Saturday 1 August 2020

PREVIEW: RME ADI-2 Pro FS R Black Edition AD/DA Converter

Look what showed up at my door (finally!) the other day. This is the latest hardware update to the RME ADI-2 Pro line of converters, the RME ADI-2 Pro FS R "Black Edition" - quite the mouthful! Let's just use "FS R BE" for the rest of this article for short in reference to the name.

Currently, the street price is around US$2,000.

Back in mid January, RME debuted this latest iteration of the ADI-2 Pro at NAMM 2020. Since 2018, I have been in contact with Matthias Carstens (co-founder of RME) off and on about the ADI-2 Pro FS that I have been using for measurements on this blog. Shortly after NAMM, I contacted Matthias about getting an upgrade as I could use two of these ADI-2 Pro devices - one for the soundroom and another for my test bench - when the company starts shipping these.

This "FS R BE" was purchased and shipped through the American distributor Synthax (no such thing as a freebie, or "longterm loans" here at the Musings). I suspect the COVID-19 pandemic over the last few months has delayed the introduction and availability of new products - increased costs of transportation and delayed supply chains. Let's say it took awhile for this box to get across the Atlantic! Anyhow, I'm glad that the machine eventually arrived "home" safely...

Here's what's inside the box:

Notice the breakout cables in the plastic bag for coaxial S/PDIF and AES/EBU.
As you can see, the contents and presentation of the box are very similar to my RME ADI-2 Pro FS previously (most of what was discussed in 2018 will in fact apply). As you can see, there's the 12V/2A switching power supply, power cable, and very thorough spiral coil manual (in English and German; can be downloaded here). For those new to the ADI-2 Pro, remember that there is a locking mechanism on the power input to be mindful of. Remember, we're looking at "pro" audio gear here and extra attention has been given to things like this; it would be bad to have the power cable come loose during a recording session, right?! Battery-powered operation as previously discussed should be no problem.

Cosmetically, other than being all black with white lettering - "Black Edition" - the dimensions and outer metal construction appear almost identical to my "Pro FS". This "FS R BE" unit however obviously has inherited the remote control from the RME ADI-2 DAC FS, its DAC-only cousin. We'll talk more about this later.

The dual 1/4" TRS "Extreme Power" headphone jacks provide low distortion across wide-ranging impedances. Output impedance is rated as 0.1Ω. It's capable of +22dBu (9.75Vrms) at "Hi-Power", current limit of 260mA, and maximum of 1.5W per channel. Should be plenty of juice for your headphones.

When you look at the description for this device, you see that it's designated as "2-in/4-out". This refers to the fact that the rear outputs and headphone PH1/2 jack share the same signal, while PH3/4 is independent with separate DAC output (hence 4 analogue channels out total). There are all kinds of options including balanced headphone mode where PH3/4 and PH1/2 carry separate channels, and "mute" modes where the DAC will turn off the rear outputs when you plug in headphones. This is simply an example of the amount of flexibility that has been programmed into this device. Also the reason why there's a thick spiral ring manual written in small font included!

For the uninitiated audiophile, it can be overwhelming the density of features packed into this box. If you use it regularly like what I do as a tool, you'll get used to the knobs and nested menus soon enough.

As you can see, the 22-button remote is a broad plastic slab (powered by a single CR2025 lithium battery) with buttons customized for the ADI-2 family of devices. There's of course the basic power/standby, volume control, mute, input selection (coaxial, optical, USB), and also extended features like treble/bass tuning,  L/R balance, loudness (LD) on/off and a few others you can see in the manual. Notice on the remote control there's a top left button identified as "SEL". Holding this button down will let you tell the remote which ADI-2 machine it's being paired with. On the back of the remote, currently there are the "ADI-2 DAC" and "ADI-2 Pro FS R" identified (the little LED beside "SEL" will change color for the target device).

Coming from a "pro audio" lineage, the analogue outputs on the back are balanced XLR and robust dual 1/4" TS line outs. If you need RCA output, use TS-RCA adaptors. As in previous ADI-2 Pro iterations, there's a pair of Neutrik combination balanced XLR/TSR connectors for analogue in. Note that these are line-level inputs so for microphones you will need a proper mic pre-amp.

