Wednesday, 3 December 2014
Three-Point Editing
In the Assembly phase of the edit process, you saw how to select a part of a video clip and add it to the sequence at the position in time where the playhead is on the timeline. The basic editing process involves marking 3 points in time to set the start and length of the clip to edit into the sequence and the position it will be added; this is why it is called Three-Point Editing.
However there are lots of different ways you can mark the three points; for example the In and Out points can be on the timeline itself and Premiere will use the Playhead in the Source Monitor as the third point. This is really useful if you know exactly where you need to place a clip in your edit and you want to replace a specific part of what is already in the sequence.
One reason you might do this is if you want to keep a line of dialogue in the edit but you want to show it from a different camera angle. If you've filmed your scene from multiple camera setups and have an alternative take of the line, you can "patch" it over the top of the edit to show a different point of view.
In the example edit below, there are three clips on the timeline. An establishing shot showing the male character on the phone, a cut to a woman sitting in a chair who stands up and leaves frame right; and a final shot showing the woman entering from the left to finish off the scene.
In the second shot, both actors have lines and take turns in the dialogue. On set, the sound recordist has wired the woman's microphone into one channel of the sound recording (the Left channel) and the man's microphone to the other channel (the Right). This is a very common way of recording the sound of two actors in the scene. Dialogue doesn't have to be recorded in stereo, a mono track will do. So its possible to use a lapel mic for each actor for the clearest sound and record the two actors at the same time straight into the camera. At the same time there will probably be a Boom mic operator recording one or other of the actors into a separate sound recorder.
You can actually see the different lines of dialogue in the sound track. Each time the woman speaks, there is a peak in the Left channel on the timeline and each time the man speaks there's a peak on the Right channel.
So the dialogue is switching between the two characters, but the picture in this clip is always of the woman. What if we want to cut to the man delivering one of his lines?
Because we have a visual clue from the sound waveform, we know exactly where the line of dialogue is; all we have to do is select that part of the timeline and replace it with a shot of the man saying the same line.
To select on the timeline, we can use IN and OUT points exactly as we would in the Source Monitor. To mark In and Out on the timeline, click on the timeline window to make it active then just press the I Key and O Key where you need the points to be.
In this case that means just before the man's last line of dialogue and just after, using the sound wave as a guide. The selected portion of the timeline will be highlighted.
We now have two of the points we need for an edit. For the final, third point, Premiere will use the Playhead in the Source Monitor. Now (as long as the Track Targets are set to V1 and A1) when you do an Overwrite edit the new clip will replace just the part of the timeline that is highlighted.
Using the IN and OUT points this way round makes it really easy to replace a line of dialogue with a different shot taken from another angle.
- In Point: marks the start of the clip you will be adding
- Out Point: marks the end or length of the clip
- Playhead: sets the point in the sequence where the clip will be added
However there are lots of different ways you can mark the three points; for example the In and Out points can be on the timeline itself and Premiere will use the Playhead in the Source Monitor as the third point. This is really useful if you know exactly where you need to place a clip in your edit and you want to replace a specific part of what is already in the sequence.
One reason you might do this is if you want to keep a line of dialogue in the edit but you want to show it from a different camera angle. If you've filmed your scene from multiple camera setups and have an alternative take of the line, you can "patch" it over the top of the edit to show a different point of view.
In the example edit below, there are three clips on the timeline. An establishing shot showing the male character on the phone, a cut to a woman sitting in a chair who stands up and leaves frame right; and a final shot showing the woman entering from the left to finish off the scene.
In the second shot, both actors have lines and take turns in the dialogue. On set, the sound recordist has wired the woman's microphone into one channel of the sound recording (the Left channel) and the man's microphone to the other channel (the Right). This is a very common way of recording the sound of two actors in the scene. Dialogue doesn't have to be recorded in stereo, a mono track will do. So its possible to use a lapel mic for each actor for the clearest sound and record the two actors at the same time straight into the camera. At the same time there will probably be a Boom mic operator recording one or other of the actors into a separate sound recorder.
You can actually see the different lines of dialogue in the sound track. Each time the woman speaks, there is a peak in the Left channel on the timeline and each time the man speaks there's a peak on the Right channel.
So the dialogue is switching between the two characters, but the picture in this clip is always of the woman. What if we want to cut to the man delivering one of his lines?
Because we have a visual clue from the sound waveform, we know exactly where the line of dialogue is; all we have to do is select that part of the timeline and replace it with a shot of the man saying the same line.
To select on the timeline, we can use IN and OUT points exactly as we would in the Source Monitor. To mark In and Out on the timeline, click on the timeline window to make it active then just press the I Key and O Key where you need the points to be.
In this case that means just before the man's last line of dialogue and just after, using the sound wave as a guide. The selected portion of the timeline will be highlighted.
We now have two of the points we need for an edit. For the final, third point, Premiere will use the Playhead in the Source Monitor. Now (as long as the Track Targets are set to V1 and A1) when you do an Overwrite edit the new clip will replace just the part of the timeline that is highlighted.
Using the IN and OUT points this way round makes it really easy to replace a line of dialogue with a different shot taken from another angle.
Wednesday, 26 November 2014
Multicam Editing in Premiere
Live events such as sports and concerts are often filmed with many different cameras from many different angles at the same time. This gives the editor material to cut together but also ensures that as much of the action as possible is recorded. Even in feature film projects there are situations where the action can't be repeated, especially for action sequences such as car crashes, explosions. Shots like that are a good candidate for Multi-camera filming.
