>I am working on an orchestral project with 226 tracks, I was running this project with the focusrite at 1024 buffer size.
>I was hoping I could run it a bit lower with the babyface but anything lower than that and I have audio dropouts

Two remarks when comparing these two interfaces:

1. Focusrite Scarlett Solo is a (if I counted correctly) 6-channel interface (2x mono in, 2x stereo out). By this, less audio data (bandwidth) has to be transferred permanently over USB compared to a Babyface Pro FS, which is a 24 channel interface.

2. Even if you use the same ASIO buffer size for both interfaces, the resulting input and output latency might be different due to differences in the driver (USB transfer modes being used, amount of safety buffers, efficiency of the driver).
Careful when comparing the input and output latency in Cubase. Cubase gets the values from the ASIO driver.
But unlike RME not every vendor puts the correct values into the driver.
If you want to compare this, there is no other way than to measure the RTL using a tool called RTL_utility.
https://oblique-audio.com/rtl-utility.php
You need to plug a loop between an analog output and analog input.

But now, according to your problem.
From your screenshot, I can see only part of Cubase performance window (F12).
The interesting part on the right side, which can indicate overload situations in Cubase, is sadly hidden.
The performance bars look theoretically harmless, but this is only a snapshot, when you took the screenshot.
Problematic are in most cases, the peaks which we cannot see.

For comparison reasons, I used my "artificial" Cubase Test project with 400 tracks and 2 VST per track and in the sum. In total, 802 VSTs: https://www.tonstudio-forum.de/blog/Ent … cks-de-en/

The recording interface is an UFX III with even 188 channels.
My computer HW is 9 years old, but has been upgraded regular.
It now has Intel Xeon E5-1680v4 (8 cores from Q4 2016 (~7y old)
64 GB ECC RAM, Only SSDs inside (normal SSD, no NVMe), 2x 10GBit Network card, Nvidia RTX 4070

Rough comparison of our CPUs using the Passmark mixed benchmark delivering multi- and single-core performance values.

https://www.cpubenchmark.net/cpu.php?cp … mp;id=2869
Passmark multi-core performance value: 14054, Single Thread Performance: 2305

Your computers' CPU is more up to date:
https://www.cpubenchmark.net/cpu.php?cp … mp;id=5717
Passmark multicore performance value: 61095, Single Thread Performance: 4780

So … Your multicore performance is factor 4.3 and single thread performance factor 2 higher than mine.
Based on that, we can simply agree that you have much more performance than me.
But … these are numbers from Benchmarks, and for benchmarks the timing is not critical *when* the results arrive.
The big difference with audio processing is, that it has near-real-time performance requirements.

Your issue might have either one or maybe different root causes that pile up.
The only chance that you have is to check, step by step, whether your system has optimum settings for serving such an audio load with near-real-time requirements.

It could also be the case that the project organization in your DAW project is suboptimal.
Last recently Cubase got even the possibility for up to 16 inserts.
AFAIK every insert runs in one thread, so the more inserts you use, the longer the needed time to calculate audio only for this particular track. The more inserts/VST and the higher the CPU demands of each VST, the higher the demand for high single thread performance.

You should measure with LatencyMon, whether your system is setup well to serve an audio load.
Do this on a fresh booted PC and when no other applications are running, and no keyboard and mouse activity during measuring. It could be the case, that DPC latency is too high, that some driver are blocking CPU cores for too long, so that audio processes, that are scheduled to run of this core need to wait for too long until the drives detaches itself from this core.

I see that your Intel CPU has Power and Efficiency cores.
It could also be the case that audio related jobs are not using the Power cores.
You could use tools like Bitsums Process Lasso Pro to ensure, that audio related processes use the Power cores.
I recommend getting the lifetime license: https://bitsum.com/

There are already several threads in this forum which deal with the topic of audio drops / "performance problems".
I would like to ask you to search for them as it costs too much time to tell the same things again and again.
You can make a Checklist based on that information ensure that your BIOS and Computer settings are well for audio processing. And you can re-measure whether your DPC latency became better.

Here is an entry thread for you which contains further threads
https://forum.rme-audio.de/viewtopic.ph … 00#p214600

Or simply use the forum's extended search and search for the keywords "LatencyMon" "powermizer" or a combination of it and my forum name "ramses". Then you get a list of related threads.

This should give you a solid basis to check your system. If something is unclear feel free to ask.