1 Topic by luis_! (edited by luis_! 2021-08-30 21:43:35)

  • Registered: 2021-08-30
  • Posts: 2

Topic: Babyface Pro performing inferior to Steinberg UR 22..?

Hi all,

I just got a used Babyface Pro (no FS - which I only realized after having it for two hours big_smile) and I'm not sure what to make of it.

Before, I used a Steinberg UR22, and the sound of the Babyface is, as hoped for, a lot better - unlike the latency and the overall performance. I'm normally working with a 128 sample buffer size, and the Steinberg managed most of my projects really well. Now with the Babyface I have to increase to 512 samples to get similar results.
The DSP performance looks similar between both interfaces on the same sample rates (you can view that live in Bitwig), but on the Babyface there are occasional spikes, which are audible as cracks/pops. After trying it for a few minutes with one of my larger projects, the Babyface actually crashed and wouldn't give out audio, even though there was a moving level meter on the device.
Also I compared both interfaces with a complex Omnisphere preset. The Babyface had occasional cracks with 128 samples buffer, while the UR 22 didn't have any issues at 96 samples - what?

My system is up to date, I'm using a Lenovo Legion 5 with an SSD running Windows 10, a Ryzen 7 4800 H and an Nvidia RTX 2060. My USB Ports are all 3.2. I am using the ASIO Fireface USB Driver. I updated the Babyface's firmware to the current version.
My audio settings are 44.1 kHz with - normally - 96-128 samples.

I have tried turning off my Wifi card and setting the Babyface to not be the standard audio device (was mentioned in the manual as a potential issue). Both didn't change anything.

I have been reading through a lot of posts here on this forum but nothing really gave me an insight as to explain why this device, aimed at the professional customer, would perform so inferior to a really basic entry level interface.
While the interface is definitely usable at 512 samples, that definitely isn't that magic performance boost that everybody here is always raving of, and which got me to get an RME interface as well.

Without changing anything about the system, the UR 22 just performs better than the Babyface. Does anybody have an idea about potential bottlenecks? Or did I just buy a broken interface?

Thank you in advance,
Luis

(Edited for clarification)

2 Reply by MC

  • MC
  • MC
  • Administrator
  • Offline
  • From: Germany/Thailand
  • Registered: 2007-04-06
  • Posts: 26,226

Re: Babyface Pro performing inferior to Steinberg UR 22..?

The UR22 has only 2/2 channels to transmit via USB. The BF Pro has 12/12 channels to transmit. With USB not performing optimally this might already explain the differences that you see. If you search this forum you will find people often have problems with AMD USB. Solutions range from latest BIOS updates up to using a PCIe to USB card to make it work.

Regards
Matthias Carstens
RME

MC's Website

  • vinark
  • DAW Mastermind
  • Offline
  • Registered: 2009-06-05
  • Posts: 3,022

Re: Babyface Pro performing inferior to Steinberg UR 22..?

And what latencies does cubase report for the same setting on both interfaces?

Vincent, Amsterdam
https://soundcloud.com/thesecretworld
BFpro fs, 2X HDSP9652 ADI-8AE, 2X HDSP9632
  • ramses
  • DAW Mastermind
  • Offline
  • From: UTC+1
  • Registered: 2012-10-14
  • Posts: 14,261

Re: Babyface Pro performing inferior to Steinberg UR 22..?

vinark wrote:

And what latencies does cubase report for the same setting on both interfaces?

Good point, usually RME has a lower RTL (Round Trip Latency, Input+Output ASIO latency) when comparing at the same sample rate and ASIO buffersize. So it could be, that the RTL is similar even if you use 512 buffers with RME.

BR Ramses - UFX III, ADI-2 Pro FS R BE, M-1620 Pro D, 12Mic, XTC, RayDAT, Win10, X10SRi-F, E5-1680v4
  • Registered: 2021-08-30
  • Posts: 2

Re: Babyface Pro performing inferior to Steinberg UR 22..?

The RTL is 4 ms for 64 samples and 13 ms for 256 samples on the RME, can't measure on the UR 22 right now. I'm pretty sure the latency values in Bitwig (my DAW) were similar.

LatencyMon gave pretty good results and said, my PC should be able to run real time audio processing.