Because multi-camera is used so often, Premiere has specific tools to edit this kind of project. The first step is to make sure all the clips are in time with each other by synchronising the videos. Professional film shoots will use high-end cameras which can share timecode signals and will start and stop recording at exactly the same time so all the clips will automatically be in sync.
For low budget shoots, on the other hand, the best if not very accurate way to get several videos in time with each other is to have some kind of visual signal recorded by every camera. The ideal solution is a film clapper board but, failing that, just a crew member clapping their hands will do.
The important thing is that all the cameras can see the sync signal -- if that means pointing the camera at the signal then panning it to catch the action then that needs to be planned in to the shot.
To start editing, import all the clips into Premiere as usual. It is a good idea to create a new Bin for each shot to keep the clips from all the cameras together. Once the clips are organised they need to be loaded into the Source Monitor one by one to line them up on the visual sync.
Use the playback controls J K L or the LEFT and RIGHT Arrow keys to get the video as close to the hand clap as possible. Once you are on the right frame, put a marker on it by pressing M then do the same thing for the rest of the videos. Make sure that the markers are all as close to the same frame as you can get them.
With all the clips marked, select them all from the Bin by holding the SHIFT key and clicking each one till they are all highlighted. The order you select them in is important! The last clip you pick will be the master clip. While you can switch from one video to another in the edit, the sound will come just from the master clip unless you tell Premiere to cut the sound as well. Usually you would pick the clip with the best, clearest audio to be the master.
With all the clips marked and selected, RIGHT Click on any of the clips and select Create Multi-Camera Source Sequence from the menu.
A pop-up window will appear with the various options for creating the new sequence. Most of these only apply to Professional cameras which are able to synchronise themselves. To sync to the markers, just make sure that option is selected then, from the menu next to it pick the name of the marker you want to use. If you haven't named your marker, this will be something like Unnamed Marker 1
If this has worked, you should now have some extra items in your project.
Click on the new Multicam sequence to open it in the Source Monitor. You will see all your videos arranged in a grid. Some of the pictures may be blank since the videos all start at slightly different times. Immediately do an Overwrite edit to put the whole clip onto the timeline then move the playhead back to the start of the clip.
Because multi-camera is used so often, Premiere has specific tools to edit this kind of project. The first step is to make sure all the clips are in time with each other by synchronising the videos. Professional film shoots will use high-end cameras which can share timecode signals and will start and stop recording at exactly the same time so all the clips will automatically be in sync.
For low budget shoots, on the other hand, the best if not very accurate way to get several videos in time with each other is to have some kind of visual signal recorded by every camera. The ideal solution is a film clapper board but, failing that, just a crew member clapping their hands will do.
The important thing is that all the cameras can see the sync signal -- if that means pointing the camera at the signal then panning it to catch the action then that needs to be planned in to the shot.
To start editing, import all the clips into Premiere as usual. It is a good idea to create a new Bin for each shot to keep the clips from all the cameras together. Once the clips are organised they need to be loaded into the Source Monitor one by one to line them up on the visual sync.
 |
| First Clip of a Multicam sequence cued up to the Visual Sync frame. |
Use the playback controls J K L or the LEFT and RIGHT Arrow keys to get the video as close to the hand clap as possible. Once you are on the right frame, put a marker on it by pressing M then do the same thing for the rest of the videos. Make sure that the markers are all as close to the same frame as you can get them.
With all the clips marked, select them all from the Bin by holding the SHIFT key and clicking each one till they are all highlighted. The order you select them in is important! The last clip you pick will be the master clip. While you can switch from one video to another in the edit, the sound will come just from the master clip unless you tell Premiere to cut the sound as well. Usually you would pick the clip with the best, clearest audio to be the master.
With all the clips marked and selected, RIGHT Click on any of the clips and select Create Multi-Camera Source Sequence from the menu.
A pop-up window will appear with the various options for creating the new sequence. Most of these only apply to Professional cameras which are able to synchronise themselves. To sync to the markers, just make sure that option is selected then, from the menu next to it pick the name of the marker you want to use. If you haven't named your marker, this will be something like Unnamed Marker 1
If this has worked, you should now have some extra items in your project.
Click on the new Multicam sequence to open it in the Source Monitor. You will see all your videos arranged in a grid. Some of the pictures may be blank since the videos all start at slightly different times. Immediately do an Overwrite edit to put the whole clip onto the timeline then move the playhead back to the start of the clip.
Before we can actually start editing, we need to make some changes to the Programme Monitor to see and record the multi cam sequence properly. First, set the Monitor into Multi-Camera mode by clicking on the small wrench icon and selecting Multi-Camera from the menu that appears.
Next, to make things a bit easier, we will add the Record button to the button bar at the bottom of the Monitor window. To customise the buttons, click on the small + to the lower right of the window to bring up a palette of the available buttons. Then drag and drop the Record button to next to the existing playback controls.
With everything set up, we are ready to edit. Unlike single camera editing, with multicam, we will be editing while the video is playing back. In exactly the same way that a TV studio director will switch from one camera to another during a live broadcast, Premiere lets you choose your camera angle "on the fly".
To capture the camera switches, we need to be recording the playback. Press the Record button, then press the Play button to start the sequence. As the playback happens, you will see all the angles on the left of the Programme Monitor and the selected angle on the right. The preview on the left will have the selected angle marked with a red outline as you are recording.
You can click on another clip to change to a different view while you are recording. However a quicker way to switch cameras is by using the number keys along the top of the keyboard. The keys 1 - 9 correspond to one of the cameras in the multicam sequence. Pressing the right number will switch straight to that camera.
