Workflow for multi-take music performance videos?

Hello,

I self-produce multi-camera recordings of classical music (concerts + multi-take sessions), from which I edit together the best audio takes in my DAW (Reaper) and then sync with video.

My workflow is tedious:

1. Record many takes of audio (DAW), and video, stopping/starting 4 cameras manually. Sometimes we leave them running bec it's too invasive to run around stopping/starting. We have no timecode.

2. Divide audio into takes (e.g. take 5 has 5 attempts, so we write take 5_1, 5_2, 5_3, etc.). In Resolve, match up the video to the audio. The timestamps on the videos/audios may help at least get the order right, especially when we left camera x off for a few takes.

3. In DAW, edit/mix/master the audio.

4. In Resolve, put in the edited audio and splice in video takes get a flow that has no sync issues.

Can anyone recommend a better workflow?

Thank you!!!
Larry

Hello again,

To be more specific, the main issue is that I get some big audio files which I break up into takes, and then I edit the takes into a nice audio flow...and then am stuck hunting through video takes to try to find corresponding bits for sync.

I thought of syncing all the audio with the video in Resolve, but I might want the better crossfade editor of Reaper....

So, how do people typically produce/sync (± classical) music videos?

Thank you!
Larry

Typically with the use of Timecode with the picture editor cutting with camera sound and the audio multitracks reconformed from the EDL for 4 point audio editing.

Typically with the use of Timecode with the picture editor cutting with camera sound and the audio multitracks reconformed from the EDL for 4 point audio editing.

Thanks so much, I have been reading about this and will keep reading. Is there ever a reversed workflow where audio multitracks are edited first and the video is somehow reconformed to match the audio? We care more about audio being "perfect" and are OK with changing some audio edits (or using B roll) if the sync requires it.

Thanks again for your reply.

Hello,

I do a lot of recording of dance and theater, multicamera (usualy 4) and multitrack (usualy 8).
The easyest and most secure way is to use timecode from my recorder (Zoom f8) to my camera through homemade XLR cables, channel 1 is TC, channel 2 is premix. If necessary, I use cheap Rode WirelessGo to send the TC to remote camera.
In DVR, you can easely synch multicam with TC (marking each camera and each track with the proper angle) and obtain a multicam file tat you can edit and mix keeping all in sync.
Working with audio waves is much more risky and prone to errors.
With free run TC showing time of the day, there is no ambiguity to resync files.
Sometime I give the multitrack file to the musician or the composer who make the proper mix, keeping the TC, so I can replace the multitrack by the right mix before editing.

Jacques

I think recording audio and filming a video are two separate activities and trying to combine them will always be a PITA.

Once you have the Master Audio, film the "lip sync" performance.

Is there ever a reversed workflow where audio multitracks are edited first and the video is somehow reconformed to match the audio?

Some do approach Multicamera CD production by recording audio performances and completing the subsequent 4 point music edit as usual, and then rehearsing with the completed audio edit and using playback during the video recording.

Hello,

Thanks for your replies. I wanted to take the time to do 2 of these edits, share my new process, and see how I should be doing this better - see below.

Btw I believe timecode wouldn't help that much because I'd still have to hunt around each video take. What we need is measurecode (the measure number in the music) in addition to timecode!

Lip sync is sadly not an option for these - we would have to "lip" sync very fast violin notes and other very difficult things.

0. Edit the audio, trying to make as few cuts as possible (or only very short cuts that don't change bigger-picture timing).
1. Assemble/sync multi-camera sequences representing all takes. Add Dynamic Zoom and transforms as needed.
2. Put the edited/mastered audio in a new timeline.
3. Put the multi-camera sequences in the new timeline.
4. In the DAW, look at which audio take is used for the first bit of audio. Find that bit of audio in the multi-camera video and insert+sync it. Repeat for every audio edit.
5. For the first video take used, open the multi-camera sequence and grade each camera angle.
6. Create stills for each camera angle. Open subsequent multi-camera sequences and apply those still grades to each camera angle.
7. Optionally, add custom grading to any clips that still don't match.
8. Since I'm creating several videos from a project that need to share an overall look: create shared nodes for any timeline-level color grades. Add them manually to each timeline.

