Getting back into sampling, I sent the Amen break out of my MOTU interface, through a Behringer mixer, then back into the MOTU's line inputs. Inverting the new sample's polarity gives me a file that's about -33LUFS integrated, peaking -19dBFS, and has energy mostly around 10kHz. No processing should be affecting the sound i.e. every plugin and setting off or flat. The two signals are perfectly in-phase based on prior calibration and visual inspection.
What do I do with this information? Is it normal to have this loud of a file for gear that should otherwise ideally be flat? Is it normal for the difference to be biased toward high frequencies? -33LUFS integrated, sorry don't know the RMS, seems to be a big difference. What is acceptable for this kind of test?
EDIT: I just reread some of the stickies but I'm not sure I know the answers to my questions. What I got from it is that -40dB is faint - particularly for loud and relatively non-dynamic music - but -25 is obvious. I didn't relisten to all the examples but I do remember listening to them a few years ago. Am I on the right track here? With 400ms averaging, 2-4kHz is about -40dB and 10kHz is about -30dB. I'm guessing that I shouldn't be noticing much difference, in practice, but the resampled one does sound brighter.
Post by Michael Lawrence on Jan 5, 2020 16:46:00 GMT
Just so I'm clear on what the test was, you are comparing the original file to a file that went through your interface's DAC, through an (analog?) mixer, and then back through your IO's ADC? If so, any residual would be attributable to both conversions in addition to whatever the mixer is doing. Some residual up around 10 kHz could certainly be attributable to a less-than-pristine analog circuit design in the Berhinger mixer. What does the residual sound like, though? Is it hissy or "digital" sounding? Does it drift in and out, or remain constant in level?
The other thing to keep in mind is that a null test is extremely sensitive to level offset - if you're off by even 0.1 dB, you won't even be able to null down to -40 dB, so it's worth playing around with the levels in very small increments if you haven't done so already. Don't assume a unity-gain conversion, especially with an analog device in between - it's pretty unlikely to be the case. Let us know what you find out!
Hi, Mr. Lawrence - thanks for replying. Yes, the signal path is DAC->Mixer->ADC Line-In. I made a link to the wavs here but I'd describe the difference in the two files as one being brighter and less tight than the original. This is in alignment with another observation I made about MOTU's signature sound but, without doubt, the Behringer is likely also culpable.
While I did try to level match the two signals both while recording and after the fact, any difference I can achieve by massaging them further doesn't seem drastic.
I'm not sure why I did this test because I've previously noted by observing the resulting wavs that my loopback is changing the audio. Maybe it's because I'm listening through a bunch of Amen breaks which have been processed with various pieces of gear and wanted to know what 'my' sound color is. Ideally, I wouldn't have to use a mixer for loopback but I'm already planning to replace this MOTU and don't want to fiddle with the S/PDIF option.
EDIT: I did this test because I'm going through the Maschine manual and wanted to test out its line-in sampling.
Well, I got an RCA cable. Windows audio is now going out S/PDIF 1-2 and I can monitor it either directly or through my DAW. The good news is that REAPER also automatically sums to mono when assigning the two track channels to a single output - less routing.
Now, instead of monitoring everything through Main Out 1-2, I can basically send audio to each speaker individually including individual mono sums to my subwoofer and mixcube. I no longer need a hardware mixer! Too bad I didn't do this before buying this Behringer.