Signal Levels is this what is expected...

simonp54

Experienced
I was playing around this morning and wanted to confirm this is "expected behaviour"

first scenario... inject a sine wave 1000Hz -20dBFS = output = -20dBFS... so far so good....

upload_2015-12-5_10-9-52.png

Now if i add this...
upload_2015-12-5_10-11-13.png
the output is -14dBFS.... again as I expected... (signal is twice as loud)

however... when i do this...
upload_2015-12-5_10-12-9.png
output is -10.4dBFS... not what i expected. I would have thought -8dBFS would have been correct here?

it also continues...
upload_2015-12-5_10-13-27.png
output = -7.9dBFS... would have thought we would be at -2dBFS by now?


so its a kinda "non musical" question... but i was surprised so thought I would ask...

Cheers
 

Attachments

  • upload_2015-12-5_10-8-44.png
    upload_2015-12-5_10-8-44.png
    178.9 KB · Views: 6
ok so "adding signal rows... the dB increase will get smaller each time"... does that mean that each feed is still at an equal level in the output (aka from a mix perspective)?

i.e. in my 3 to 1 example (the input signal being 3 instances of -20dBFS) the output is constructed of 3 signals of 3.2dB effective increase?

i.e. -20 (in 3 times) -10.4 (out) (3 sets of 3.2dB increase?)

and in the four example
-20 (in 4 times) -7.9 (out) (4 sets of 3.025dB increase?)

what im trying to make sure is that the output of the "mix point" has a consistent make up of the input signals... im not really bothered about the up lift... i can adjust that back with the "level of the output of a mixer for instance...
 
ok so "adding signal rows... the dB increase will get smaller each time"... does that mean that each feed is still at an equal level in the output (aka from a mix perspective)?

i.e. in my 3 to 1 example (the input signal being 3 instances of -20dBFS) the output is constructed of 3 signals of 3.2dB effective increase?

The rows will all have the same level.

The "X sets of N dB increase" thing isn't an accurate way to describe what's happening. Rows 2, 3, 4 add around 6, 3.5, 2.5 dB respectively. This is because dB is describing a ratio.
 
Don't think in terms of dB. Assume you are putting 1V in (for convenience). If you add two rows you get 2V out. That's a 6 dB increase. If you add three rows you get 3V out. That's a 9.5 dB increase. If you add four rows you get 4V. That's 12 dB.
 
Ah....this thread takes me back to acoustics courses in graduate school, where we'd have word problems such as if one fan is 27dB, how loud would 4 fans be ?

Should of taught the course in relation to Axe FX grids and then maybe I'd have stayed awake better lol (I did end up with an A though just the same)
 
Speaking of dB, image the fun trying to explain to the average lay patient their audiogram results (from a hearing exam) which charted in the dB scale, so thusly log.

If their hearing threshold went from lets say 30 dB to 40 dB for a given frequency, they see it as no big deal. They think of it in terms of their tv remote control volume, where increasing the volume level from number 30 to number 40 isn't a big change, as its a linear scale. Just try explaining the concept of log values to a 85 year old patient if you really want to pull your hair out.
 
Ah....this thread takes me back to acoustics courses in graduate school, where we'd have word problems such as if one fan is 27dB, how loud would 4 fans be ?

And even more confusing re. the OP the answer isn't 27 + 12 = 39.
 
Don't think in terms of dB. Assume you are putting 1V in (for convenience). If you add two rows you get 2V out. That's a 6 dB increase. If you add three rows you get 3V out. That's a 9.5 dB increase. If you add four rows you get 4V. That's 12 dB.
Thanks this makes it very very clear to me now... ;-)
 
And even more confusing re. the OP the answer isn't 27 + 12 = 39.

Think about the poor TA who had office hours after handing back the graded exams... bunch of irritate students (all of which skipped lecture) demanding points back on their test because 27+12=39

Then we went onto having to work in pressure for SPL instead of dB, and figuring resonant frequency of open vs closed tubes et al., and about half the students in the course ended up dropping it before the next exam.

Fun stuff really
 
This is how I figure out how dB output will change:
20*Log10(Number of things that are the same volume)

Examples:
20*Log10(2)=6
20*Log10(3)=9.5
20*Log10(4)=12
20*Log10(5)=14
20*Log10(6)=15.6

I use that equation for gain staging when recording. If I'm recording 3 guitar tracks and they're all panned the same, I set their channel levels all to -9.5. If I'm recording a project with 24 tracks, I set all their initial levels to -27.6
 
Back
Top Bottom