I managed to hugely improve the situation by changing my Windows Power Plan settings. There was a setting allowing the USB ports power to be throttled or something like that.

I will still try to streamline my system more to get the best possible performance, looking at a BIOS update now.

A PCI-Express card is not a viable option for a laptop, if I'm not mistaken..?

Thanks for your answers!
Luis

6 Reply by ramses (edited by ramses 2021-09-02 04:54:56)

  • ramses
  • DAW Mastermind
  • Offline
  • From: UTC+1
  • Registered: 2012-10-14
  • Posts: 14,261

Re: Babyface Pro performing inferior to Steinberg UR 22..?

> LatencyMon gave pretty good results and said, my PC should be able to run real time audio processing.

How long did you test. What is "pretty good" ? Could you identify e.g. drivers causing higher DPCs ?

> I managed to hugely improve the situation by changing my Windows Power Plan settings. There was a setting allowing the USB ports power to be throttled or something like that.

You need to use a power plan for full performance. Cubase can be set to use a power plan of its own which also disables CPU core parking (which also cases higher DPC), at least on a desktop PC.

> A PCI-Express card is not a viable option for a laptop, if I'm not mistaken..?

A laptop is flat, it doesn't offer PCIe slots on the mainboard.

Only Thunderbolt is external PCIe but then you need either a thunderbolt recording interface (UFX+). or a thunderbolt based extension cabinet with one or more PCIe slots which would allow for another USB card under the assumption that the chipset of your laptop causes USB issues.

This would be an expensive solution, no guarantee, that this works / works better. But AFAIK AMD based laptops usually do not have a thunderbolt option anyway.

BR Ramses - UFX III, ADI-2 Pro FS R BE, M-1620 Pro D, 12Mic, XTC, RayDAT, Win10, X10SRi-F, E5-1680v4
  • Registered: 2025-07-04
  • Posts: 7

Re: Babyface Pro performing inferior to Steinberg UR 22..?

Hi Luis!

Did you found a solution? I own an UR22MK2 and wanted to buy a babyface pro fs because i want less latency. So i am afraid to read about your story.

Feedback appreciated!!

Alfonx

I produce Dance, House, Pop and sometimes Techno smile
https://soundcloud.com/alfonx/sets/alfonx-tracks
  • mkok
  • DAW Mastermind
  • Offline
  • Registered: 2020-11-04
  • Posts: 701

Re: Babyface Pro performing inferior to Steinberg UR 22..?

I have the Babyface pro fs (this thread is about the non fs unit) I have no problems at all on a 5 year old i9 PC. On small projects I run at 64 buffers with the best latency I’ve had from any device. On large projects I run at 128 buffers which has lower latency than a lot of interfaces at 64 buffers.

Babyface Pro Fs, Behringer ADA8200, win 10/11 PCs, Cubase/Wavelab, Adam A7X monitors.
  • Kubrak
  • DAW Mastermind
  • Offline
  • Registered: 2023-08-03
  • Posts: 918

Re: Babyface Pro performing inferior to Steinberg UR 22..?

I run Babyface Pro fs on many years old i5 (10th gen.) with buffersize 64 samples... And no problem.

FF UCX II, Digiface USB, Babyface Pro FS

10 Reply by ramses (edited by ramses 2025-07-04 17:05:29)

  • ramses
  • DAW Mastermind
  • Offline
  • From: UTC+1
  • Registered: 2012-10-14
  • Posts: 14,261

Re: Babyface Pro performing inferior to Steinberg UR 22..?

These are the RTL which you can expect from an RME recording interfaces on Windows at 44.1 kHz:

https://www.tonstudio-forum.de/blog/ent … cts-en-de/

Even with an ASIO buffer size of 128 samples, I am still below 10ms RTL, in my case 7,3 ms with UFX III.

It's a different driver (MADIface) but values should be similar,
except that a buffer size of 32 samples is not possible with the USB driver, only with the MADIface driver.

BR Ramses - UFX III, ADI-2 Pro FS R BE, M-1620 Pro D, 12Mic, XTC, RayDAT, Win10, X10SRi-F, E5-1680v4
  • Registered: 2025-07-04
  • Posts: 7

Re: Babyface Pro performing inferior to Steinberg UR 22..?