You won't see any cuts in the timeline until you get to the end of the sequence and the playback stops. When it does you should see lots of clips which reflect the camera selections you made.
 |
| Programme Monitor during recording. The clip being recorded is outlined in red. |
 |
| At the end of the recording, the cuts show up in the timeline. |
If you look at the audio track A1, you'll notice that the sound is not cut -- the whole track is taken from the master clip. This is usually what you want.
With the edit in place, you can go back and use the Trimming tools to fix the timing of the cuts. Just be aware that you only want to move the edit point so the only kind of trim you need will be Roll trim which will keep the timing and synchronisation correct in the rest of the sequence.
Also if you change your mind about which angle you want, you can Right click any part of the sequence and select a different angle by choosing Multi-Camera -> Camera # where # is the number of the camera you want to use instead.
 |
| Right click and choose a camera from the menu to pick a different angle. |
Monday, 24 November 2014
Audio Sweetening in Premiere Part 1
In the early stages of editing, especially during the Assembly and Trimming phases, its common to concentrate on the visuals making sure the pacing and continuity of the scene is right. Once the structure is in place, though its important to make sure that the sound is working at least as well as the picture.
You might not do a full audio mix at this stage but you will want to make sure the sound edits together smoothly and that the various tracks, Dialogue, Sound FX, Foley and Music are balanced and mixed properly. This stage of the edit is often referred to as Audio Sweetening.
Premiere Pro has all the audio tools needed to do a decent sound mix and you can work on individual clips, whole audio tracks as well as Sub-mixes.
First job is to set up the interface for audio editing. You can do this by going to the Window -> Workspaces -> Audio menu item.
The Audio layout is very similar to the default layout but it adds an Audio Mixer panel which you will need for balancing tracks and Sub-mixes later on. For now we will concentrate on clips in the timeline.
By default, each audio track has a volume overlay displayed on each clip (the white line in the picture below)
The line shows the volume adjustments for each clip. You can grab the line and move it up or down to change the volume of each clip separately. If you hover the mouse over the line, Premiere will show you what the current volume adjustment is:
Sound volume is measured in Decibels, which is written as dB. A sound playing at its recorded volume, in other words without any volume adjustments, is said to be at 0 dB and as the volume goes below zero, the sound gets quieter.
However Decibels are not a linear scale, in fact every 6 dB drop in volume makes the sound half as loud so -6 dB is half the normal volume -12 dB is half as loud as that ( or 1/4 of the normal volume ) and -18 dB is only 1/8 as loud as the sound was when it was recorded; and a sound at minus Infinity dB is perfectly silent.
Of course we are ignoring the volume setting on the speakers which is a separate issue.
As you drag the volume adjustment line down or up, Premiere will show you how much the volume has changed:
Adjusting each clip one-by-one down the timeline, you can balance the sound through-out the sequence so there aren't any distracting jumps in volume.
This is fine if each clip is recorded evenly without any dips or jumps in volume; but what if the volume of a clip changes, for example if someone turns away from the mic slightly? It is possible for you to let the volume adjustment change over time buy adding keyframes to the clip.
To add a keyframe, all you have to do is use the Pen tool ( P on the keyboard ). With the pen selected, you can click on the volume adjustment line to add a keyframe whereever there needs to be a change in volume. With at least two key frames, you can set the volume at the start and finish of the clip to be different and Premiere will fade between the two levels. Add the keyframes with the Pen Tool (P) then press V for the Selection Tool and drag down the part of the clip that needs to be quieter.
The default keyframes that Premiere adds cause sudden changes in volume which don't always sound as smooth as they should. To fix this, you can Right Click on a keyframe and change the Interpolation type to Bezier which will make the keyframe change in a more controllable way:
The Bezier option adds a handle to the keyframe which you can move around to change the bend of the line so the fade-in or fade-out is smoother.
That's it for adjusting clip volume, in Part 2, we'll have a look at adjusting whole tracks with the Audio Mixer.
You might not do a full audio mix at this stage but you will want to make sure the sound edits together smoothly and that the various tracks, Dialogue, Sound FX, Foley and Music are balanced and mixed properly. This stage of the edit is often referred to as Audio Sweetening.
Premiere Pro has all the audio tools needed to do a decent sound mix and you can work on individual clips, whole audio tracks as well as Sub-mixes.
First job is to set up the interface for audio editing. You can do this by going to the Window -> Workspaces -> Audio menu item.
And then, just to be on the safe side you can reset the layout to organise the windows with Window -> Workspace -> Reset Current Workspace
The Audio layout is very similar to the default layout but it adds an Audio Mixer panel which you will need for balancing tracks and Sub-mixes later on. For now we will concentrate on clips in the timeline.
By default, each audio track has a volume overlay displayed on each clip (the white line in the picture below)
The line shows the volume adjustments for each clip. You can grab the line and move it up or down to change the volume of each clip separately. If you hover the mouse over the line, Premiere will show you what the current volume adjustment is:
Sound volume is measured in Decibels, which is written as dB. A sound playing at its recorded volume, in other words without any volume adjustments, is said to be at 0 dB and as the volume goes below zero, the sound gets quieter.
However Decibels are not a linear scale, in fact every 6 dB drop in volume makes the sound half as loud so -6 dB is half the normal volume -12 dB is half as loud as that ( or 1/4 of the normal volume ) and -18 dB is only 1/8 as loud as the sound was when it was recorded; and a sound at minus Infinity dB is perfectly silent.
Of course we are ignoring the volume setting on the speakers which is a separate issue.