In the past I tried grading just each clip used in the final timeline. But then if I have to change camera angles later, the grade needs to be changed. It's also harder to apply grades to multiple clips this way. That said, I have the same issue with Dynamic Zoom and Transform - if I later change a camera angle, I have to redo this. I could instead apply these by editing the multicam sequence, but then I risk doing postproduction work on angles that might not get used.

I do multicam video, (usually 6 or 8 cams,) along with multi-track audio, (up to 20 tracks). I make promo videos for local bands which consist of short excerpts of songs pieced together into a few minutes run time, similar to a movie trailer. I actually film while they do live shows, thus I let my cameras run until the batteries get low, change them on the fly and keep filming. This creates video files of around an hour or so each. Two of my cams are GoPros, so they create a bunch of smaller files, but I just deal with it. I do let all cameras record onboard audio, but I only use that for syncing up in post. To obtain multi-track audio, I have a bank of microphone splitters that all stage mics go into. One side of the split goes to my multi-track recorder and the other side goes to the house PA system.
Anyway, in post, I do the following.

For the Audio portion:
1. I load all multi-track audio files into Fairlight and proceed to enhance, compress, EQ, add effects, whatever it needs, to a final mix. (I am also a sound engineer.)
2. Before rendering in the delivery page, I listen through to choose which song or passage I want for the first excerpt in the video.
3. Once I have a certain passage in mind, (it might be just a verse or a chorus) I link all clips (tracks) to keep them in sync.
4. Once linked, I drag the beginning and end points to resize all audio tracks together down to the size of the excerpt, let’s say maybe 30 or 45 seconds worth.
5. Then I render that selection out to a single stereo wav file.
6. Returning to the same Fairlight timeline, I drag all clips back out to full length and resize again to the next section or excerpt.
7. Rinse and repeat for each song excerpt that I want to use in the video.
8. I might end up with six or eight new wav audio files that are ready for delivery.

For the Video portion:
1. My plan of attack is to create a separate multicam timeline for each song excerpt. Then place each of those timelines inside of one master timeline.
2. Before editing, I spend a bit of time figuring out which video files will go with each audio clip that I made above, making notes about camera angles and songs, and sometimes renaming files to avoid confusion. Some of the longer clips will be used in more than one multicam timeline.
3. I import clips as needed and create proxies. Most of my cams are 4K, but I render to HD 1080p for Facebook or YT. Thus I set my timelines to 1080.
4. For the first multicam timeline, I select all needed video clips along with one finished audio clip. I right click and choose “Create Multicam Timeline”. I sync with audio because I can’t afford to buy timecode just yet. Syncing with audio is not perfect. If one camera is farther from the source than another, it may have a bit of lag, but it’s close enough that I can tweak it manually by nudging clips.
5. I drag that timeline into my master timeline and start editing. I generally do color grading first. I choose “Open In Timeline” and go to the color page. Once graded, I grab stills to use in the next song timeline.
6. Once color corrected, I return to the master timeline and shorten those clips to match the length of the final audio clip. Then I proceed choosing angles and add some dynamic zoom here and there.
7. Anyway, you get the idea. Follow the same procedure for the other song clips/timelines. I stack all timelines end to end until I have a finished promo reel.

You can see a sample on my facebook page. Link in my profile. I'm just a beginner and not very good at color correction yet. But I enjoy it. I'm learning as I go.

If anyone knows a simpler method, please let me know.

Hello,

Working on another, this is still super painful - there has to be a better way!

Has anyone tried syncing the unedited audio with unedited video and using Fairlight for all the audio edits? My only question is, then how would I keep track of takes? (Right now in Reaper I set up a region for each take and export them using region IDs, so the artists can pick takes.)

Larry

Remember the film, The Graduate? Remember the single word of advice for the lost protagonist? Well, this time the single word of advice that will guide the rest of your creative life and is ignored at your peril: Timecode.

If you embrace that word, you still need to be able to identify where you are in the space-time continuum so provide your coordinates and define each movement precisely. Your epoch of sight and sound to be captured together depends upon the complexity and can encompass, for example, a song or a verse or a chorus or an instrumental interlude.

