Syncing two tracks: methods, features and suggestions

Hi all,

OS. Slackware Gnu Linux 14.2 (the present stable release of slackware)
Audacity Version: 2.1.2

I’m trying audacity to synchronize an “italian” audio track against the original one in “english”. I have the english version of this movie (audio/video) and I want to add the italian language track.
The problem seems that the “out of sync” is not just “constant” nor “linear”, but changes “randomly”: if you synchronize an “event” at the beginning and an other near the end of the track, you will not obtain a fully synchronized track. Hope I explained that clear enough (sorry for my english).
I solved by working from the beginning of the track, synchronizing a first part, a second, and so on until the end of the whole track.
I used the Audacity “change tempo” effect to “stretch” or “compress” the part I was synchronizing. I used also labels to mark the “referring events” which the sync is based on.
The following points explain what I did exactly:

1- load “master track in english” (to not altered because it perfectly fit the video stream)

2- load the italian track to edit

3- looking for the first ref event: let’s name R1eng that event on the english track and R1ita the same event on the english track.
This event has to be as near as possible to the begining of the movie, it could be a big clean noise suhc as a gun shot, a “slam” of a closing door or something similar. Also a word, for example in my case I chose one of the first words spoken at the beginning, it was “alive” in english and “vivo” on the italian track).

4- I selcted that “R1ita” on the italian track and added a label for it (tracks - add label to selection). Then selected R1eng on the english track and add its label.

5- Finally I shifted the italian track so that R1ita and R1eng were exactly at the same time. For this the label R1eng it is very useful thanks to the “lighting alert” when the mouse pointer.
Then I had to add a silence at the beginning of the ita track or to cut its first exceeding part…
Now we have the first event in sync: R1 is at the same time on both the tracks and occur exactly at the same time of the set label R1eng. We can rename it “R1” and delete R1ita label.

6- I now check the tracks, look for an other event out of sync “R2” and set labels “R2ita” and “R2eng”.

7- This time we can’t shift the italian track, otherwise the R1 would go out of sync. We have to keep R1 in place and “stretch” or “compress” the italian track so as R2ita = R2eng.
To achieve that I used the following operations:

  • select the portion R1-R2eng
  • effects, change tempo
  • annotation of the time duration R1-R2eng
  • this will must be the time duration of R1-R2ita so that R2ita=R2eng…
  • exit from the change tempo dialog window by pressing “cancel”: we opened this effect just for reading present duration with the needed precision “seconds,cents”…
    There must be a smarter way to obtain that value, but I didn’t know none…
  • select R1–R2ita, open change tempo effect and set the new duration “seconds,cents” annotated before and finally apply the effect.

8- Now R2ita should coincide exactly with R2eng. We can delete R2ita label and rename R2eng to R2.
And so on, looking for R3 out of sync, labels, keep in place R2, drag R3ita so that is in sync with R3eng, etc, etc…

Following the above operations I obtained a new ita track synchronized with the original english video. Anyway I think there should be an easier method, especially to drag the end of the track portion out of sync and make coincide it to the corresponding event on the “master” track.

Change tempo effect is too uncomfortable to set (time annotation, portions selections and so on slow down all the work). I think it could be much easier:

  • set the two labels, for example R2ita on the ita track and R2eng on the english track
  • select the portion of ita track from R1 to R2ita
  • select a “change tempo tool”
  • click on R1ita extremity of the selected portion and hold down
  • drag the R1ita and place it at the same time of R2eng label

In this way there would no need to know any numerical duration of the portion R1-R2. Would be all just visually done and much speedy. I suspect should already be a similar way to work… But I don’t know audacity, I haven’t used it since a week ago. So I’m there to asking for suggestions or explanations.

Could anyone suggest a better method to do this tracks sync operation with audacity?


PS. 1
Other software suggestions would be welcome too, if you had any idea about a more suitable tool for this purpose.

PS. 2
I think this topic it isn’t related to a specific OS. Anyway I’m using Linux, so I posted in this section. It could not be the most appropriate (perhaps in windows section or mac-os was seen by more users).
In case moderators please move this to the right section.

PS.3
Sorry for the length of the post!

Sorry for the length of the post!

That’s OK. Way better than a three-word complaint.

As I read through that my mind flashed back to a friend of mine who does translation for the Court in Los Angeles. He does English/Spanish/English. Every so often, they’ll get a “fresh” judge that doesn’t understand Spanish doesn’t get to the point until the end and you just have to wait.

That and occasionally need to deal with “Street Spanish.”

English/Latin languages work like that and there’s no convenient way around it. If your question was easy, we would have knocked it out a long time ago. Whatever you can get to work is the way it works.

Koz

If your question was easy, we would have knocked it out a long time ago. Whatever you can get to work is the way it works.