On the left side of the rear panel, there is a myriad of digital inputs including USB2.0 (UAC2 compatible), TosLink in/out, and the electrical D-sub "Digital I/O" connector for break-out cables to handle AES/EBU and coaxial digital.

The "FS R BE" with its older brother the "Pro FS".
For the most part, bit-perfect playback and recording is what I would use the device for. Remember though that there's a powerful >2 GFLOPS DSP inside (same as the Pro FS) to manipulate the sound including a 5-band parametric EQ, crossfeed, bass/treble, and loudness compensation. The full suite of DSP functions is available up to 192kHz samplerate with limitations above that. I find it hard to imagine 192kHz being a significant limit unless you routinely upsample to 352.8kHz+ rates on playback and for some reason still need the built-in DSP instead of performing processing with your upsampling device.

Let's talk about the main functions and what's new compared to the previous Pro FS...

1. There has been an update to the DAC chip internally. It has been bumped up to the Asahi Kasei Microdevices (AKM) AK4493 (from the AK4490 of the "Pro FS"). I know that RME has put extra effort into squeezing out the least distortion from this chip so I will be looking forward to what I find when I measure the device. There's also +2.5dB headroom which will be good for those who listen to loud, dynamically compressed stuff that trigger strong intersample peaks.

The temporal performance should remain excellent using RME's "SteadyClock III" circuitry and I fully expect jitter to be a non-issue like before.

One very important point about this device which is rather unique among the DACs (and ADCs) I have seen is that RME implemented four hardware reference levels that allow the device to maintain high dynamic range across a broader output level. With Matthias Carstens' permission, here's a graph showing this behaviour in the ADI-2 DAC that he posted on Audio Science Review back in May:

SINAD vs Volume Control - ADI-2 DAC AutoRef Edit.PNG

That auto-reference level ("AutoRef") behaviour switches every 6dB (as indicated by the dBu reference values), keeping the dynamic range optimal as volume is reduced. As per the graph, this is clearly quite an improvement compared to when it's turned off.

This is useful when you use the ADI-2's digital volume control with headphone listening or connected directly to an amplifier. Depending on your headphone and amplifier sensitivity, typically the volume will be turned down to something like -12dB to -30dB for listening. With "AutoRef" mode turned on, the DAC adjusts to your listening level in a way that still keeps noise and distortions as low as possible. This is a practical and I think undervalued feature that I will look for in future DACs.

2. The ADC portion hasn't changed substantially with this update, it's still using the 4-channel AK5574 converter in dual-mono mode. Specifications list the analogue input THD at -1dBFS to be -116dB with THD+N at -114dB (0.0002%). Remember that this kind of resolution is absolutely unnecessary for audiophile duties like vinyl rips or home recordings but important for measurements. Realistically, the vast majority of devices I'm measuring will not achieve distortion anywhere near that limit especially if we're talking about things like amplifiers. Indeed, one of the "use cases" for these ADI-2 Pro devices is for performing measurements like what I'm doing on this blog. Of interest, RME has implemented a 384kHz digital compensation filter to allow for "ruler flat" frequency response measurements to 90kHz. I'll need to remember to use this when the situation arises.

While typically during measurements, I'll keep the input level set at a steady reference (usually +13dBu or +19dBu), similar to the comments above about the DAC, the RME's discrete hardware reference levels (+24dBu, +19dBu, +13dBu, +4dBu @ 0dBFS) will optimize dynamic range across input settings. Again, a practical feature that maintains optimal resolution found on pro audio machines.

I have been asked about my measurements with the ADI-2 Pro FS compared to Audio Precision gear over the years. Obviously, I'm doing this as a hobby so the need for absolute calibrated accuracy is clearly not the same! What I'm aiming for are objective measurements that will allow me to verify expected behaviour of devices (good enough to cleanly capture impulse responses and filter settings for example), allow me to make relative comparisons (good enough test-retest reliability to graph device performance for comparisons), and find anomalies that might be audible. I have no reason to believe that the RME ADC fails in achieving these goals over the years of use.

