multi channel Dirac Live plugin outputting via USB to an Octo DAC 8 Pro. L-R amp is a Purifi Eval 1 to LS 50 Metas, the center goe to an Aiyima A7 which powers a Kef HTC 3001SE, the sub channels go to dual SVS SB 2000s, and the surrounds go to a Behringer A500 which powers OG LS 50's. JRiver handles all the routing over USB just fine, and does bass management as well. Even if you use powered speakers, a multichannel DAC like the Octo can drive them AND the amplifiers for any passive speakers. I believe all the LS 50 Wireless IIs have analog inputs as well.

Again I don't see the need for the Digiface.

I was under the impression that each digital SPDIF output from the Digiface or a comparable device could be configured as two channel? A proper multichannel DAC like the Topping DM7 would be the easier solution, but don't you think that an extra ADA conversion in the active KEF speakers would degrade the sound quality?

 

[digitalfrost]

I have a similar setup and also use the Digiface. I think this should work. I feed two stereo DACs for satellites/stereo subwoofers from it. What you need to know:

You can theoretically use all 4 outputs but it depends in bitrate/samplerate how many AND the outputs change internally in TotalMix when you change sample rate. So just must either route the same input to the multiple outputs or make sure you stay at one samplerate and resample everything else.

I cannot speak for Mac, but on the PC the max channels a Windows WDM soundcard can have is 8. Which outputs these 8 map to is depending on sampling rate. If you need more channels than this you can, but you need to move to ASIO and use ADAT instead of SPDIF. Relevant part from the documentation:

Standard: 32 channels 24 bit, up to 48 kHz
Double Speed (S/MUX): 16 channels 24 bit 96 kHz
Quad Speed (S/MUX4): 8 channels 24 bit 192 kHz

I suggest you read the manual carefully https://www.rme-audio.de/downloads/dface_usb_e.pdf

Maybe this output from TotalMix makes it clearer:

Software playback are the channels as represented to the operating system.
Hardware outputs are the virtual outputs of the Digiface card, which physical plug this is on the interface in the real world depends on the sampling rate. To be clear: You don't need to reconnect plugs, but the hardware outputs channel name will change.

I'm currently running at 96khz. I have my main speakers connected to ADAT 5/6 out, this is the 2nd physical output ont he card, and I route WDM channel ADAT 1/2 to this output. This way, if I make a mistake by default WDM 1/2 (software playback) would be routed to ADAT 1/2 so my subwoofers would get the singal. I.e. if something is wrong with crossover or something and I play stereo, the signal only goes to the subs.

The subs are connected to ADAT 1/2 (the first physical port on the card) out but I route WDM channels ADAT 5/6 to this (you can also call this RL and RR in 7.1...)

Now you will notice that ADAT 9/10 also gets signal. This is currently unused, but if I were to change the sampling rate to 48khz, my satellites would move from hardware output ADAT 5/6 to ADAT 9/10, since theoretically because of the 48khz I could have more channels in between 1/2 to 9/10. Since I use SPDIF it doesn't matter because it's stereo anyway but internally the card represents everything as ADAT. It does not affect the physical connections, just internal representation. ADAT 1/2 never changes since it's by definition the first two channels in software and the first physical port on the card. But using ADAT it could be the first 8 channels, thus the next available channels for 2nd output are 9+10.

I've measured this system multiple times and I have no issues with clockdrift if use the Digiface as the central piece. For measuring I connect an ADAT mic pre to the input of the Digiface and sync it to the clock of the source. You can choose for every output if you want to run SPDIF or ADAT.

The input automatically detects what is running. For measurements I use a 2nd virtual WDM soundcard, I think something like channel 25/26 ADAT map to the physical input 4 on the card.

If you use a 2nd USB microphone for measurements it will not work correctly because you will get clockdrift between devices due to ASRC from the USB.

Some software like Audiolense can compensate for that but I'm not sure how well it works. I guess most people using something like the UMIK have this problem so maybe Dirac can deal with it as well.

If I had to do this again today, I would probably choose a Fireface, but I'm still very happy with this setup, it just works. Also the Fireface is much more expensive, but depending on what you spend for DACs who knows what the total will be.

I can surely understand the wish to avoid unnecessary AD/DA conversions. The measurement issue could point to a Fireface it's really a single device that will do everything you need.

Another thing to point out, I also choose the Digiface because of galvanic isolation. I still use unbalanced connections to the amps and if you have any amp with a ground plug (most modern class D amps are and probably also all active speakers) you get problems if you involve a computer (with another ground) and unbalanced connections. Nowadays I would simply buy equipment with balanced connections...

Both are good solutions.

 

[phoenixdogfan]

I was under the impression that each digital SPDIF output from the Digiface or a comparable device could be configured as two channel? A proper multichannel DAC like the Topping DM7 would be the easier solution, but don't you think that an extra ADA conversion in the active KEF speakers would degrade the sound quality?

I think those ADA conversions arecompletely transparent. It's a holdover of the old audiophile digiphobia idea that AD and DA conversion adversely affect sound quality. Might have been true in the 1980's, but not true today. It's kind of like the old audiophile hang up on using PEQ. If you look on these pages, you will find PEQ, AD conversions, and DSP are all employed freely for the simple reason they allow systems to work and/or work better,

I was mistaken about the Digiface, btw. Those four I/O are each capable of handling two channels, for eight channels in total. If you do decide to use it, you might want to check out if all the outputs are slaved to one master clock, or else you will have four dacs each playing out of sync with one another.