nav-left cat-right
cat-right

Audio Pull up/down

This post is extremely post production related.  So I apologize in advance if you are not in audio post.  I would suggest however if you are a student, learn as much about video as you can now, because chances are you’ll end up in post anyway!

I. Frame Rates

The are several frame rates that you will come across in the post production world.

The most common ones are 29.97(NTSC), 23.98(NTSC), 24(Film), and 25(PAL).  When a picture gets transferred from one speed to another the audio will have to change speeds as well (**Note: this is not always true now with some advanced video transfers, but for the sake of this tutorial I will assume traditional speed changes are taking place)

29.97fps and 23.98fps are both slowed down .1% from real time (from 30 and 24 fps respectively). This means that 1 hour of timecode is actually .1% longer than an hour of real time.  This gave birth to drop frame timecode which is used in television. I will not go into drop frame here.

PAL and Film run at the same speed as real time.  That means that an hour of timecode will equal an actual hour.

The table below shows the speed changes between the different frame rates. (please note: all PAL conversions in this tutorial assume that the picture department plays 25fps picture at 24fps or visa-versa. There are other ways to convert PAL to NTSC and PAL to Film that do not incur a speed change)

29.97 23.98 24 25 Destination
29.97 No Change 0.1% up 4.1% up
23.98 No Change 0.1% up 4.1% up
24 0.1% down 0.1% down 4% up
25 4.1% down 4.1% down 4% down
Source

II. Examples

Example #1 Film -> NTSC
To do pull ups and pull downs in Pro Tools we use the Sample Rate Conversion (SRC) feature on the import audio window. For our examples we will assume your source and destination are all going to be 48k.

1. Make a new session at the bit depth you are working at.  In the session setup window (command – 2 on the number pad) change your frame rate to 24.

2. Open the import Audio Window (shift-command-I) and choose the file you want to pull down.

3. Press the convert button, and check the box that says “Apply SRC” this will enable the “Source Sample Rate” dialogue.

4. Scroll to the bottom of the “Source Sample Rate” dropdown and you will see a category that says “48kHz (Pull Up/Down) Choose the appropriate setting which in this case is “47,952 – 0.1% Down (film to NTSC)”

5. Click Done and it will ask you where to save the file, and then it will import into your session.  You know have a file that is 48k, but plays .1% slower than the original.

III. Throwing PAL into the mix! PAL to Film Speed

(NOTE: In PAL 2 pops are supposed to be 48frames from FFoA.  If this is done correctly, you do NOT need to cut off head and tail pops as I explain below.  I have gotten a slew of projects where that is not the case, so I use the method below.  See comments for a discussion on the topic.)

Here it get’s a little more complicated.  Before you can do the conversion you have to take off any head or tail pops.  So cut your mix to first frame of action and last frame of action (when appropriate).  Then take that file and follow the steps below.

1. Make a new session at the bit depth you are working at.  In the session setup window (command – 2 on the number pad) change your frame rate to 24.

2. Open the import Audio Window (shift-command-I) and choose the file you want to pull down.

3. Press the convert button, and check the box that says “Apply SRC” this will enable the “Source Sample Rate” dialogue.

4. Scroll to the bottom of the “Source Sample Rate” dropdown and you will see a category that says “48kHz (Pull Up/Down) Choose the appropriate setting which in this case is “46,080 – (PAL to Film)”

5. Click Done and it will ask you where to save the file, and then it will import into your session.  You know have a file that is 48k, but plays 4% slower than the original.

6.  Now you will have to pitch your files up 4% to make up for the speed change. (Note: this is only done with 4% changes as a .1% change is not audible)

7. Add head and tail pops back on the file, and you’re all set to play you PAL audio at film speed!

NOTE: When importing audio into Final Cut watch for this bug.

35 Responses to “Audio Pull up/down”

  1. tim says:

    Why do you suggest removing head & tail pops before speed changing? In my experience this is the exact time when the head AND tail pops are of the most value ie to quickly check that the speed change is correct.

    If you speed change the whole session including head & tail pop, then you can verify first that the 2 pop is where it should be relative to the new video (and if it isn’t, select everything & move it all until the head pop is correct with the leader) then skip to the end & verify your audio tail pop matches the visual tail pop.