For the sake of discussion, what I know is that the current Audio Precision APx555 B-Series uses the AKM AK5394A (datasheet here) ADCs up to 192kHz samplerate and above that, has Analog Devices ADCs for measurements up to 1MHz. The discontinued AK5394A I have been told has a flat noise floor for frequencies over 40kHz which is clearly desirable for producing measurements. Remember that the hardware is more than simply just the ADC chips used. The AP has precise autoranging, very clean analogue signal generators, low-noise notch filtering, comprehensive interfacing, and of course regularly updated modular software to tie these functions together. I won't reproduce it here, but you can see the block diagram for the "Performance Sine Analyzer Mode" (see slide 32) which includes parallel ADC steps, amplitude level changes, and the various analogue and digital filters used in the process. That's why you pay the big bucks for a calibrated AP machine, and the know-how support that comes with it - list price for the APx555 is US$28,300 when it came out in 2014, now in it's B-Series iteration. Audio Science Review has shown the 1kHz 1V loopback of the APx555; I'll have a look at something like this when I measure the "FS R BE".

Remember guys, there is such a thing as "good enough" resolution! Let's not simply chase numbers which for mature technologies like modern digital hi-res audio devices are beyond audibility with already achievable thresholds of low THD+N, high SINAD, nano/pico/femtosecond jitter, high channel separation, flat frequency response, etc. Measurements for me are to help verify high-fidelity performance and confirm claims of audible differences. As the results of the recent THD Blind Test show, subjective preferences do not always favour high-fidelity sound for any individual. For objectivists, this means we should not simplistically just compare numbers without reference to the context and relative to audibility. For subjectivists, the fact that the highest levels of THD did not result in optimal preference suggests that they should still look at the numbers because on average, non-linear distortion eventually does affect preference.

Don't forget that the RME ADC can record up to 768kHz and DSD256 (11.2MHz); way more than "needed" for any audio purpose I can imagine. Those who have a sweet spot for DSD might find vinyl rips using DSD128/256 to be very satisfying with this machine. I've personally used SoundIt! Pro for native DSD recordings and it sounds great. Try that on an Audio Precision... :-)

3. IR Multi-Remote-Control "MRC". This is the most obvious change/upgrade and I'm guessing the "R" in the name is for Remote?

RME, in making the ADI-2 DAC FS a few years ago with the RCA output and remote control seemed to signal a move into the high quality consumer space for them and now with the "FS R BE", I see them taking another step forward in this direction for those of us who have need of an ADC. I can now have a remote controlled, black (which integrates into the sound system better), compact-sized, very high quality DAC and ADC in my system for both playback of music (through an inexpensive USB streamer) with the convenience of simply connecting a few cables within seconds for ADC duties like vinyl ripping.

As I have described elsewhere, RME's design philosophy appears to be one of providing settings and options to the user as much as possible. This open, detail-oriented approach applies to the remote control as well.

I hate needing to use too many remotes in the soundroom and there will be those who will not like the plastic MRC. No problem, the MRC command codes can be found here. In fact, I've already programmed my old, long discontinued Logitech Harmony 1100 for the "FS R BE":

The old Harmony Remote Software 7 was able to find the RME brand listed under "Amplifier". I typed in the Model name "RME ADI-2 Pro FS R", followed some instructions to help the software identify the device (pressing the power and "1" buttons to teach the Harmony), and within a couple of minutes, the software was able to download the rest of the command set off the cloud. I was actually rather surprised with how quickly I got this working without having to manually teach every button to the learning remote.

Logitech Harmony 7 software. On the right, first page of the customized soft buttons for the most important features I'll use in the soundroom. Volume control already hard-coded to the remote's physical buttons.
There are in fact more commands that the "FS R BE" understands than available buttons on the MRC which you can assign on your universal remote. For example there are separate remote codes for the different DAC upsampling filter settings if you wish to fool around with that, direct access to set-up configurations, mono switch, etc...

Let's end off this preview...
So, until I have some time to run measurements, here's where I'll be employing the RME ADI-2 Pro FS R Black Edition for the next while:

Streaming music off my Roon server with another inexpensive Raspberry Pi "Touch" Streamer that I put together with my son the other day running RoPieee XL (will put up a quick updated "How-To" in the next week). Notice the diminutive size of the RME compared to the TEAC UD-501 and Oppo UDP-205 underneath. The relatively large 7" Pi Touch screen with the album cover and dancing amber RME spectrogram look great with the room lights off :-).