As you drag the volume adjustment line down or up, Premiere will show you how much the volume has changed:
Adjusting each clip one-by-one down the timeline, you can balance the sound through-out the sequence so there aren't any distracting jumps in volume.
This is fine if each clip is recorded evenly without any dips or jumps in volume; but what if the volume of a clip changes, for example if someone turns away from the mic slightly? It is possible for you to let the volume adjustment change over time buy adding keyframes to the clip.
To add a keyframe, all you have to do is use the Pen tool ( P on the keyboard ). With the pen selected, you can click on the volume adjustment line to add a keyframe whereever there needs to be a change in volume. With at least two key frames, you can set the volume at the start and finish of the clip to be different and Premiere will fade between the two levels. Add the keyframes with the Pen Tool (P) then press V for the Selection Tool and drag down the part of the clip that needs to be quieter.
The default keyframes that Premiere adds cause sudden changes in volume which don't always sound as smooth as they should. To fix this, you can Right Click on a keyframe and change the Interpolation type to Bezier which will make the keyframe change in a more controllable way:
The Bezier option adds a handle to the keyframe which you can move around to change the bend of the line so the fade-in or fade-out is smoother.
That's it for adjusting clip volume, in Part 2, we'll have a look at adjusting whole tracks with the Audio Mixer.
Audio Sweetening in Premiere Pro Part 2
In Part 1, you saw how you can adjust the volume of a clip in the timeline either by a constant amount or by an adjustment that varies while the clip plays.
Once the clips on each track are the same volume as each other, the mix between tracks needs to be balanced. Typically you would organise your sound track on the time line by the type of sound. So, in a simple mix for example, all the Dialogue would go on track A1, the Sound Effects would go on A2 and the background Music would go on track A3.
To keep the different sounds separate and easy to hear, you would usually set the Dialogue to peak at -6 dB, the SFX to be half as loud at -12 dB and the Music to be the quietest at around -18 dB. This way the Dialogue is not going to be drowned out by the Effects or Music but all the ambient sounds are still going to be audible under the actors lines.
Also, having the loudest track at -6 dB means there is still room to add really loud spot effects like an explosion or a gun shot which should be even louder than the dialogue.
Mixing the tracks is done in the Audio Track Mixer window:
The Mixer consists of a set of Volume sliders, one for each audio track as well as a Master control which sets the volume of the final output to the speakers. Its a good idea to change the labels on each slider to match what you have on each track. This also changes the names of the tracks themselves.
As the sound track is playing, the Level Meter gives you feedback about the changing volume of the sound. The loudest part of the track is marked by a Yellow peaking line:
Using this you can work out which track is the loudest and balance it out with the others.
Having control over the overall track volume is really useful but what if you have more than one Dialogue track or you have multiple Sound Effects for different characters or shots? For example it is really common to have a separate voice track for each actor especially if they are mic'd up with a lapel mic each and you're using a boom as well.
On top of that, if you loose a line during filming, you may have to re-recorded it afterwards in the studio and dub it onto the edit in post. You could easily end up with two voice tracks per actor and a scene could have any number of characters.
As another example, imagine you're editing a restaurant scene. You will need background SFX such as multiple conversations, plate and cutlery sounds, footsteps, chairs moving and countless other sounds which will probably all be recorded separately, so you quickly end up with lots of SFX tracks in your edit. Add all the Dialogue tracks and you have a big problem: how can you balance the Dialogue with the rest of the sound mix when there are so many different volume levels to set?
This is where Sub-mixes come in...
A Sub-mix is just a special kind of sound track which aggregates or combines two or more tracks into one with its own sound control. To create a Sub-mix all you have to do is Right Click the header of the timeline and select Add Audio Submix Track
Once the clips on each track are the same volume as each other, the mix between tracks needs to be balanced. Typically you would organise your sound track on the time line by the type of sound. So, in a simple mix for example, all the Dialogue would go on track A1, the Sound Effects would go on A2 and the background Music would go on track A3.
To keep the different sounds separate and easy to hear, you would usually set the Dialogue to peak at -6 dB, the SFX to be half as loud at -12 dB and the Music to be the quietest at around -18 dB. This way the Dialogue is not going to be drowned out by the Effects or Music but all the ambient sounds are still going to be audible under the actors lines.
Also, having the loudest track at -6 dB means there is still room to add really loud spot effects like an explosion or a gun shot which should be even louder than the dialogue.
Mixing the tracks is done in the Audio Track Mixer window:
The Mixer consists of a set of Volume sliders, one for each audio track as well as a Master control which sets the volume of the final output to the speakers. Its a good idea to change the labels on each slider to match what you have on each track. This also changes the names of the tracks themselves.
As the sound track is playing, the Level Meter gives you feedback about the changing volume of the sound. The loudest part of the track is marked by a Yellow peaking line:
Using this you can work out which track is the loudest and balance it out with the others.
Having control over the overall track volume is really useful but what if you have more than one Dialogue track or you have multiple Sound Effects for different characters or shots? For example it is really common to have a separate voice track for each actor especially if they are mic'd up with a lapel mic each and you're using a boom as well.
On top of that, if you loose a line during filming, you may have to re-recorded it afterwards in the studio and dub it onto the edit in post. You could easily end up with two voice tracks per actor and a scene could have any number of characters.
As another example, imagine you're editing a restaurant scene. You will need background SFX such as multiple conversations, plate and cutlery sounds, footsteps, chairs moving and countless other sounds which will probably all be recorded separately, so you quickly end up with lots of SFX tracks in your edit. Add all the Dialogue tracks and you have a big problem: how can you balance the Dialogue with the rest of the sound mix when there are so many different volume levels to set?