Thank you ramses, Kubrak, mkok! smile

Yes, that is what i had expected when it comes to latency and RME.
In the beginning of a new project (Ableton live suite 12 on intel i7 8700k, 32GB RAM) i usually use 128 samples 44.1kHz.
(energy power plan "high/max")

But then later i have 60 Tracks in the Project (lots of plugins,...) then i have to switch to 512 samples. This is no big problem but if i want to add melodies via keyboard the "latency of death" is here.

I know i can choose "low latency while monitoring" in Ableton but however my via midi recorded notes are often out of sync. So i had the idea of buying something new - like the RME babyface.

So i have exact the same approach like luis and he is not happy with the results. So i would be unhappy too.
Of course i can buy, test  it and send it back ... so i can test it on my own.

Greets and beats!
Alfonx

I produce Dance, House, Pop and sometimes Techno smile
https://soundcloud.com/alfonx/sets/alfonx-tracks

12 Reply by ramses (edited by ramses 2025-07-04 17:45:56)

  • ramses
  • DAW Mastermind
  • Offline
  • From: UTC+1
  • Registered: 2012-10-14
  • Posts: 14,261

Re: Babyface Pro performing inferior to Steinberg UR 22..?

Can you freeze the tracks to reduce the CPU load?
When freezing tracks you render it, play it as wav file, but still can pan or reduce volume in the mix.
Thats the usual way of handling a lot of tracks / plugins once the plugin settings are set ...

BR Ramses - UFX III, ADI-2 Pro FS R BE, M-1620 Pro D, 12Mic, XTC, RayDAT, Win10, X10SRi-F, E5-1680v4

13 Reply by Alfonx (edited by Alfonx 2025-07-04 18:10:50)

  • Registered: 2025-07-04
  • Posts: 7

Re: Babyface Pro performing inferior to Steinberg UR 22..?

ramses wrote:

Can you freeze the tracks to reduce the CPU load?

Yes, you are right! i could, perhaps i should do this. But freezing 50 Tracks (e.g. without the melodies) is really annoying and time consumption.  But - yes, this culd be one solution. I should really try it in my current project.

Of course its just one time "time consumption" but then they are all freezed and this should be the benefit. I'll try this..

I produce Dance, House, Pop and sometimes Techno smile
https://soundcloud.com/alfonx/sets/alfonx-tracks

14 Reply by ramses (edited by ramses 2025-07-04 18:36:12)

  • ramses
  • DAW Mastermind
  • Offline
  • From: UTC+1
  • Registered: 2012-10-14
  • Posts: 14,261

Re: Babyface Pro performing inferior to Steinberg UR 22..?

Additionally problematic is when using multiple VST in a track as insert because they have to be processed by one CPU core.
This is a time-consuming task and needs a CPU with high single thread performance.
Especially in such situation, freezing can help a lot.

BR Ramses - UFX III, ADI-2 Pro FS R BE, M-1620 Pro D, 12Mic, XTC, RayDAT, Win10, X10SRi-F, E5-1680v4
  • Registered: 2025-07-04
  • Posts: 7

Re: Babyface Pro performing inferior to Steinberg UR 22..?

ramses wrote:

Additionally problematic is when using multiple VST in a track as insert because they have to be processed by one CPU core.
This is a time-consuming task and needs a CPU with high single thread performance.
Especially in such situation, freezing can help a lot.

Hi Ramses!
I think i have the solution, it’s kind of silly…
Yes, i do have 60 tracks, synths and so.
But the main problem is ozone on the master with a huge amount of latency. If i delete the ozone plugin from the master chain the latency problem is gone.

Probably this was the same Problem that luis had. Its not the sound device its ableton/plugin latency problem.
It’s a pity you can not „really“ turn off plugins in Ableton projects. But i can save my ozone s preset and load it when i need it.

So - i think i can spend my money on other things than that rme babyface

I produce Dance, House, Pop and sometimes Techno smile
https://soundcloud.com/alfonx/sets/alfonx-tracks
  • Kubrak
  • DAW Mastermind
  • Offline
  • Registered: 2023-08-03
  • Posts: 918

Re: Babyface Pro performing inferior to Steinberg UR 22..?

Generally, it is not problem of plugins and/or interface. It is problem of computing strenght of computer.