At its simplest and ideally, the stars in your universe are positioned and adequately lit, your sensors are in position with movements charted and propellants ready. God speaks and angels respond:
“Roll Sound” — “Sound Rolling”
(Your performers physically ready themselves to think of one thing, to do one thing; they’re at the starting line, in the blocks)
“Roll Camera” — “Camera Rolling”
(Your performers take a breath, feel their muscles ready to launch)
God names the Scene with the name of “The Song” and numbers the flight of fancy “Take 1” and someone in the Scene makes heard and seen a “Clap” of thunder like the cocking of a pistol
“Action”
A moment of anticipation, silence, intense focus, and someone raises a baton or lifts their head, and then each performer Fires and the race is on!
When they reach the finish, a long pause, a reflection, many conclusions formed and some spoken, the performers reach their own truth, “I messed up” “I felt good” “I can do better” “That was better” “Our best” “Can we go again?” “For fun”
And God speaks
“Cut” and the scribe writes:
“Good take”
God replies “one more time please.”
“It’s a wrap”

When it’s all over the historian puts all the evidence on the Timeline which has a Track for every voice, every instrument (or every microphone). And then begins the creation of the coherent story, pruning and pruning, and then sweetening what’s left until the witnesses are in awe.

But it all is made more easily possible because of Timecode and organization. Before Timecode, it took a miracle to tell a story. We still believe in miracles. Can you afford to ignore your best tools?

“Are you feeling lucky?”

Rick, reading your post, I was spell-bound. I agree that timecode is king. I don't think Resolve has great native tools for managing the audio from one take with the video from another, though. Is that what the OP is actually looking for? Not sure.

Rick, reading your post, I was spell-bound. I agree that timecode is king. I don't think Resolve has great native tools for managing the audio from one take with the video from another, though. Is that what the OP is actually looking for? Not sure.

Thanks, Steve. As for mixing the best audio with the best video from a different take, I learned the hard way in the early days not to do this even if it’s done all the time by others. There will be something that doesn’t quite work in my experience. Our brains are amazing detectives and making a jumble of audio and video will feel off.

I’ve lifted entire sections where the video and audio of one take are better than the video and audio of another take. Often the transition of the video appears to be perfect to the viewer because I switched angles. The switch in the audio requires great care to find the right frame to make the cuts. I’ve found you have success if you cut just before the beat. I’ve cut a few bars of music or half the song this way and both were imperceptible to the viewer. That’s why I shoot with two cameras, external audio, Timecode.

If you can’t have at least two angles, you can fake a second angle by zooming in or out and/or reframing at the point of the cut

The initial video sync actually isn't an issue. The thing I don't understand is how best to set this up in Premiere. Right now I:

0. Put audio in Reaper and split it into takes. Optionally, create a new multitrack project with a track for each take, so I can easily audition takes.
1. Edit audio in Reaper with very precise crossfades
2. Create multicam sequence in Resolve, synced using waveforms
3. Put edited audio soundtrack in Resolve
4. For each audio cut I made in Reaper, find the corresponding segment of video in the multicam sequence

#4 is what takes by far the most time. I basically end up redoing all the work of #1.

This time I tried a different workflow:

1. Create multicam sequence in Resolve, synced using waveforms
2. Duplicate multicam sequence for each take and trim, so there's one multicam sequence for each take
3. Optionally, create a Resolve timeline with a track for every take, so I can easily audition takes
4. Pick takes and stitch them together in Resolve with very precise crossfades

However, #4 is quite annoying to do even in Fairlight, for these reasons:

a. In Reaper, overlapping clips creates automatic crossfades. In Resolve, you have to create crossfades manually, and they get deleted if you accidentally move the clips so they don't overlap.
b. In Reaper you can adjust crossfade timing down to 1 sample, whereas in Reaper it's only down to 1 frame.
c. I might be wrong but I don't think Resolve can do equal-power crossfades. I hear a dip in SPL during crossfades.

By the way, Reaper can export Edit Decision List (.edl) format. I just don't know how to set up my audio/video files (and/or my Resolve project) so that I can take EDLs to/from Reaper.

b. In Reaper you can adjust crossfade timing down to 1 sample, whereas in Reaper it's only down to 1 frame.

Fades and Crossfades including the transition may be adjusted at the sample level on the Fairlight Timeline.

c. I might be wrong but I don't think Resolve can do equal-power crossfades. I hear a dip in SPL during crossfades.

Default fades are defined via Preferences > User > Editing > Default Fades or within the Fairlight menu's Batch Fades Settings window to set Equal Power or Equal Gain.

By the way, Reaper can export Edit Decision List (.edl) format. I just don't know how to set up my audio/video files (and/or my Resolve project) so that I can take EDLs to/from Reaper.

