NEW Need HELP

SpectraLayers
1

I edit my GoPro Videos just for fun as well as some 4K with a very advanced video setup. This is just a hobby. I am good at editing in DaVinci Resolve Studio but have no real clue on Spectra Layers work.

I have seen some videos but NONE of them start from the very beginning. For instance “here is a clip of sound we will work with to remove a voice” and all the steps needed.

Bring in the clip, tell you what to look for in the example and how to remove it? Everything I have seen is advanced and does not cater to someone just starting?? They start from someone who has done this over and over. There is a million videos of editing video start to finish but not for sound.

I am clueless and the manual is useless as most are. Are there any video courses out there for new people to help with this or step by step articles? I’m sure ounce I do a few I will get the hang of it as I am good with most processing.

If there is not I would be willing to Zoom with one of you guys and pay for your time sending payment through Zelle or something??

Any suggestions would help.

Bob

Rochester, NY

1 Like
2

Have you seen this site?

3

Hi Sound,

you are using video editing terminology: “clip” for example

so I’m guessing you are more aligned video editing tasks

can you let us know what it is you want to do?

Clean up (Noise Reduction or NR) audio?

remove wind and handling noises from camera audio?

separate noise from voice?

surely you have seen this:

4

or maybe you want to work on music; separating instruments from a mixed down song?

I have written a lot of posts to this forum since summer 2024 about my teething using SLP.

let’s start at the very top: are you running SLP in standalone or ARA in another program? Personally, I only use SL in standalone.

5

second step: (assuming you have SLP running and audio devices working, etc)

once you SLP running; you will want to have something to work on

navigate to “File”> “Open”

you can open a multitude of file types

I’d say you need to have an audio file ready to open in SL; wherever that is

you might also need to set up some folders for what you are working on, obviously that’s up to you

so: navigate your system to open an audio file; I usually use .wav

once you have an audio file open in SL, you ready to work

I’d say the easiest thing to work on is dialog of someone talking with unwanted background noises like wind, buzzes and mic handling noise

I would skip running any modules to start using SLP; I would start by learning to select areas in the Spectral Display Area.

The very first thing to understand when using TOOLS in SL is setting the SELECTION TYPE. This is very easy to overlook as Robin’s manual doesn’t point that out too well, IMO. What does that mean? sadly, it’s time to RTM:

“(1)Replace Selection: Replace the current selection with a new selection.
(2)Add to Selection: Add the new selection to the current selection (or press Shift ).
(3)Subtract from Selection: Subtract the new selection from the current selection (or press
Alt ).
(4)Intersect with Selection: Intersect the new selection with the current selection (or press
Shift + Alt ). This will keep the common areas between the two selections.

These four settings are the four “circle icons” just under The Edit menu when you are using a tool that needs the selection type. If you don’t see the circles, it’s because the currently active tool doesn’t require the selection type.

Hopefully you have a zoom level that is easy to see your audio. If you are having sluggish performance, set the FFT size in the display panel to around 2048 or less…higher FFT settings will have a delay when navigating, so first learn to navigate with lower FFT settings :slight_smile: (FFT settings are in the Display Panel; upper right)

So, before getting into the tools one by one, activate the the RECTANGULAR Selection tool (press “M” on keyboard). With the rectangular selection tool active, the the selection type circles’ buttons will be visible.

(1)Replace Selection: is default in SLP. This allows one selection at a time. *In SLP12, you need to be in Replace to resize the selection…each time user selects, the previous selection is cleared…try it out

(2)Add to Selection: As the name says, make all the selections user wants and the selections just add up without removing the last selection made as happens with Replace selection. This is how we make complex/ composite selections.

(3)Subtract from Selection: as the name says; this is how users remove parts of selection areas

(4)Intersect with Selection: I never use this, so I can’t say :slight_smile:

ok, get good at making selections to start :slight_smile:

6

Hey Bob,

Sent you a DM.

1 Like
7

OK, so what to do with the selections?

well, that’s where layers come in

manage layers in the layers panel on the right side of the UI

So, if carrying on from my last post, you opened an audio file and tried selections; well what got saved? You save your editing job as a .slp file; so you need to do that before committing any changes. File menu> Save Project or Save Project as… to make sure you are working with .slp and not directly on the audio file.

Once you saved your .slp as whatever name (syntax), you can start editing.

There are all kinds of ways to use layers. I’ve posted about it a lot on this board that I prefer to work non-destructively. This is because SLP only has a limited number of undos (64 is the max) and once user closes a .slp file, the undos are lost. If users just starts cutting away with the eraser tool, for example, that removed eraser tool information is not recoverable.