And beside that, also workflow organisation thing. It is recomended to have mastering as totally separate project. Just two rendered tracks (if working in stereo) and mastering chain. And no mastering plugins in mixing project. It has two benefits, less CPU is needed for the both, mixing and mastering. And mixing and mastering may be separated in time and space and done in two consequentive steps.

FF UCX II, Digiface USB, Babyface Pro FS

17 Reply by ramses (edited by ramses 2025-07-05 11:44:56)

  • ramses
  • DAW Mastermind
  • Offline
  • From: UTC+1
  • Registered: 2012-10-14
  • Posts: 14,261

Re: Babyface Pro performing inferior to Steinberg UR 22..?

Alfonx wrote:
ramses wrote:

Additionally problematic is when using multiple VST in a track as insert because they have to be processed by one CPU core.
This is a time-consuming task and needs a CPU with high single thread performance.
Especially in such situation, freezing can help a lot.

Hi Ramses!
I think i have the solution, it’s kind of silly…
Yes, i do have 60 tracks, synths and so.
But the main problem is ozone on the master with a huge amount of latency. If i delete the ozone plugin from the master chain the latency problem is gone.

Probably this was the same Problem that luis had. Its not the sound device its ableton/plugin latency problem.
It’s a pity you can not „really“ turn off plugins in Ableton projects. But i can save my ozone s preset and load it when i need it.

So - i think i can spend my money on other things than that rme babyface

It's recommended to separate mixing and mastering — this also helps reduce the number of plugins involved.

BTW .. for the same reason, I no longer use Ozone; it introduces too much processing overhead.

You might want to try bx_masterdesk instead.
It's an excellent tool that's easy to use — and currently available at a very low price.
https://www.plugin-alliance.com/en/prod … rdesk.html

There are also more advanced versions available including true peak limiter and a pro version.
https://www.plugin-alliance.com/en/prod … _peak.html
https://www.plugin-alliance.com/en/prod … k_pro.html

The "pro" version I do not like so much, because the display contrasts are not so nice and too many controls.
The other two versions give me better control / overview.

From ease of use perspective I would recommend bx_masterdesk or maybe better the True Peak version
which has a true peak limiter integrated which could be very useful.

The Plugin Allicance installation manager makes the download/installation process of pa plugins super easy.

BR Ramses - UFX III, ADI-2 Pro FS R BE, M-1620 Pro D, 12Mic, XTC, RayDAT, Win10, X10SRi-F, E5-1680v4
  • mkok
  • DAW Mastermind
  • Offline
  • Registered: 2020-11-04
  • Posts: 701

Re: Babyface Pro performing inferior to Steinberg UR 22..?

The basic rule is if you want to play a vst instrument with low latency then don’t have any processing in the path, either on the vsti channel or the master. I’ve seen this so many times on the Cubase forum. Definitely worth mentioning again.

Babyface Pro Fs, Behringer ADA8200, win 10/11 PCs, Cubase/Wavelab, Adam A7X monitors.
  • Registered: 2025-07-04
  • Posts: 7

Re: Babyface Pro performing inferior to Steinberg UR 22..?

mkok wrote:

The basic rule is if you want to play a vst instrument with low latency then don’t have any processing in the path, either on the vsti channel or the master. I’ve seen this so many times on the Cubase forum. Definitely worth mentioning again.

Yes. But in general it’s running good with my projects except this situation with ozone on the master. I have a full project and of course i want to simply add something here and there and usually everything is working in this situation. I just forgot this one plugin that causes a big latency.

I produce Dance, House, Pop and sometimes Techno smile
https://soundcloud.com/alfonx/sets/alfonx-tracks
  • Registered: 2025-07-04
  • Posts: 7

Re: Babyface Pro performing inferior to Steinberg UR 22..?

Hi Ramses,
Hi Kubrak,
i know the way of separating mixing and mastering. But in 2025 you have so much possibilities, so why not having a mixed project and easily switch on/off the masterprocessing? So i do. So i can change one track if i hear its messing up something.
I just forgot to have a look at the latency value of my plugins.

@ramses: Thank you, i already have a lot of PA. And sure the masterdesk would be very useful. But i am really really glad to have ozone, i definitely don’t want to dive deeper in this mastering stuff!! I already gave some songs away for mastering but it’s crazy expensive and it’s just my hobby. So i decided to take a step back and use ozone. It’s okay for me. Really. I am no audio engineer and don’t want to be one.