This is where Sub-mixes come in...
A Sub-mix is just a special kind of sound track which aggregates or combines two or more tracks into one with its own sound control. To create a Sub-mix all you have to do is Right Click the header of the timeline and select Add Audio Submix Track
This adds an extra, empty track on the timeline and an extra volume control in the Audio Mixer Window. You don't add clips to this new track. Instead you set the tracks in the Mixer to go to one of your Sub-mix tracks. All of the Sub-mixes then go to the Master and end up going to the speakers. Grouping similar tracks in the timeline into Sub-mixes in the Mixer gives you a new Volume control which lowers or raises the volume of all those tracks at once.
To set the individual tracks to go to a particular Sub-mix, pick the name of the Sub-mix from the menu at near the top of the Mixer window.
This way you can adjust the volume of all of the Dialogue or all of the Sound Effects in one go, no matter how many tracks of each you have, and still have control over the individual tracks to balance everything out. Just make sure you rename the Sub-mixes to keep things organised.
Monday, 27 February 2012
QuickTip: Fixing Noisy Sound Clips with Audition
Hi again,
One of the commonest problems on any film shoot is poorly recorded sound. Often sound levels are too low meaning a lot of 'gain' has to be added to make the sound louder. Alternatively, faulty equipment or a bad recording environment can ruin the sound. Either way you can end up with very noisy audio which can be very distracting.
Luckily there have been major improvements in sound processing tech which mean (some of) these problems can be fixed. For example, Adobe Audition has powerful noise reduction filters which can rescue badly recorded sound. Here's how you do it:
First, import your noisy sound clip into Audition:
You'll see a standard 'waveform' sound display. Its sometimes easier to see the noise damage by turning on the Spectral Frequency Display by pressing SHIFT D
The example sound file is a ticking clock. The 'Ticks' are the tall spikes in the frequency display. The noise shows up as grainy dots all over the background. Before Audition can remove the noise, you have to tell it what the noise sounds like by selecting a section of the clip where there is only noise and no 'foreground' sounds. Press T to turn on the Time Selection Tool then drag out a box on the sound between two sound peaks.
With just the noise selected, go to Effects > Noise Reduction / Restoration > Capture Noise Print or press SHIFT P and Audition will analyse the noise to identify specific frequencies.
Once you have a noise print, select a bigger portion of the clip, one including the recorded sound and go to Effects > Noise Reduction / Restoration > Noise Reduction (process) and a new window will open up.
Press the Play button in the bottom left of the new window to listen to the processed sound. The filter will only affect the section you have highlighted. By pressing Play a few times, you can listen to the whole sound to get a 'before' and 'after' of the noise reduction. If the sound is really noisy, you can adjust the sliders to remove more or less noise but remember that overdoing it will start to make the sound 'tinny' and synthetic. Once you are happy with the result press the Select Entire File button, then click Apply and the sound is fixed; just save it back out to your Final Cut Pro project folder and import it back into Final Cut and you're done!
P.S. Audition is only installed on the iMacs in DMW 4 (room G123b)
One of the commonest problems on any film shoot is poorly recorded sound. Often sound levels are too low meaning a lot of 'gain' has to be added to make the sound louder. Alternatively, faulty equipment or a bad recording environment can ruin the sound. Either way you can end up with very noisy audio which can be very distracting.
Luckily there have been major improvements in sound processing tech which mean (some of) these problems can be fixed. For example, Adobe Audition has powerful noise reduction filters which can rescue badly recorded sound. Here's how you do it:
First, import your noisy sound clip into Audition:
You'll see a standard 'waveform' sound display. Its sometimes easier to see the noise damage by turning on the Spectral Frequency Display by pressing SHIFT D
The example sound file is a ticking clock. The 'Ticks' are the tall spikes in the frequency display. The noise shows up as grainy dots all over the background. Before Audition can remove the noise, you have to tell it what the noise sounds like by selecting a section of the clip where there is only noise and no 'foreground' sounds. Press T to turn on the Time Selection Tool then drag out a box on the sound between two sound peaks.
With just the noise selected, go to Effects > Noise Reduction / Restoration > Capture Noise Print or press SHIFT P and Audition will analyse the noise to identify specific frequencies.
Once you have a noise print, select a bigger portion of the clip, one including the recorded sound and go to Effects > Noise Reduction / Restoration > Noise Reduction (process) and a new window will open up.
Press the Play button in the bottom left of the new window to listen to the processed sound. The filter will only affect the section you have highlighted. By pressing Play a few times, you can listen to the whole sound to get a 'before' and 'after' of the noise reduction. If the sound is really noisy, you can adjust the sliders to remove more or less noise but remember that overdoing it will start to make the sound 'tinny' and synthetic. Once you are happy with the result press the Select Entire File button, then click Apply and the sound is fixed; just save it back out to your Final Cut Pro project folder and import it back into Final Cut and you're done!
P.S. Audition is only installed on the iMacs in DMW 4 (room G123b)
QuickTip: Sweetening Sound Effects
There are a few rules you have to follow when editing audio.
One of the most important is that you should always cut sound clips on a 'Zero Crossing'. What's a zero crossing? -- well, digital sound is just a series of numbers, CD sound for example is stored as numbers ranging from -127 to +127. The number Zero is the same as silence. The further away from zero the number values get, the louder the sound.
