The approach is simple:

Record the replacement dialog while recording *and the video playing* at half speed. So the actors mouth moves at half speed or whatever speed you define, and thus making it easier to synchronize the voice with the image.

So the narrator who is replacing the dialog speaks slowly at half speed as well,matching the video.

When played back at full speed, everything should sound normal but with a much more precise ADR!

 

What you really need is a DAW with ADR (Automated Dialogue Replacement) capability. There may be others on this forum who have recommendations about this.

Another approach is going back to technology from a few decades and use a reel-to-reel tape recorder to record the actor at 7.5 or 15 ips and then playback at 15 or 30 ips, depending on the recorders tech.

Oh! So there is DAW that support ADR? or Dubbing rather? (that is what I am doing, dubbing)

Anyone have any recommendations?

@ito-b ... please let us know how this works out it you thus find a way to record at a slower speed. It's a novel and thus intriguing approach.

Do you have a narrator who is able to speak that slowly with the correct inflection? You can't record at half speed. You can playback at 2x speed or play video at 1/2 speed and have your narrator speak in sync. Then speed the narrator up 2x, but it will never sound normal.

@Former user +1

I agree.

On paper, it sounds like an ingenious idea.

In practice, I bet it's all but impossible to actually pull-off and still sound natural.

Not only would your Narrator would have to speak at half speed, but also pitch their voice down one octave. The idea seems straight forward, but is in practice ludicrous.

Not only would your Narrator would have to speak at half speed, but also pitch their voice down one octave. The idea seems straight forward, but is in practice ludicrous.

Simple: The editing software would compensate for any pitch variations...kind of like we do with time stretching today while keeping original pitch. (I am thinking about Zplane elastic, a high quality algo)

And it doesn't have to be at half speed. It could be a percentage, like 70% of original speed.

I am surprised I am the only one who though of it.

I think no one does this, because it isn't effective. Seems like a lot of extra unneeded work * . People been doing ADR forever without needing to resort to this concept. Find someone better at ADR. Or a better editor. It's very simple to push/pull specific words around in a sentence to match lip movement, in an edit. It's done all the time.

* if it's just for fun, experiments, have fun with it. :)

@ito-b ...

...kind of like we do with time stretching today while keeping original pitch. (I am thinking about Zplane elastic, a high quality algo)

In that case, you do not need to record at less than 1x. Record the voice at 1x in any audio recorder like Vegas Pro, Sound Forge Pro, etc while the actor follows the speed-reduced video original - and then use Zplane elastic (elastique TimeStretch in VP) to speed up the new recording to match the normal speed original video, adjusting pitch as needed.

BTW, this is not an ADR process. ADR is a specific computer process where the ADR algorithm compares and matches the audio characteristics of the new recording (e.g waveform positioning of the peaks and troughs) to the audio characteristics of the original recording. From your description, what you are aiming to do is a variation on the many decades old pre-ADR process of 'post-syncing'.

Your optimism about this approach is commendable, but you really should experiment with your own voice first before maybe paying an actor for a recording session that is experimental at best.

Again, please let us know how it works out.

@ito-b ...

...kind of like we do with time stretching today while keeping original pitch. (I am thinking about Zplane elastic, a high quality algo)

In that case, you do not need to record at less than 1x. Record the voice at 1x in any audio recorder like Vegas Pro, Sound Forge Pro, etc while the actor follows the speed-reduced video original - and then use Zplane elastic (elastique TimeStretch in VP) to speed up the new recording to match the normal speed original video, adjusting pitch as needed.

BTW, this is not an ADR process. ADR is a specific computer process where the ADR algorithm compares and matches the audio characteristics of the new recording (e.g waveform positioning of the peaks and troughs) to the audio characteristics of the original recording. From your description, what you are aiming to do is a variation on the many decades old pre-ADR process of 'post-syncing'.

Your optimism about this approach is commendable, but you really should experiment with your own voice first before maybe paying an actor for a recording session that is experimental at best.

Again, please let us know how it works out.

 

Sorry, I don't know the terms.

What is ADR?

What I am doing is taking a video, removing the original audio track, and recording my own in another language, with as close to possible matching the actor's lip movements.

What is called what I am doing?

ADR = Automated Dialogue Replacement. As outlined in my earlier comment, ADR utilizes computer intelligence to match dialogue recorded at a later time to the dialogue recorded with the original video recording. A typical example is where dialogue is recorded on location but the sound quality is poor/unsalvageable due to wind noise, location noise etc. At a later time in a better sound environment - perhaps a sound booth - the dialogue used on location is re-recorded while the same actor attempts to lip-sync to the original recording, but ADR fine-tunes that second read with the aim of exactly matching the timing of the original recording. I've also seen it very probably used in TV ads where a TV ad filmed in one country is used in another country, and a local actor in that 2nd country is used to replace the original actor's dialogue so that the dialogue for the 2nd country is in the 2nd country's local accent.

If you are not using ADR as it seems, then you will be using the decades old post-syncing (or dubbing as it is sometimes called) process where it is simply the 2nd dialogue read being done with the actor trying to match as closely as possible the original dialogue, but there is no computer assistance thereafter, maybe just an audio editor physically cutting tape in much earlier times.

ADR = Automated Dialogue Replacement. As outlined in my earlier comment, ADR utilizes computer intelligence to match dialogue recorded at a later time to the dialogue recorded with the original video recording. A typical example is where dialogue is recorded on location but the sound quality is poor/unsalvageable due to wind noise, location noise etc. At a later time in a better sound environment - perhaps a sound booth - the dialogue used on location is re-recorded while the same actor attempts to lip-sync to the original recording, but ADR fine-tunes that second read with the aim of exactly matching the timing of the original recording. I've also seen it very probably used in TV ads where a TV ad filmed in one country is used in another country, and a local actor in that 2nd country is used to replace the original actor's dialogue so that the dialogue for the 2nd country is in the 2nd country's local accent.

If you are not using ADR as it seems, then you will be using the decades old post-syncing (or dubbing as it is sometimes called) process where it is simply the 2nd dialogue read being done with the actor trying to match as closely as possible the original dialogue, but there is no computer assistance thereafter, maybe just an audio editor physically cutting tape in much earlier times.

Yep, so I am doing dubbing I guess since the audio I am recording is in a different language and this impossible to ADR.

Thanks for the epxlanation!

The term 'ADR' (Automated (or Automatic) Dialogue Replacement has been around since the early days of movie making. It was far from 'Automatic' though.