I produce Dance, House, Pop and sometimes Techno smile
https://soundcloud.com/alfonx/sets/alfonx-tracks

21 Reply by ramses (edited by ramses 2025-07-06 09:30:27)

  • ramses
  • DAW Mastermind
  • Offline
  • From: UTC+1
  • Registered: 2012-10-14
  • Posts: 14,261

Re: Babyface Pro performing inferior to Steinberg UR 22..?

Alfonx wrote:

Hi Ramses,
Hi Kubrak,
i know the way of separating mixing and mastering. But in 2025 you have so much possibilities, so why not having a mixed project and easily switch on/off the masterprocessing? So i do. So i can change one track if i hear its messing up something.
I just forgot to have a look at the latency value of my plugins.

@ramses: Thank you, i already have a lot of PA. And sure the masterdesk would be very useful. But i am really really glad to have ozone, i definitely don’t want to dive deeper in this mastering stuff!! I already gave some songs away for mastering but it’s crazy expensive and it’s just my hobby. So i decided to take a step back and use ozone. It’s okay for me. Really. I am no audio engineer and don’t want to be one.

Hi Alfonx,

I think there is a misunderstanding on your side. bx_masterdesk is IMHO much easier to use compared to Ozone and at the same time doesn't spoil so many computer resources.

I got this tool recommended from a Jazz guitarist, who has a lot of experience mastering his recordings.
The design goal of this tool is simplicity.

It was a well-intentioned recommendation on my part, and at $29.99 (€25.46), it doesn't cost much either.

I use the tool successfully in every recording and am very satisfied with it because it does so many things right straight away.

The user manual is also very well written and provides an excellent overview of how to work with this tool in just a few pages.

https://www.plugin-alliance.com/en/prod … manual.pdf

BR Ramses - UFX III, ADI-2 Pro FS R BE, M-1620 Pro D, 12Mic, XTC, RayDAT, Win10, X10SRi-F, E5-1680v4

22 Reply by ramses (edited by ramses 2025-07-06 09:30:37)

  • ramses
  • DAW Mastermind
  • Offline
  • From: UTC+1
  • Registered: 2012-10-14
  • Posts: 14,261

Re: Babyface Pro performing inferior to Steinberg UR 22..?

In addition, some measuring which could be of interest to you.

I just tested bx_masterdesk true peak with my 400 track test project with 803 VSTs, see also here in the blog:
https://www.tonstudio-forum.de/blog/Ent … cks-de-en/

Cubase 14.0.10: Frequency and Compressor in each of the 400 tracks -> 800 VST
A few VSTs in the sum: Multiband Compressor, Magneto II, bx_masterdesk, in total 803.
Projects sample rate: 44.1 kHz, ASIO buffersize: 32 samples (minimum).
Result: also with bx_masterdesk in the stereo sum no audio drops during playback.

There was even a Macrium Reflect Backup running in the background creating a disk image which increases CPU load and I/O.

The screenshot now after the Macrium backup, so that the computer load reflects only the DAW load (see the CPU and memory usage meter at the upper right).

https://www.dropbox.com/scl/fi/fwnc89me … d&dl=1

https://www.dropbox.com/scl/fi/fwnc89meabvluygu32gaq/2025-07-06-400-tracks-load-test-803-VST-backup-running-bx_masterdesk.jpg?rlkey=nmf7re3mi1jm01ptaw7hw7p56&st=p8k0iohd&dl=1

Memory consumption: Cubase 14.7 GB, Total: 24.3 of 64 GB.

Installation of Win10 22H2, system setup, LatencyMon Results, etc documented here:
https://www.tonstudio-forum.de/blog/ent … lts-en-de/

The system is basically from 2014, so 11 year old in total, but got a lot of upgrades (CPU, GPU, RAM, 10GBit LAN, etc).
The CPU is from 2016, I got it cheap from eBay years later for ~€160. Otherwise it would have been too expensive (~€1750).

If you want to compare the CPU performance with your CPU, for this purpose the passmark mixed benchmark is nice:
https://www.cpubenchmark.net/cpu.php?cp … mp;id=2869
The you see Multithread and also Single Thread rating.