In most programmes, audio waveforms are drawn with Zero along the middle of the track, with +127 at the top and -127 at the bottom (assuming you're using CD quality sound). If you zoom right into the waveform, you'll see that the sound wave goes up and down passing through Zero over and over again. In the screenshot from Final Cut Pro 7 below, Zero happens to be along the dark purple line and a single frame of video is highlighted in black:
Each point where the waveform passes through Zero is a 'Zero Crossing' and this is where you should ideally cut the audio.
But there are a couple of problems. Firstly, FCP 7 only lets you make cuts exactly on video frames. Chances are there won't be a zero crossing exactly at the beginning or end of the frame. The next problem is that wherever you have two sound clips together that don't both start and end at Zero, you'll get a nasty 'click' on the sound track where they join. So how to fix this?
The easiest and most effective way is simply to add a subtle 'Fade-In' and 'Fade-Out' effect to every single Sound FX clip. That way you can guarantee that every clip will start and end on Zero and there will be no chance of getting a glitch on the sound track.
You could use audio transitions from the Effects menu to add the Fade-In and Fade-Out but there is another way which is worth knowing about. Make sure you have clip overlays turned on and put a Fade-In and Fade-Out on one clip like this.
Next select the clip and go to Edit > Copy
Now select all of the other sound effects clips by pressing A for the arrow tool and drawing a box around them in the timeline. When they are all selected, go to Edit > Paste Attributes
The Paste Attributes window will open up. We only want to copy and paste the 'Levels' keyframes we put on the first clip. We also want to make sure that the keyframes are squashed or stretched to the length of the other clips otherwise they might drop off the end of the clip. Make sure "Scale Attribute Times" is ticked and then tick "Levels" as the only attribute we want to paste:
Click OK and you should end up with Fade-In and Fade-Out on every Sound FX clip with no more worries about bad clicking on the audio track.
One of the most important is that you should always cut sound clips on a 'Zero Crossing'. What's a zero crossing? -- well, digital sound is just a series of numbers, CD sound for example is stored as numbers ranging from -127 to +127. The number Zero is the same as silence. The further away from zero the number values get, the louder the sound.
In most programmes, audio waveforms are drawn with Zero along the middle of the track, with +127 at the top and -127 at the bottom (assuming you're using CD quality sound). If you zoom right into the waveform, you'll see that the sound wave goes up and down passing through Zero over and over again. In the screenshot from Final Cut Pro 7 below, Zero happens to be along the dark purple line and a single frame of video is highlighted in black:
Each point where the waveform passes through Zero is a 'Zero Crossing' and this is where you should ideally cut the audio.
But there are a couple of problems. Firstly, FCP 7 only lets you make cuts exactly on video frames. Chances are there won't be a zero crossing exactly at the beginning or end of the frame. The next problem is that wherever you have two sound clips together that don't both start and end at Zero, you'll get a nasty 'click' on the sound track where they join. So how to fix this?
The easiest and most effective way is simply to add a subtle 'Fade-In' and 'Fade-Out' effect to every single Sound FX clip. That way you can guarantee that every clip will start and end on Zero and there will be no chance of getting a glitch on the sound track.
You could use audio transitions from the Effects menu to add the Fade-In and Fade-Out but there is another way which is worth knowing about. Make sure you have clip overlays turned on and put a Fade-In and Fade-Out on one clip like this.
Next select the clip and go to Edit > Copy
Now select all of the other sound effects clips by pressing A for the arrow tool and drawing a box around them in the timeline. When they are all selected, go to Edit > Paste Attributes
The Paste Attributes window will open up. We only want to copy and paste the 'Levels' keyframes we put on the first clip. We also want to make sure that the keyframes are squashed or stretched to the length of the other clips otherwise they might drop off the end of the clip. Make sure "Scale Attribute Times" is ticked and then tick "Levels" as the only attribute we want to paste:
Click OK and you should end up with Fade-In and Fade-Out on every Sound FX clip with no more worries about bad clicking on the audio track.
QuickTip: Normalising Sound Clips
In this quick tip I'll be reviewing the workshop session from last week about 'Audio Sweetening'.
A very common problem in post-production is editing together sound tracks with different volumes or 'levels'. Sound recorded on different shoots or with different equipment may be louder or quieter than they should be. You could adjust the levels on each clip until they match -- but there is a better way.
The Normalisation Gain tool will analyse the sound volume and automatically adjust the levels so that the volume is constant through-out your edit. Here is how to use it:
In the timeline select all the clips you want to fix
You can see from the waveforms that the volume of the sound in each clip don't match. Next go to Modify > Audio > Apply Normalization Gain
A small window will open which allows you to set the target volume in decibels (dB). Your dialog track should usually be set to around -12dB. Move the slider to -12bB and click OK.
If you play back the edit the sound levels should now match with no really loud or really quite dialog. The waveforms in the timeline haven't changed though. That's because the Normalize Gain tool doesn't alter the sound directly, instead it adds a Gain Filter to the sound. This means the changes aren't permanent and each clip can be easily adjusted if you want. Double click a clip to open it in the Viewer and click the Filters tab to see the Gain Filter settings:
Remember, negative dB values make the sound quieter, and every 3dB doubles the effect.
A very common problem in post-production is editing together sound tracks with different volumes or 'levels'. Sound recorded on different shoots or with different equipment may be louder or quieter than they should be. You could adjust the levels on each clip until they match -- but there is a better way.
The Normalisation Gain tool will analyse the sound volume and automatically adjust the levels so that the volume is constant through-out your edit. Here is how to use it:
In the timeline select all the clips you want to fix
You can see from the waveforms that the volume of the sound in each clip don't match. Next go to Modify > Audio > Apply Normalization Gain
A small window will open which allows you to set the target volume in decibels (dB). Your dialog track should usually be set to around -12dB. Move the slider to -12bB and click OK.