Reaper EDLs are formatted for Samplitude/Sequoia.

Resolve requires CMX 3600 EDLs and third party applications such as Vordio or AATranslator are required if using Reaper and a CMX 3600 EDL is required.

b. In Reaper you can adjust crossfade timing down to 1 sample, whereas in Reaper it's only down to 1 frame.

Fades and Crossfades including the transition may be adjusted at the sample level on the Fairlight Timeline.

c. I might be wrong but I don't think Resolve can do equal-power crossfades. I hear a dip in SPL during crossfades.

Default fades are defined via Preferences > User > Editing > Default Fades or within the Fairlight menu's Batch Fades Settings window to set Equal Power or Equal Gain.

Thanks. I set this to Equal Power, but new crossfades still appear linear (X-shaped). Also, classical audio editing I am very often nudging clips by 1ms at a time (rolling or not) to finesse crossfades, but when I nudge in Reaper, the crossfade disappears. Is there a way to fix this?

By the way, Reaper can export Edit Decision List (.edl) format. I just don't know how to set up my audio/video files (and/or my Resolve project) so that I can take EDLs to/from Reaper.

Reaper EDLs are formatted for Samplitude/Sequoia.

Resolve requires CMX 3600 EDLs and third party applications such as Vordio or AATranslator are required if using Reaper and a CMX 3600 EDL is required.

Thanks, the conversion looks straightforward. Can you just help me understand the workflow, since I'm new?

1. In Resolve, set up a trimmed multicam sequence for each take.
2. Export multicam sequences as WAV audio - separate for each track, somehow?
3. In Reaper, edit the multitrack and save EDL.
4. Somehow...change the EDL to refer to the multicam sequences instead of the audio?

Is this right?

Not sure about the workflow with Reaper as it is usually only a multitrack backup to the Pyramix system.

For Reaper you will have to either use Vordio or AATranslator to convert the Reaper Project in to a CMX EDL for use with Resolve.

My suggestion is still to edit with camera sound and reconform from an EDL (or AAF) using Timecode and complete the standard 4 point audio edit with the multitrack takes of each performance and then import all assets in to Resolve for delivery.

I record live classical music concerts with 5 cameras and usually 8 mics. As these are mainly amateur musicians the potential for 'slips' in the live concert I also record the rehearsals. To get the best musical performance I'm editing the audio in Reaper but using the superb ReaClassical extension that is designed for accurate 'source-destination editing', a la Pyramid/Sequoia. This keeps the all the original tracks in a 'Folder' stack during every edit. It is obviously desirable to not 'Ripple' the timeline when doing these audio edits which requires some time manipulation in Reaper to still match the original video at each edit. Dropping one of the videos into Reaper to check is possible if required. I ideally would now do an AAF format export out of Reaper as Resolve/Fairlight can import this but Reapers not capable and even Vordio seems to fail with that. AAF Translator may do it though.
I therefore export the mix or mix plus all the stems into Resolve, using the Fairlight page to put in place against the previously worked on Multicam edit I've be making in Resolve. Mixing the separate audio stems in Fairlight is something I've been doing since 1996, but alas Resolve's new version is different but still a good audio mixing tool. A bit of 'parallel compression,' does wonders to reduce the dynamic range of classical music in a totally transparent manner by the way.

In the above, I should have pointed out that I only have one video take, the concert and the rehearsals are in sound only. Do get ReaClassical if you edit in Reaper, as a source-destination editor it's the equal of Pyramix and also does RRP files for commercial CD production beautifully.

Thanks to Reynaud and Postfade!

I realized there are two types of video edits - single-take videos (live or otherwise) with occasional patches but overall timing maintained (these are fairly easy), and totally edited videos (start with take 5, cut to take 2, cut to take 11, …). Unfortunately I'm usually doing the latter.

It's easy enough for me to sync better audio with the video so I can make edit decisions based on the final, detailed audio.

Reynaud, I want to follow your advice but don't completely understand this:

My suggestion is still to edit with camera sound and reconform from an EDL (or AAF) using Timecode and complete the standard 4 point audio edit with the multitrack takes of each performance and then import all assets in to Resolve for delivery.