So far, what I said back in 2018 about sound quality of the "Pro FS" still stands today. Simply accurate, clean sound that shows off the best, and worst of the audio production through the quiet balanced XLR outputs. Full range frequency response, realistic tonality, highly dynamic, excellent transients, it beautifully conveys nuances, wide soundstage, and so on and so forth...

As a DAC, this device will not be the limiting factor in one's high-fidelity playback system. In fact, remember last week I talked about intermodulation distortion and had a download file for you to test out? Check this out:

On top is the FFT from the "perfect" digitally created synthetic test signal. Below are the 2 channels from the RME ADI-2 Pro FS R BE captured as loopback: computer --> USB in --> DAC analogue playback --> ADC analogue in --> computer.

As you can see, there's a tiny intermodulation signal down at <2.5kHz below -130dBFS. That is all. When/if you hear that "Star Trek tricorder" intermodulation distortion on your system, you can bet that it's not because of a DAC of this quality producing the distortion! Look to your preamp, amp and speakers/headphones to explain the audible anomaly. This is an example of what "accurate" looks like objectively; and sounds like by way of the "sound of silence" way below -100dBFS.

Obviously, measurements of the RME ADI-2 Pro FS R Black Edition will be coming in the weeks ahead as I listen more and run the machine through its paces.

Happy August everyone! I hope you're staying healthy and enjoying the music...

PS: What happened to InnerFidelity and AudioStream!? I certainly hope the contents were archived somewhere and can be recalled... Would be a shame to lose all that work and discussions.


  1. HI Archimago, can this device be use as a crossover for a 2 ways? Like Mitchco do with the Hilo Lynx.

    1. Hi Yan,
      I have not tried it myself since I don't have this kind of set-up but I think it should be doable.

      With USB disconnected (so you can change the driver mode), there's a "CC-Mode" (Class Compliant Mode) setting in SETUP - Options - Device Mode / DD - CC-Mode which you can change over from "Stereo" to "Multichannel". This will provide output to 4 analogue channels plus S/PDIF digital out and is also accessible in ASIO.

      From there, I'm guessing in JRiver which is what Mitch used in the article, should be able to select and play out as multichannel "Analog 1"-"Analog 4" for the 2-way.

      If anyone else has tried this and can confirm, I'd like to know as well!

    2. Two more things:

      1. In multichannel mode, maximum sample rate is 192kHz which is consistent with the limit of S/PDIF.

      2. To tap the analogue outs, 2 channels will be from the rear XLR/TS + 2 channels PH3/4 headphone out. Alternatively, perhaps more consistently, take the outputs from the 2 front headphone outs (PH1/2 + PH3/4).

    3. Thanks for taking time.

      I'm still not sure what I will buy. RME and Hilo cost a lot, miniDSP SHD cost less but I don't want to pay for Dirac (included in miniDSP's price) and Acourate.

    4. Note that I already have a miniDSP 2x4 HD, but I have high sensitivy speakers and there's an annoying hiss (No hiss before the miniDSP introduction).

    5. Unfortunate about the hiss, Yan. That's no good.

      If you get the RME, perhaps turning down the reference output to -5dBu (0.4Vrms) along with the better SNR of the DAC, you should be free from the hum with the sensitive set-up. Obviously dependent on the rest of the system like the amp. Plus you can upsample everything to 192kHz.

      I see ASR has results for the 2x4:

      Yeah, noise level would/should be substantially better with the RME.

    6. Yep 2x4 HD has higher noise floor.

      Amplifiers are from emotiva, but anyway there were no hiss before I add the miniDSP.

      I just want to make sure that I can route signal to the outputs I want with the RME.

      Thanks again.

  2. Hi Archimago

    Ah, good to read you bought also the Pro FS R. I bought also one last week and starting some measurements and later some listening. Had also the Pro FS (non R) before. And for a further AKM DAC comparison the Topping D90.

    Now it is Holiday time here, so will not do much in the next 2 or 3 weeks, but after that, will do. And yes, that new “Auto Range” feature of the this RME device is a very, very pleasant function, that I will also get into it.