If you play back the edit the sound levels should now match with no really loud or really quite dialog. The waveforms in the timeline haven't changed though. That's because the Normalize Gain tool doesn't alter the sound directly, instead it adds a Gain Filter to the sound. This means the changes aren't permanent and each clip can be easily adjusted if you want. Double click a clip to open it in the Viewer and click the Filters tab to see the Gain Filter settings:
Remember, negative dB values make the sound quieter, and every 3dB doubles the effect.
Thursday, 16 February 2012
Camera Tracked Titles
Hi Everyone,
Here as promised is an After Effects tutorial to recreate the 'floating title' effect from last week's workshop. Before we begin, have another look at the final result as a reminder:
Stage 1: Tracking the Camera Move
Open After Effects and make a new Composition matching your video settings
Import your video file and drag it onto the composition.
You'll need to add a Camera Tracker to the video by clicking on the 'Tracker' tab of the bottom right panel. If there is no 'Tracker' tab, go up to the Window menu and make sure 'Tracker' has a tick next to it.
There are two basic options: 'Track Motion' and 'Stabilize Motion'. Stabilize lets you take shaky, hand-held video and stabilize it to look like it was shot on a tripod. It works quite well but is not going to give you as good results as a proper tripod shot -- there are no short cuts in cinematography!
We're going to use the first option. Click the 'Track Motion' button and the rest of the panel will switch on. Because there is only one video track the 'Motion Source' option should automatically select your video. The only option we need to change is 'Track Type'. Set this to 'Perspective Corner Pin'
You should now have a tracker called 'Tracker 1' on your video and some extra buttons will have appeared under the picture.
The tracker is made up of four labeled 'Track Points' at the corners of a white square. The track points are used to mark an object in the video so that After Effects knows what colours and shapes to track from frame to frame. Its important you pick a feature in the picture which is easy to track. The best features are high-contrast corners or dots. For example the white painted window frames against the red bricks make really strong tracking targets, especially if you put the track point right on the corner of the window frame.
You can move each track point by clicking in the smaller square inside it. Move the mouse inside the square until you get a black arrow head and direction arrows. If you see a white arrow head, move the mouse a bit more until it is over the edge of the little square. You can then click and move the tracker. You should see a magnified section of the video to let you easily pick a good tracking target. Move it until the corner of the window paintwork is in the middle of the square like this:
Do the same for the other Tracking Points until you've got a large rectangle lined up with the background. Remember each tracker needs to be on a clearly visible corner in the picture.
When all the tracking points are positioned properly, its time to start tracking. Down in the Tracking panel at the bottom right there are some 'Analyse' buttons. Pressing the 'Analyse Forward' button will start After Effects looking through all the frames of video. The tracking points should stay stuck to their track targets in the video and follow them like they were pinned to the picture. At the end of the analysis, the path followed by the Track Points will be shown as a grey 'squigly' line:
If everything has worked, After Effects has tracked the video and created keyframes recording the movement of the track points through-out the video. If you click on the arrows on the video track to open up the properties, you'll see the Tracker and Track Points along with their keyframes in the timeline:
Section 2: The Titles
Now that we have a track of the video its time to create the titles. We'll be using a Text layer with Layer Styles to very quickly create solid looking text with shadows and highlights. By adjusting the lighting on the layer styles to match the real lighting in the video, a much more realistic and convincing effect can be achieved.
First off, make a new Text layer: Layer > New > Text
You can adjust the basic font settings in the Character panel which should be on the middle right of the screen. Because we'll be shrinking the text layer into the background of the video, we want to start off with the text as big as possible so that it fills the tracked rectangle. You can change any of the yellow coloured values just by clicking on them and moving the mouse left to decrease the value or right to increase it. To help the text to 'sit' in the picture better you should make sure you colour it to match the colours in your video. The buildings already have a dramatic colour scheme. The red bricks contrast nicely with the battleship grey of the metal cladding. You can use the 'eye dropper' tool to pick the grey colour for the text which will make it stand out against the brickwork in the background. Make sure you select the text before you make changes to the font.
After adjusting the size and colour of the text, you should have something like this:
At the moment its a bit dull and flat, but its easy to make the text look more 'solid' with some layer styles. Go to the menus and find Layer > Layer Styles > Bevel and Emboss, this will put shadows and highlights around the text edges to make look less flat. Next put a drop shadow behind the text with Layer > Layer Styles > Drop Shadow. Both of these effects take light direction into account. To make the text look like its really in the video, we need to adjust the light direction to match the real world. Its obvious from the video that the sun is coming in from the right and the shadows are pointing to the left. The properties for the effects are accessible in the bottom panel next to the time line. Click on the arrows to open the properties for each effect and change the Angle value to match the lighting.
Play around with the other properties, especially Distance and Opacity until you get something like:
Stage 3 Corner Pin
To create the illusion that the text is floating in the video, we have to "map" the text layer onto the rectangle made by the Tracker. The Corner Pin effect does just what it says -- it pins the corners of a layer on to four points on the screen. If the points move, the layer will follow. The Tracker has already created keyframes for its four corner points under the Feature Centre property. All we have to do is add a corner pin effect to the text layer, then literally copy and paste the keyframes from the tracker points to the corner pin effect. First, click on the text layer to select it then go to Effect > Distort > Corner Pin
Open up the properties of the text layer and expand the Effects section, then Corner Pin.