I think I lack the proper background info. Can you point me to anything that will help me understand how reconforming from an EDL works? Also,  when the footage has no timecode, is there a way to add it (approximately) based on time of day, or by manually syncing in Resolve (sliding clips around till it's perfect, then somehow embedding this as timecode in the source footage)?

The reconform process finds the difference between two picture tracks, creates a change list and then applies changes to audio and picture tracks. This may be a EDL or AAF depending on the tool.

Resolve currently provides a sub-set of this functionality with the Timeline Comparison window. Post-production audio workstations have this functionality built-in whereas the music oriented audio applications tend to require add-on software.

It is also possible to manually reconform with a 4 point audio editor to simplify and streamline the reconform process as the manual method is both tedious and time consuming. Opening multiple EDLs containing the various Takes and recording dates stacked within a single Project ensures this process is more efficient (especially with independent playback, scroll and zoom of individual EDLs).

If no Timecode is available a conform tool provides the option to match sound roll names or file names to match source audio to a picture edit.

Alternatively if you do not have the required software toolset and appropriate metadata, complete the multitrack audio edit first, and then start the picture edit with the final audio edit as template. This is the last resort once the other preferred options have been exhausted.

Thank you so much for the helpful replies.

I am now using this workflow:

1. Reaper: create regions for takes: Take 1, Take 2, etc.
2. Reaper: add several seconds buffer around each take (±3–5s) and export regions
3. Reaper: create a new project and edit the audio takes together
4. Reaper: export Samplitude EDL (.edl) and edit to contain just the one track.
5. Terminal: use a Python script to convert Samplitude EDL to CMX 3600 format. (see below)
6. Resolve: create multicam sequences for each take: Take 1, Take 2, etc., matching durations of step 1, and assigning Reel numbers 1, 2, etc. to each sequence.
7. Resolve: Import converted EDL with File > Import > Timeline.

This has already saved a TON of time and is working ALMOST perfectly!

But the later clips in the imported sequence are misaligned by as much as 5 frames. Any ideas why? Must be a bug in my Python script?

Python script:

Code: Select all

# pip3 install timecode

import argparse
import csv
import io
import re
from timecode import Timecode

def samples_to_timecode(samples, sample_rate, fps, start_hour=0):
    # Convert samples to total seconds
    total_seconds = samples / sample_rate

    # Initialize a Timecode instance with the given fps and start time
    tc = Timecode(fps, start_timecode=f'{start_hour:02}:00:00:00')

    # Add the total seconds to the timecode
    tc.add_frames(tc.float_to_tc(total_seconds))

    return tc.__str__()

def parse_edl_line(line):
    reader = csv.reader(io.StringIO(line), delimiter=' ', skipinitialspace=True, quotechar='"')
    for row in reader:
        return row

def extract_title(infile):
    with open(infile, 'r') as file:
        for line in file:
            if "Title:" in line:
                return line.split(':', 1)[1].strip()
    return "Unknown Project"

def extract_sample_rate(infile):
    with open(infile, 'r') as file:
        for line in file:
            if "Sample Rate:" in line:
                return int(line.split(':', 1)[1].strip())
    raise ValueError("Sample rate not found in EDL file.")

def write_clip_info(out_file, clip_info):
    # Extract reel number if clip name ends with digits
    match = re.search(r'(\d+)$', clip_info['name'])
    reel_number = match.group(1) if match else 'AX'

    out_file.write(f"{clip_info['event_number']:03}  {reel_number}       V     C        {clip_info['source_in_tc']} {clip_info['source_out_tc']} {clip_info['play_in_tc']} {clip_info['play_out_tc']}\n")
    out_file.write(f"* FROM CLIP NAME: {clip_info['name']}\n")

def convert_samplitude_to_cmx3600(infile, outfile, track_number, fps, start_tc_hour, adjust_overlaps=False):
    title = extract_title(infile)
    sample_rate = extract_sample_rate(infile)

    with open(infile, 'r') as in_file, open(outfile, 'w') as out_file:
        out_file.write(f"TITLE: {title}\nFCM: NON-DROP FRAME\n\n")
        process_lines = False
        event_number = 1
        previous_clip_info = None

        for line in in_file:
            if line.startswith("Track ") and process_lines:
                break
            if line.startswith(f"Track {track_number}:"):
                process_lines = True
                continue