    1. Nice hearing from you Juergen,
      That D90 with the AK4499 looks and I'm sure sounds sparkling accurate :-).

      I was actually considering that model when I got the DX3 Pro for its Bluetooth 5.0 abilities but the lack of convenient headphone out excluded this from using on my computer workstation, so I went with the less expensive option.

      I hope you and yours are keeping well! Enjoy the holiday time...

      I took some time off early last month and was at least able to spend a few evenings at a resort a few hours away. Massive reduction in travels here with the Canada-US border still restricted with no end in sight.

  3. All Innerfidelity URLs point to Stereophile instead, what the hell!? The wayback machine still have it stored, luckily…

    1. Hey Fluffy,
      Amazing that it all goes down just like that with essentially no warning other than the links going off from the Stereophile site a few days ago suggesting something's up.

      As a reader, while I appreciate that ultimately it's their (AVTech Media) choice, I think this is rather disrespectful considering that content included the comments and discussions from the public. At least give the readership a couple of weeks to consider if there's anything else they might want to read one last time. It'll let contributors archive their "final" posted articles as it looked before disappearing. Maybe review/download a few of Tyll's headphone measurements for one's own archives, etc...

      Seems a little hasty. Desperate even if there's primarily a financial component attached to this and they can't even keep the virtual lights on for a few more days with some clear warning to the readership!

      As you say, Wayback Machine can help. The June 29, 2020 snapshot of InnerFidelity looks pretty complete and I was able to see a few of Tyll's measurement PDFs:

  4. Direct link to the Pro infrared commands:

  5. After Tyll Hertsens left Innerfidelity was not the same. It is about the money and I am sure that Audiostream and Innerfidelity did not have the readership after Tyll left. With the headphone market being huge one might thing it could have had a long life, but Tyll's measurements were a big part of it. It is always about profitability for sure. I am not into steaming as I tried Tidal and Spotify and the quality was not there for me. Herb at Stereophile is still a big headphone guy so they will continue with some reviews.

    1. Hi Jim,
      Absolutely, I've always felt that the longevity of articles is tied to whether observations are published and made with objectivity in mind. As much as there may be a small segment of audiophiles who are attracted to purely subjective descriptions about how something sounds and maybe to the prose of some artistic writer, IMO, the value of this kind of reporting/reviewing is highly limited for technologically-based products.

      Tyll's measurements and thoughtful incorporation of the objective with the subjective along with his affable personality and intellect made InnerFidelity what it was (and the successful volume of hits). Can't say as much about AudioStream of course, but the last year or so has been simply dismal for both sites. Who knows, maybe the advertisers themselves demanded the sites be pulled down...

      Well, I see Michael Lavorgna is back from retirement with a few more articles on Twittering Machines. ;-)

    2. I have about 6 of Michael's recordings and they are all excellent. He could have done a nice job with either of those sites given the chance. thanks for the head up and I will check his site out again.

    3. Hi Jim,
      Are you talking about 6 of Lavorgna's recommended recordings? I didn't know he recorded any of his own music...

      I always thought he was given plenty of chances to be successful with those sites!

    4. So I went and took a look at Twittering Machines ... oh dear...
      Here’s Lavorgna reviewing the Intona USB isolator: “The mid bass was tighter with improved definition in the overall bass. Midrange clarity and focus seemed to be enhanced“
      I closed the tab.

    5. :-) Charles

      Yeah, he's back but so far just seems like more of the same with odd idiosyncratic beliefs and superficial reviews of things which generally follow along with manufacturer claims.

      Not sure I've ever seen his reviews go against manufacturer claims or not seem to "respect" price as correlating with a "better" sound.

  6. Right now I'm torn between getting the RME PRO (or DAC) versus something like the Topping D90/A90 combo. I personally like the look, build, and feel of the Topping product more. I really prefer only having one 2-in-1 dac/headphone amp device so the RME fits that criteria, but I don't feel like I would be using much of the other features like the DSP, which I would probably just use my computer itself for since that is what I use exclusively as a source.

