Taperssection.com
Gear / Technical Help => Post-Processing, Computer / Streaming / Internet Devices & Related Activity => Topic started by: ThePiedPiper on July 16, 2018, 07:42:34 PM
-
I have a show that I was playing it safe with the levels during the recording, and it's very low, so I'll be raising it approx 14db in Audition. I want to do some (minor) EQ. My question is what is the "best practice" here? EQ the show BEFORE raising the levels or EQ the show AFTER raising the levels? Also I imagine most dither and then re-sample (in that order) is that correct?
-
I have a show that I was playing it safe with the levels during the recording, and it's very low, so I'll be raising it approx 14db in Audition. I want to do some (minor) EQ. My question is what is the "best practice" here? EQ the show BEFORE raising the levels or EQ the show AFTER raising the levels? Also I imagine most dither and then re-sample (in that order) is that correct?
I always get the levels up before EQ. In a break from standard practice I add dynamics after EQ. I find that the gentle compression that I add to my recordings works better if I EQ first.
Can't comment on resample or dither - I record in 24bit/44.1kHz so I leave it as is.
-
If resampling to a higher rate, there may be some advantage to doing so prior to EQ.
If resampling to a lower rate, best practice would be to do so after EQ.
In general its best to make gross level adjustments prior to EQ, but if you are doing the level adjustment and EQing in the same editing software and saving a new file after all your editing manipulations, it doesn't really matter if you EQ first or adjust levels first, as the signal is represented within the much larger 32-bit floating-point calculation space within the software editor until is is saved as output.
Bit reduction should should always be your last processing step, other than tracking if you do that outside of your editing software. Dither when reducing bit depth, which is usually done as a single processing step (dither being applied just prior to truncating the excess bits).
-
Thank you Goodcooker and Gutbucket for the knowledge. It's the eternal hunt for "the best methods". I appreciate any and all knowledge.
-
If you’re just raising levels it doesn’t really matter. With any sort of dynamics, you should always EQ first, as you don’t want frequencies that you are cutting to have an effect on the compression
-
you don’t want frequencies that you are cutting to have an effect on the compression
True statement.
Also, as gutbucket said with more specificity above, the correct order (when resampling 'down', i.e. 48 kHz to 44.1 kHz) is to resample then dither/bit-depth reduction. Just be sure to add your fades before dithering, since the value added by dithering is most apparent in the quietest moments of your recording (often times the fade out).
-
If you’re just raising levels it doesn’t really matter. With any sort of dynamics, you should always EQ first, as you don’t want frequencies that you are cutting to have an effect on the compression
I agree with this and with gutbuckets general statements. I will add to the practical side something I learned the hard way:
If my original source is more than -10dB then I level it up (Amplify in Audacity) to -2dB; then process including EQ, THEN recheck the levels as occasionally the EQ'ing will make it go over 0.0dB. THEN compress if doing that step.
-
I'm taking notes! I hope people keep adding to this thread and it becomes a "guide" to some nice EQ & Mastering practices.
:headphones:
-
Since the levels can change up or down with EQ, I have always EQd first and then raised levels (normalize in my work flow). Not sure if this is the "correct" way to go but that has been my rational.
-
Normalization.
Is there an advantage, or disadvantage, using normalization vs. amplify?
-
I tend to work iteratively. That is not to say "like an idiot", but rather by homing in on where I want to be via multiple steps instead of making a single correction for each thing. The first corrections are the gross ones addressing the largest and most obvious changes needed (and serves to make it more enjoyable to listen to while making other corrections), the last corrections are minor fine-tuning corrections which take into account the effect of other corrections I've made.
I never really thought about it this way, but I'm more or less going through a general procedural order, then going back and making smaller polishing corrections by sort of going in reverse order through the previous steps. Something like this-
>Gross level corrections- adjusting to something more reasonable if they are very low (no need to get overly precise at this step)
>Sync/time-stretch sources if required
>Stereo balance corrections
>Gross error corrections (dropouts, ittermitancies, whatev)
>Gross EQ corrections (fix bloated bass, dull treble due to windscreens, presence-range correction, and whatever else obviously needs correction).
>Dynamics corrections- knock down errant peaks (manually or with a limiter), pull up the level of quiet parts if called for (volume envelopes and/or parallel comp), reduce the level of loud audience reaction between songs if called for (volume envelopes usually for me, some use a limiter for this), overall compression if called for.
(^Dynamics changes can change EQ balance, so there can be less differentiation between these two steps than it may seem. Especially with multiband compression)
And sort of back though in sort of reverse order..
>Small error corrections (spectral editing of clicks, whistles, candy wrappers, coughs, whatev)
>EQ polishing
>Balance check
>Fades
>Final level adjustment (master bus fader gain / limiting)
>Downsampling if required
>Bit reduction (with dither) if required
>Save and output
Not all projects need all this stuff, and not everyone will want to do all this stuff. Just skip the steps which aren't required. Tracking/tagging however you do it, either within the editor or afterwards in CDWAV or whatev (this will not alter the audio data other than chopping into separate files, so not necessary to do this prior to downsampling and bit reduction steps, making this step sort of an exception that way).
-
Normalization.
Is there an advantage, or disadvantage, using normalization vs. amplify?
Simple peak normalization is amplifying, only with the normalization routine monitoring the levels for you so you don't have to go back and checking that peak levels are where you want them after the gain correction. You specify what you want the highest peak to be and the normalization routine figures out how much to amplify for you.
"RMS normalization", normalization routines which includes brick-wall limiting, and other variants on normalization introduce other complexities. If you stick with simple peak normalization, no worries. Best to check what value the highest peak hits prior to exporting the files regardless of how you go about adjusting final levels.
-
Good advice here. I will add that I like to add gain and general L-R balance first if needed. But, I'll normalize it 2-4 db below zero, because EQ can add gain and push a mix into clipping if you normalize too much first. Always check the loudest parts after adding EQ to make sure you're not clipping.