BR Ramses - UFX III, ADI-2 Pro FS R BE, M-1620 Pro D, 12Mic, XTC, RayDAT, Win10, X10SRi-F, E5-1680v4

23 Reply by Alfonx (edited by Alfonx 2025-07-06 12:32:47)

  • Registered: 2025-07-04
  • Posts: 7

Re: Babyface Pro performing inferior to Steinberg UR 22..?

ramses wrote:


I think there is a misunderstanding on your side. bx_masterdesk is IMHO much easier to use compared to Ozone and at the same time doesn't spoil so many computer resources.

Okay sorry-yes, i should have a closer look at it. I own bx masterdesk, but not bx masterdesk true peak (have collected plugins for no reason years ago - but in the last years i changed my view and was only buying i definitely needed).
Perhaps bx masterdesk is okay for my „preview“ approach. And if i am finally satisfied with the mix i swap masterdesk with ozone. I already tested the ableton live limiter, really cool and efficient!!

I am looking at your second posting - jesus!! 400 tracks, 800 vsts… these are big numbers. I guess you do orchestral musik?

I produce Dance, House, Pop and sometimes Techno smile
https://soundcloud.com/alfonx/sets/alfonx-tracks

24 Reply by ramses (edited by ramses 2025-07-06 13:46:47)

  • ramses
  • DAW Mastermind
  • Offline
  • From: UTC+1
  • Registered: 2012-10-14
  • Posts: 14,261

Re: Babyface Pro performing inferior to Steinberg UR 22..?

Alfonx wrote:

Okay sorry-yes, i should have a closer look at it. I own bx masterdesk, but not bx masterdesk true peak

bx_masterdesk is sufficient. I've been using it for a long time — I only picked up the TP version because it was on special offer.

Alfonx wrote:

(have collected plugins for no reason years ago - but in the last years i changed my view and was only buying i definitely needed).

I think this is where most of us started. There are only a few I still use regularly — this is one of them.

Alfonx wrote:

Perhaps bx masterdesk is okay for my „preview“ approach. And if i am finally satisfied with the mix i swap masterdesk with ozone. I already tested the ableton live limiter, really cool and efficient!!

Use whatever works best for you. I stopped using Ozone due to its high CPU load — I really disliked that. What also put me off was that all my presets suddenly vanished, and I couldn’t find any way to restore them. After that, I saw no reason to keep using the software.

Alfonx wrote:

I am looking at your second posting - jesus!! 400 tracks, 800 vsts… these are big numbers. I guess you do orchestral musik?

No, it’s an artificial benchmark. When optimizing a PC, you need a way to verify whether DAW performance has actually improved. So I tested how many tracks I could run with two VSTs inserted per channel before audio dropouts occurred. Then I scaled it back slightly to land at a round number — 400 tracks.

I am playing guitar and like recording as a hobby. Initially I played usual cover songs and created some own songs, but it turned out to be too time consuming. For this reason I make most of the time overdubs to existing records that I like and which challenge me and give me some room to play. This is also good for me, not so sit too long in front of the PC. You surely know how much time you are spending otherwise on your own songs when mixing and mastering.

Besides guitar playing I also had a clear focus on optimizing my recording setup and the PC which is also big fun to me.
I gained much experience in the past 27 years and finally achieved an excellent setup in the past ~10 years using RME components.

This setup works reliably for both single and double sample rates (44.1 and 96 kHz). It serves as a solid benchmark and a kind of sanity check after system changes to confirm everything still runs smoothly.

Achieving this load on my system requires that the CPU cores aren’t being blocked by poorly written drivers and that system settings are properly optimized. Otherwise, the system simply can’t respond fast enough under high DAW stress.
That’s one of the main reasons I went with server/enterprise-grade components this time — mainboard, CPU, etc.
With my last consumer grade mainboards I had issues and I hoped to get this way better components.

On a 'weaker' system — not necessarily in terms of raw CPU power, but in terms of efficiency and low DPC latency — you could compensate with a larger ASIO buffer (what I had to do), but that comes at the cost of increased USB audio latency, which is clearly audible.

BR Ramses - UFX III, ADI-2 Pro FS R BE, M-1620 Pro D, 12Mic, XTC, RayDAT, Win10, X10SRi-F, E5-1680v4