    I prefer to initially check this way as its quick & you arent trying to spot a drift in sync (eg if a 24fps speed change was actually meant to be a 23.976) – the tail pop is either right or it isn’t… I have had a couple of cases over the years where it took serious detective work to discover how or why something didnt sync correctly (in one case the material delivered to me had been speed changed twice!) This is where keeping the 2 pops AND the guidetracks in the session is invaluable – you need definitive sync references to compare.

    Cutting new 2 pops on a speed changed session means you have no way of verifying your speed change is correct!

  2. Major says:

    Thanks for you Comment Tim!

    The only time I suggest cutting off the 2 pop is when doing a 4% or 4.1% change. In this case the pop will not line up with the picture 2 pop because it will be longer than 2 seconds before the picture. I suggest cutting to FFOA because that is the only absolute reference point after the speed change.

    At least the pictures I have gotten don’t have the visual 2 pop (or rather the space between the 2 pop and FFOA) changed 4% as well. Is stretching or shrinking the 2 pop along with the picture common practice that i have not come across, it is entirely possible?

    When doing a .1% changes the 2 pop is still relevant and I definitely keep it.

  3. Ric says:

    What does 4% equate to in cents?
    Looking for how many "cents" = 4%
    Thanks

  4. Major says:

    4% is about 70 cents

  5. rabbit says:

    You are saying that there is no speed change between 29.97fps and 23.98fps.
    So if i capture 29.97fps from tape and import it into a project w, 23.98fps timeline, the 29.97fps audio be in sync with 23.98fps source? 29.97fps file doesn't need to be conformed in anyway?
    Thank you

  6. Major says:

    That is correct… 29.97 and 23.98 are the same speed. In Pro Tools you can run 23.98 in a 29.97 session and vice-versa. The frame rate indicator on the video track will still turn red (as the session frame rate and the video frame rate are different) but everything will play at the correct speed.

  7. Erik says:

    What about going from 25 to 23.98. seems like it would be the same process but with a 4.1% pull down. If that's correct, what would the pith correction be?

  8. Major says:

    Yup 25fps to 23.98fps is 4.1% down then you'd have to pitch it back up 4.1% (or just 4% nobody can hear the .1%)

  9. saulsas says:

    I have a dude, in your first example. I suppose that you converting from 24 to 29.97 fps. Right? in this case my new sesion rate setup must be 29.97? thanks

  10. Major says:

    Actually Session setup frame rate doesn't matter. But I usually set the session to the target frame rate(in your case 29.97), so I can check it against picture etc.

  11. saulsas says:

    Nice, thank you

  12. Carlos says:

    When you work in PAL, pops must be located at 48 frames (and NOT 2 seconds) from FFoA and LFoA respectively, so sync is maintained after the pull.
    Nice blog, by the way.
    🙂

  13. Major says:

    Hey Carlos…
    Yeah I thought that's what the deal was, but then I started getting things that had a 2 pop 50 frames from FFoA… so each time I do it, I ask now where they want the pop. I don't work in PAL that often, but it seems to change from project to project. I guess some of the editors over here don't know the "correct" way…. so I just use FFoA as a reference point, and ask them where they want their pop.

  14. Carlos says:

    Yeah, that happens a lot even in places where PAL is the standard. Communication with the editing department is crucial.
    If everything fails, syncing to FFoA's and LFoA's is not a bad idea at all. The problem could be with roll #1, wich usually fades from black to the initial credits, and so FFoA is not as certain. The same problem applies with tail pops and LFoA's in the final reel.
    Greets!

  15. Dave says:

    I just received some audio for a project that was shot at 25fps on a canon 5d. On the first day of the shoot, they had the framerate on the sound devices recorder set to 30fps instead of 25 fps. is there a sample rate conversion that i should use to correct this? all audio was recorded at 48khz.

    thanks

    Dave

  16. Major says:

    Should be fine. The files from the first day are just stamped with 30 frame TC instead of 25. This doesn't effect the speed of the audio as long as everything was recorded at 48k.

  17. Rob says:

    I have source audio material at 23.98 and I need to convert it to 25fps. Tried several ways and the audio does not maintain sync. Please advise.

