Qu-Drive Recording Level?

[[XAP]BobParticipant]

I’ve got my QU out for a gig tomorrow, so I’ll try and run a test tonight.

[Anonymous]

I struggled with *low looking level files* from USB QU Drive and I think I wrote a post somewhere about this a year or so ago and from memory someone shot me down with a machine gun?

So I only use USB Streaming where I can push files up in the display of my Mac to represent what I am seeing from my desks’. Always aware of noise floor.
And I only “very” rarely use normalisation.
However I have come from the 16 bit world.
There a lot of info on this subject https://www.google.co.nz/search?q=normalisation+or+not+to+normalise+audio&ie=utf-8&oe=utf-8&gws_rd=cr&ei=rnpfVYHsPImI8QXToYDACg
and also different levels of standards where zero is
https://en.wikipedia.org/wiki/DBFS
I guess I’m in the other camp where I would prefer to see wave files in a more normal analogue veiw.
A nice healthy veiw rather than having to zoom into see dribble and then having to work with that.
The last thing I want to do before working on files is to normalise istly…
I must admit the QU floor noise series is pretty clean!

[[XAP]BobParticipant]

Erm – I just did some tests …

I took a 1kHz sine wave at -20dB out of Mix1 – routed to Alt-Out

Alt Out looped back to Input 7, preamp gain set to 0dB, dialled the Alt-Out volume control to measure 0dB on the preamp meter according to the iPad (better scale markings) – that was “about 9”, not by the tag.

I raised the preamp gain, and the metered level agreed with the Preamp changes (give or take a dB) until the peak light lit – the meter was still below the peak – for another

I had the main output send out to my HiFi amp, so was listening to the clean sine wave…
The peak light came on at 36dB gain, at 39dB the whole meter was taken up – the sine wave was still quite clean.

Increasing the gain further was clean until harmonics at around 47dB preamp gain.

However the *only* wav file the mac couldn’t copy was channel7…
I’ll try and get it off via Linux, but the mac is having a “computer says no” moment.

Channel14 (in which I had an SM58) recorded nicely, and I’ve bumped that by ~30dB in audacity – perfectly clean, I can hear the music my wife was playing at the far end of the house…

If I can get the files somewhere sensible I’ll post the link here…

[[XAP]BobParticipant]

OK – just done another series of tests, and the corruption hasn’t recurred, but these were all shorter and less excess running – Audacity indicates clipping in them.

Stereo Flac file: Left channel is the channel into which I fed the SigGen, Right channel is me mumbling/rambling at the desk – the mic was several feet away, pointing away from me (don’t ask).
[audio src="https://www.dropbox.com/s/ttih6jaqvezmf2j/QURecordingLevelTest.flac" /]

The levels on the left channel are unchanged, the right channel has been boosted by up to 30dB (there are three sets of files here, the first track needed 30, the next two were a little less).

No other processing on the files – so the levels shown on a DAW are what the levels are…

[[XAP]BobParticipant]

Running through the above file…

0 on the input PreAmp -> -20 metered -> -41 to -38 on Audacity
10 on the input PreAmp -> -10 metered -> -32.5 to -29.5 on Audacity
20 on the input PreAmp -> 0 metered -> -23 to -20 on Audacity
30 on the input PreAmp -> 10 metered -> -12 to -9 on Audacity
35 on the input PreAmp -> 15 metered -> -6 to -3 on Audacity
36 on the input PreAmp -> Peak Light -> -5 to -2 on Audacity
37 on the input PreAmp -> -> -4 to -1 on Audacity
38 on the input PreAmp -> Sounds OK, Audacity says it’s clipping
39 on the input PreAmp -> Audible distortion, Audacity says it’s clipping

[[XAP]BobParticipant]

Note that the audacity levels are read off the side of the waveform (dB) chart well zoomed in.

[AndreasModerator]

Thanks Bob!
So, from your tests this all makes sense for me:
Sending a -20 dB signal to an input needs to be rised by +20dB input gain to get 0dB readout which is recorded at about -18dBFS (-23..-20dB from audacity).
That’s what I had in mind and expected things to work.

[[XAP]BobParticipant]

And zero dB on the meter corresponds with -18dBFS (given that 38dB is marked as clipping, but isn’t audibly so – I assume it is a full scale sine wave at that point)

The output level on the Alt-Out surprised me, because the tag isn’t at “unity” – i.e. a -20dB signal pushed through a Mix at 0dB and into a Channel with 0dB gain isn’t -20dB measured until the Alt-Out level is at ~9/10, when the big tag which I assumed to be unity is at 7/10…

[StonepianoParticipant]

I have now switched input metering in Logic to pre-fader, helps a lot in determining the actual recorded track levels (amongst other benefits).
The reason for my levels being quite low, was simply the poor resolution of the Qu-Drive metering. The home meter screen is crap, also. Never look at those any more- the input processing screen and Qu-Pad give you a much more accurate view of your levels.

B.R.
T.K.

[volounteerParticipant]

You can see the red lines at every clip point in audacity.

As I recall audacity called touching 0dBFS as a clip.
It assumes that in the gaps there will be peaks which is erroneous.

There are NO INTERSAMPLE PEAKS UNTIL YOU CONVERT TO ANALOG.

Talking about them in the digital domain is pure nonsense although way too common especially on the internet.
And if you lower the gain before you convert there is no clipping in the analog world.

My rule of thumb was to keep everything at -12dBFS at all times.
If you want louder turn the knob on the power amp before the speakers to the right.

With 0 on the Qu being equal to -18dBFS I have an extra 6 dB even before my personal safety margin,
so that touching the top light on the Qu does not bother me.

You could keep it all a LOT lower and still have no noise problems.

I am an EE and did digital in graduate school.
There is a lot of nonsense out there which is ‘common knowledge’ that is inaccurate and misleading.

[Author]

Posts