    If the PRO could take microphone inputs I would get it in a heartbeat because it could also replace another one of my devices, the MOTU M4. The MOTU is another device I have with a bunch of features that I rarely make use of, I'm only using it as a one-channel mic preamp/ADC. One major gripe I have with the RME is that I wish it had an XLR headphone out and a built-in power supply.

    One thing I don't like about either is the patheticly sized volume knob, they all just feel so dinky compared to the older equipment I grew up using and even the DIY projects I still use to this day. Perhaps the upcoming Pre90 preamp from topping will cure that for me. Topping really should make a full-featured DX9 Pro (like dx7 pro but with the D90/A90 circuitry in a larger chassis) and that would be another contender.

    1. Decisions, decisions, eh? asdf ;-).

      As a pure high quality DAC + good headphone out, how about the Sabaj D5? I don't know much about the brand but the ASR measurements don't look bad and it's an ES9038Pro DAC-level performance.

  7. Can't wait for the Raspberry Pi "Touch" Streamer how-to. I've been thinking about it for a long time.

    1. Coming up this weekend Unknown.

      Inexpensive, easy to do in an evening and RoPieee has been excellent. Even tested up to 32/768 streaming to the RME.

  8. Hi! Thanks for this excellent indepth review. I don't think such precise level of details are available anywhere on internet..

    One question i wish to know about. I am planning to use this R Edition as recording interface and mastering. On DAC level it is having the latest chipset by AKM.

    However on ADC side, all competitive or premium mastering grade convertors use latest AKM 5578EN.
    While R Edition incorporates AKM 5574EN since previous versions.

    Being an expert, can you please describe if AKM 5578EN chip convertors are better than 5574EN?

    1. Hi Harsh,
      This is of course just the "preview". It's nice to show what's in the box and discuss the principles and what's new behind products before the actual "review" which will try to describe and objectify the limits of the machine in future posts.

      I'd certainly not call myself an "expert" as I think the word these days has been used to describe all kinds of things especially in the audio world. For one thing, I do this more for fun than as any professional pursuit! There are experts who I've been in contact with over the years working with the designs of these devices and know the machines inside out... Then there are reviewers who have examined and listened to all kinds of devices for years and in some way see their experience as giving them special insights into the sound of things - I suppose this also is a kind of expertise. Then there are actual professional recording and mastering folks and that also provides an expertise that's very different!

      In a world with so many options and ways of looking at things, there are countless things we can be experts as... I suppose in the measurements, listening, thinking about audio, and writing, the experience itself has provided some insights which I hope readers might find beneficial regardless of what one calls oneself!

      Now as for the ADC. Both AKM 5578 and 5574s are from the same generation of AKM devices. The main difference is that the 5578 is an 8-channel ADC vs. the 5574 with 4-channels.



      When used in a 2-channel ADC, the 5578 can perform an 8-to-2 channel summation mode that can improve potential SNR by 3dB (compared to the 5574's 4-to-2 channel mode) - 130dB vs. 127dB. The data sheet doesn't suggest any significant difference in distortion between the two however.

      Physically of course, this will correlate with a larger package for the 5578 and the power requirements 2x the 5574.

      As usual, the actual benefits will be dependent on the design around the ADC - this is why I said potential improvement. Since you'll be using the device for recording and mastering, I would suspect that the RME's resolution will surpass any analog recording you'll be doing!

  9. How does the RME compare to the TEAC UD-501? I know the TEAC’s headphone amp is a bit weak, but from an overall / DAC perspective?

    1. Yes Rufus, the RME's headphone output is clearly more powerful and of lower output impedance than the TEAC's. And then of course the ADC capability puts the RME in another class of devices.

      To be completely honest, as a DAC, I would not put a huge amount of weight on audible differences. I don't have the ability to A/B instantaneously between the 2 devices and I would certainly not bet on myself to be able to hear a difference in an actual blind test!

      Remember the results of this back in 2019:

      Those devices tested were way more unbalanced in terms of price and quality than 2 DACs capable of very good balanced output!

      Clearly as a more modern device with current feature-set, the RME is superior. Already shown in 2018, the hi-res tests show the RME AD-2 Pro FS to measure better than the TEAC:

      I'll leave it to you as to whether this might sound better to your ears. Go have a listen. :-)