    Rob

  18. Major says:

    Should be a 4.1% pull up (NTSC->PAL (Film style))

    if that doesn't work… you have to contact the picture department and find out how exactly they went about doing the 23.98->25 conversion… there are some methods that maintain sync.

  19. Jack says:

    OH!!! Thank goodness!!! I was struggling like crazy with audio that slightly drifted out of sync, and you SAVED THE DAY!!! success. I am so thankful for your post. MANY MANY thanks.

  20. chris says:

    I recorded audio synced to 29.97DF. The cameraman was filming in 23.97. I bounced a mix and I have been informed that it lags at a certain point. Is this possible? In final cut the audio is being synced to slate not code.

  21. jonas says:

    could this be your solution, chris?

    http://www.youtube.com/watch?v=s5VLHDLiIHw

    it deals with a certain sync-bug in final cut.

  22. jonas says:

    could this be your solution, chris?

    http://www.youtube.com/watch?v=s5VLHDLiIHw

    it deals with a certain sync-bug in final cut…

  23. jonas says:

    great post, major. i was curious about just this.

    could i ask what pitcher you would use in pro tools for the 4% up? pitch shift or time shift? polyphonic or rhythmic? and mathematically, 4% up would mean 104% pitch right? which equals to 0.68 semi-tones, am i doing it right, then?

    thanks in advance for any help!

    /jonas

  24. GINO says:

    hey major!!! is posible complete de table whith the pitch cents for all situations
    i need to pull down 25 to 24, ¿ is the pitch correction +70 cents exactly) ¿ because is aprox?
    thanks for you time!!

  25. Jim says:

    Hey Gino, We use Pitch ‘n Time for our speed changes. As always, do a confidence check on the converted files. Listen out for pure notes, high notes in the music as these tend to go bonkers with a less than perfect time change.

    Great blog!

  26. Rupjit das says:

    Hi.. I m convert audio 24 to 25 but its pitch little bit high.. & when i pitched it -71 cents then pitch problem solved but at many places sound is phasing .. So whats the actual procedure… ?

  27. Major says:

    @Rupjit
    That is the correct process… try pitch shifting the stems individually instead of the whole mix, that usually sounds better.

  28. Tom Neary says:

    This table saved my butt ! i had footage in 25fps and recording audio externally. The rest of my footage was 24fps so to stop rendering, everything has to be the same fps, managed to change the footage to 24fps but the external audio was out of sync, after looking at the table at putting my audio at 96% (taking off the 4% suggested) it works perfectly. So in short serious Kudos !

  29. Floris Werver says:

    hey guys could someone explain how to get audio shot at 25fps /48khz sync with an image (movie) shot at 40fps? I use protools and tried TCE but it just doesn’t sound good…
    i’ve tried importing it with src at -60% but it’s way to long then… *it does sound better then TCE*

    we want it to be a bit slowmo.. but I can’t seem to find the right way to go about this…
    please help!

  30. Cat says:

    Thank you very much for this – a great help for a first timer!

  31. sam says:

    Hey Guys,
    I’m starting the process of edit and then mix a 90 minute movie and I’ll need to convert it from 25f to 24f using pro tools SRC.
    Should I create separate stems of dialogue, music and sfx and do the process from them and then mix them or can I SRC the full mix one mono at a time.
    I’m geting a bit stuck on what is the best process.

    Please advise on what is the best workflow.
    Thank you for the great info on the post.

  32. Maurício says:

    Men, many thanks!!! You saved my day 🙂
    I recorded the dubbed version of a film, and today i received a tv promo of that film and had tto convert a song for that promo! With your help everything went just fine!

  33. Steven says:

    Thanks for the info. One Question- when bouncing this new 48k file, what sample rate do I set the bounce to in Pro tools?

  34. Markus says:

    I’m mixing a 23.98 session. Picture is at 23.98, but I also have to deliver one audio file of the mix that will be married to 30 FPS picture. What is my conversion procedure?

    Thanks!

    BTW – It sound like I’ll never get a 30FPS version of the picture.

Leave a Reply

Your email address will not be published. Required fields are marked *