The four corners of the text layer are set to the corners of the composition size ( 0, 0 for the top left and 1280, 720 at the bottom right). Scroll down to the video layer and expand its properties until you find the first Tracker Point's Feature Center keyframes. To select all the keyframes, just click on the name of the property. The selected keyframes will be coloured yellow.
Then just go to Edit > Copy or press CMD C go back to the Corner Pin effect and click on "Upper Left" and paste the keyframes ( Edit > Paste or CMD V ). The Tracker points and corner pin points are in the same order so copy and paste keyframes from each one in turn.
And that's it! If you play back the composition, the text should move along with the video as if it was always part of the picture. The last thing to do is render out a video to edit into your film.
Here as promised is an After Effects tutorial to recreate the 'floating title' effect from last week's workshop. Before we begin, have another look at the final result as a reminder:
Stage 1: Tracking the Camera Move
Open After Effects and make a new Composition matching your video settings
Import your video file and drag it onto the composition.
You'll need to add a Camera Tracker to the video by clicking on the 'Tracker' tab of the bottom right panel. If there is no 'Tracker' tab, go up to the Window menu and make sure 'Tracker' has a tick next to it.
There are two basic options: 'Track Motion' and 'Stabilize Motion'. Stabilize lets you take shaky, hand-held video and stabilize it to look like it was shot on a tripod. It works quite well but is not going to give you as good results as a proper tripod shot -- there are no short cuts in cinematography!
We're going to use the first option. Click the 'Track Motion' button and the rest of the panel will switch on. Because there is only one video track the 'Motion Source' option should automatically select your video. The only option we need to change is 'Track Type'. Set this to 'Perspective Corner Pin'
You should now have a tracker called 'Tracker 1' on your video and some extra buttons will have appeared under the picture.
The tracker is made up of four labeled 'Track Points' at the corners of a white square. The track points are used to mark an object in the video so that After Effects knows what colours and shapes to track from frame to frame. Its important you pick a feature in the picture which is easy to track. The best features are high-contrast corners or dots. For example the white painted window frames against the red bricks make really strong tracking targets, especially if you put the track point right on the corner of the window frame.
You can move each track point by clicking in the smaller square inside it. Move the mouse inside the square until you get a black arrow head and direction arrows. If you see a white arrow head, move the mouse a bit more until it is over the edge of the little square. You can then click and move the tracker. You should see a magnified section of the video to let you easily pick a good tracking target. Move it until the corner of the window paintwork is in the middle of the square like this:
Do the same for the other Tracking Points until you've got a large rectangle lined up with the background. Remember each tracker needs to be on a clearly visible corner in the picture.
When all the tracking points are positioned properly, its time to start tracking. Down in the Tracking panel at the bottom right there are some 'Analyse' buttons. Pressing the 'Analyse Forward' button will start After Effects looking through all the frames of video. The tracking points should stay stuck to their track targets in the video and follow them like they were pinned to the picture. At the end of the analysis, the path followed by the Track Points will be shown as a grey 'squigly' line:
If everything has worked, After Effects has tracked the video and created keyframes recording the movement of the track points through-out the video. If you click on the arrows on the video track to open up the properties, you'll see the Tracker and Track Points along with their keyframes in the timeline:
Section 2: The Titles
Now that we have a track of the video its time to create the titles. We'll be using a Text layer with Layer Styles to very quickly create solid looking text with shadows and highlights. By adjusting the lighting on the layer styles to match the real lighting in the video, a much more realistic and convincing effect can be achieved.
First off, make a new Text layer: Layer > New > Text
After adjusting the size and colour of the text, you should have something like this:
At the moment its a bit dull and flat, but its easy to make the text look more 'solid' with some layer styles. Go to the menus and find Layer > Layer Styles > Bevel and Emboss, this will put shadows and highlights around the text edges to make look less flat. Next put a drop shadow behind the text with Layer > Layer Styles > Drop Shadow. Both of these effects take light direction into account. To make the text look like its really in the video, we need to adjust the light direction to match the real world. Its obvious from the video that the sun is coming in from the right and the shadows are pointing to the left. The properties for the effects are accessible in the bottom panel next to the time line. Click on the arrows to open the properties for each effect and change the Angle value to match the lighting.
Play around with the other properties, especially Distance and Opacity until you get something like:
Stage 3 Corner Pin
To create the illusion that the text is floating in the video, we have to "map" the text layer onto the rectangle made by the Tracker. The Corner Pin effect does just what it says -- it pins the corners of a layer on to four points on the screen. If the points move, the layer will follow. The Tracker has already created keyframes for its four corner points under the Feature Centre property. All we have to do is add a corner pin effect to the text layer, then literally copy and paste the keyframes from the tracker points to the corner pin effect. First, click on the text layer to select it then go to Effect > Distort > Corner Pin
Open up the properties of the text layer and expand the Effects section, then Corner Pin.
The four corners of the text layer are set to the corners of the composition size ( 0, 0 for the top left and 1280, 720 at the bottom right). Scroll down to the video layer and expand its properties until you find the first Tracker Point's Feature Center keyframes. To select all the keyframes, just click on the name of the property. The selected keyframes will be coloured yellow.
Then just go to Edit > Copy or press CMD C go back to the Corner Pin effect and click on "Upper Left" and paste the keyframes ( Edit > Paste or CMD V ). The Tracker points and corner pin points are in the same order so copy and paste keyframes from each one in turn.
And that's it! If you play back the composition, the text should move along with the video as if it was always part of the picture. The last thing to do is render out a video to edit into your film.
Subscribe to:
Posts (Atom)