            if process_lines and not line.startswith('#') and line.strip():
                parsed_line = parse_edl_line(line)
                if parsed_line and len(parsed_line) >= 16:
                    _, _, play_in, play_out, record_in, record_out, *_ = parsed_line[:6]
                    name = parsed_line[15].rsplit('.', 1)[0]

                    play_in_tc = samples_to_timecode(int(play_in), sample_rate, fps)
                    play_out_tc = samples_to_timecode(int(play_out), sample_rate, fps)
                    source_in_tc = samples_to_timecode(int(record_in), sample_rate, fps, start_tc_hour)
                    source_out_tc = samples_to_timecode(int(record_out), sample_rate, fps, start_tc_hour)

                    if previous_clip_info and handle_overlaps:
                        overlap_samples = int(play_out) - int(play_in)

                        if overlap_samples > 0:
                            # Adjusting the end of the previous clip to the start of the current one
                            previous_clip_info['play_out_samples'] = int(play_in)
                            previous_clip_info['play_out_tc'] = play_in_tc

                            # Adjust the source time out of the previous clip
                            previous_clip_info['source_out_samples'] -= overlap_samples
                            previous_clip_info['source_out_tc'] = samples_to_timecode(previous_clip_info['source_out_samples'],
                                sample_rate, fps, start_tc_hour)

                    if previous_clip_info:
                        write_clip_info(out_file, previous_clip_info)

                    previous_clip_info = {
                        'event_number': event_number,
                        'source_in_tc': source_in_tc,
                        'source_out_tc': source_out_tc,
                        'play_in_tc': play_in_tc,
                        'play_out_tc': play_out_tc,
                        'source_in_samples': int(record_in),
                        'source_out_samples': int(record_out),
                        'play_in_samples': int(play_in),
                        'play_out_samples': int(play_out),
                        'name': name,
                    }
                    event_number += 1

        if previous_clip_info:
            write_clip_info(out_file, previous_clip_info)

def main():
    parser = argparse.ArgumentParser(description="Convert Samplitude EDL to Resolve-compatible format.")
    parser.add_argument("--infile", type=str, required=True, help="Input EDL file path")
    parser.add_argument("--outfile", type=str, required=True, help="Output EDL file path")
    parser.add_argument("--track", type=int, required=True, help="Track number to process")
    parser.add_argument("--fps", type=float, required=True, help="Frame rate (no default value)")
    parser.add_argument("--start_tc_hour", type=int, default=0, help="Start time code hour")

    args = parser.parse_args()
    convert_samplitude_to_cmx3600(args.infile, args.outfile, args.track, args.fps, args.start_tc_hour)

if __name__ == "__main__":
    main()


The initial video sync actually isn't an issue. The thing I don't understand is how best to set this up in Premiere. Right now I:

0. Put audio in Reaper and split it into takes. Optionally, create a new multitrack project with a track for each take, so I can easily audition takes.
1. Edit audio in Reaper with very precise crossfades
2. Create multicam sequence in Resolve, synced using waveforms
3. Put edited audio soundtrack in Resolve
4. For each audio cut I made in Reaper, find the corresponding segment of video in the multicam sequence

#4 is what takes by far the most time. I basically end up redoing all the work of #1.

This time I tried a different workflow:

1. Create multicam sequence in Resolve, synced using waveforms
2. Duplicate multicam sequence for each take and trim, so there's one multicam sequence for each take
3. Optionally, create a Resolve timeline with a track for every take, so I can easily audition takes
4. Pick takes and stitch them together in Resolve with very precise crossfades

However, #4 is quite annoying to do even in Fairlight, for these reasons:

a. In Reaper, overlapping clips creates automatic crossfades. In Resolve, you have to create crossfades manually, and they get deleted if you accidentally move the clips so they don't overlap. I remember how important it was not only to create the creative product itself in a musical genre, but also to make it popular, promote it and so on. That's the basis and basis without which you just can't do it. That's why I'm very glad that here https://artistpush.me/products/50-sound ... list-likes I found marketing specialists in Soundcloud, who were able to collect a lot of likes, subscribers and real listeners for me. So in addition to creating a music product, don't forget about marketing.
b. In Reaper you can adjust crossfade timing down to 1 sample, whereas in Reaper it's only down to 1 frame.
c. I might be wrong but I don't think Resolve can do equal-power crossfades. I hear a dip in SPL during crossfades.

Very informative, I will be testing different ways too.