OK, so one way to remind yourself of what you have done is to cut unwanted SELECTIONs to a muted trash layer; and then you don’t have to worry about the levels of undo from slp session to session.

What I like to do is create a safety when I start. So, duplicate the original audio you opened by right click>duplicate the layer in the layers panel; label that duplicated layer “Safety” or however you want and the mute that duplicate. That way, you can refer back to the SAFETY in the event you mess up too much.

Then create a new layer, (also in the layers panel) and label the new layer if you want. There are several ways to create a new layer; I use the button in the bottom of the layers panel.

So now you have 3 layers; your source audio, a safety and an empty layer. Color them however you want. I have a system that works for me, it might not work for you.

8

Working with layers

so, you know how to select;

and you have some layers to work with.

When working with layers, it is critical that user knows which layer (or layers) is currently active.

In the layers panel, the currently selected layer has a light highlight on the layer’s container. (Starting with version 12 users can now select more than one layer at a time, so that’s fairly new).

Essentially, select audio for one layer and then cut or copy selections from one layer to another. Either use the Edit menu commands or keyboard shortcuts. Realize that each layer has a horizontal volume slider. Practice doing that with rectangular selection and you are on your way. Then things should begin to become clearer as you practice.

One key command that is critical to my work that is not obvious in the user manual is Re-Select which is ctrl+shift+D. Often times user will need “cut to layer” action several times to achieve what user wants from the previous selection which disappears after a cut or copy action. Just press Re-Select to apply the same selection.

If you have stability issues with your computer performance, that’s another bridge to cross, beyond the scope of using SL :slight_smile:

9

for me, the biggest issue starting out was my inability to comprehend what I was looking at in the Spectral Display. Over time teething in SL and doing spectrogram study of reading transforms by just searching the net and youtube really helped me begin to see and locate and select specific sounds (I have posted links to this forum previously). Nowadays (after thousands of hours in SL) I can find pretty much anything.

I bring this up to explain a little about FFT size. For most of my work I like FFT size set to 2048 smp because navigation is quick as screen redraw does not lag on my machine. That said, when focusing on editing low frequencies, higher FFT settings brings low frequency transforms into sharp focus; the tradeoff is sluggish, delayed performance and actual “cut to layer” actions can make you feel like the software has crashed or has hung…you will get “not responding” OS messages as SL churns the numbers, so when using higher FFT sizes, endeavour to be very patient.

All that should get you started. Just start practicing and the knowledge will accrue and you will be able to fish after that :slight_smile:

1 Like
10

Well thank you for everybody for getting back to me it’s really appreciated. I will read through these again and digest everything thank you. Some of you wanted to know what I was going to be doing so I will tell you exactly.

  1. I also fly model airplanes and sometimes on filming somebody and somebody walks up and just starts talking as someone is flying should and before I get too far to tell him to keep it down I have to deal with their voice in the clip. So first off removing somebody’s voice just talking is what I want to do.
  2. Second I’m on the beach in Florida a couple times a year and film a lot of stuff but there’s a lot of wind noise. Now a lot of it is taken down with microphone add-ons , and I know I can’t remove it 100% but I would like to just tone it down a little bit so I’m not sure how you tone down that to a certain percent. I think these are the two main things I would be doing so again I will read through everybody’s information and digest that and hopefully that will help any other suggestions are always welcome.

Thank you so much!!!

1 Like
11

I forgot to add my new system has plenty of run power.

AMD Threadripper CPU 32 threads 64 cores

NVidia 5090 Video Card

128 GB RAM

12GB of storage

12

Unmix Noisy Speech for this; you will still have artefacts if you remove the speech layer completely…yet, “distant” voices with no inhales, chuckles, sighs, or close proximity mouth noises is what SL excels at. So, where you have artefacts; clone stamp from clean areas to cover the holes and maybe some Heal. Nevertheless, when editing in the spectrogram, you still need to know what to look for…takes plenty of teething to read the spectrogram.

Wind noise is easily removed in SLP. Really, the wind noise that is typically unwanted is the low end (200Hz and below) mic/ speaker breakers. Usually I want to retain the high end for wind noise when the vision has moving vegetation and fabric, etc. Use higher FFT settings to more highly target low end high sound pressure issues where required. I mean, you can get away with just cutting everything 200Hz and below. This is where your playback format needs to be addressed. Is the final product hitting a sub-woofer? If so, you gotta deal with that however you usually do.