Hi!
And thanks for your reply. My english is not so good… So I’m to ask a confirm If I understood right:
The way I edited the track, I’mean through “tempo changing” and it’s numerical setting, is the only way to “stretch” or “shrink” it…
Did you mean that?

PS.
I didn’t clearly understood your “judge story”…
Sounds like my prologue was too “verbose”! :smiley:
Wasn’t it?

Change Tempo may be the only way to force matching between the two languages. You have to be careful not to make the changes too large because the meaning of the phrases may get lost.

It can be difficult for English speakers to adapt to a different language because foreign rhythm, speed and emphasis can have important meaning, not just the words. Announcer English is difficult to learn, but anyone can use broken English, get most of the words wrong and still get the meaning through.

Change Speed would be perfect for you because you can type the new and old length and the tool will automatically match them. But, Change Speed also changes the pitch of the voices, so the voices would go up and down.

Can you get someone to live translate to the original show and record a whole new track? That may be easier than what you’re doing.

Koz

Ok, my question was a bit different…
“Change tempo” is the right way to edit my track. I have to maintain the pitch.
Audacity has that feature, so ok, it works!
But, I found too complicate the way to apply it.

You wrote:

you can type the new and old length and the tool will automatically match them

Yes, I done in that way also to apply “change tempo”, but:

  • How to know the new length?
    Well, you have to get the right value by selecting the track portion as I explained, annotate that value and then paste it in change tempo dialog window.
    Now I’m asking:
    Is there a way to apply “change tempo” by dragging a selected portion of the track, without inserting any numerical value?
    By reading your above reply, I think the answer is: no, there isn’t any “non numerical” way to apply change tempo effect.

The point I wanted to highlight was that in theory there is no need to know any numerical length value to apply that effect: two labels would be just enough to stretch the selected portion of the track to the right synchro.
Is my above sentence clear enough?
Any way to apply the effect like that?

Any way to apply the effect like that?

Not that I know of. We should wait for one of the other forum elves. They may have other ideas.

Koz

Ok, thank you! :slight_smile:
Let’s wait for some other idea…

PS.
In the meantime, do you know a similar feature any other software provides?
For example: avidemux, cinelerra, ardour, kdenlive or something other… I don’t know those tools. This question here is a bit OT, anyway I think I’m not the first user who has the idea to mix a good quality video track (synced with its original language audio track) with a “non original” language audio track.

That isn’t the only example:
let’s immagine to record two tracks, one of a singer and an other of a guitar. There could be the case: the singer keeps time right, but a guitar phrase is out of time.
Change tempo effect can do the magic by selecting the out of tempo part and stretch or shrink it.
Perhaps that’s a more familiar example for audacity users experience…

Thanks a lot again! :smiley:

I tried Ardour to test its “stretch mode” feature.
It is very close to Audacity “change tempo” effect, but it allows to stretch a selected track “region” just by click near the border of that region, hold left mouse button and drag to stretch or shorten it. When you release the button, a dialog window appears in which you can also insert the new length numeric value, but your can skeep this specification just by pressi “OK”. So it is possible in Ardour to “stretch” a selected region without inserting any numeric value related to the new length.

That could be a “new feature request” for Audacity developer.
I think Audacity is more recommended for simple tasks like the one I described. Ardour seems a very well done piece of software, but it provides a full DAW which is a bit too “expensive”, it has many features and so on… but It seems a bit slow in track importing, It doesn’t support mp3 and so on… so I think Audacity would be more fit for purpose.

How to submit this new feature to Audacity developers?

Wiki Feature Requests has:

More region/clip dragging functionality:

Time Stretching: (7 votes) mouse tools for both pitch-variable and pitch-constant time stretching. Adobe Audition has this and it’s vital for the production of sample-based music.

Is that what you would like to vote for?



Gale

Yes!
If I well understand, I think that’s the same feature I’m described in my previous posts.
Is there a way to see more detailed description of that feature request?

Thanks!

I added your “vote”. The details that we have are there. Most people who ask for it refer to the Adobe tools, and I assume the core method in Ardour is similar.


Gale

Thank you!
Is there a way to stay update and see when this new feature will be added?

The feature only has eight votes. We have very few developers available at present. So the feature almost certainly won’t be added in the near future, unless a developer already has an interest in it.

You can follow Audacity on Google+, Twitter and Facebook (see the buttons top left). Or there is RSS http://www.audacityteam.org/feed/.


Gale

Thanks a lot for those infos!

Is this any helpful?

Steps:

Open the Audacity program. …
Record your original track. …
Input additional tracks as necessary. …
Select your additional tracks to align them with your initial track. …
Use an align command to effectively move your selected track. …
Play back the track and make sure your sounds are placed correctly.

I am new here. Please be kind.