textsyncr
  • Tim BrookTim Brook
    Hi guys, just downloaded this, followed your directions which were all good (installed java, mono wav file, etc.) until I hit 'Analyze' and then AFX stalls (running latest version CC 2015.0). Any ideas?
  • Joachim HollerJoachim Holler
    What OS are you running?

    I have seen the engine having issues with specific wav compressions. Try to avoid 24 bit depth and more sample-rate wise.
  • Tim BrookTim Brook
    Hi Joachim, thanks for getting back to me. Its the latest Mac (OX 10.10.3 Yosemite). The WAV is 16bit (48kHz).

    Its quite a chunky bit of VO script though (938 words), would that impact TextSyncr?
  • Joachim HollerJoachim Holler
    Yes, it definitely could. Please try to cut out a 20 seconds piece or so, to find out if this causes the problem.

    Other than that, very bad audio quality (apart from the compression quality) can also lead to problems.
  • Tim BrookTim Brook
    I'll try a smaller segment. Does TextSyncr work with layers that have audio markers already on them (eg. I have some old voice over stuff I ran through Premiere's speech analysis to generate markers I can see in AFX)?
  • Joachim HollerJoachim Holler
    You mean with the captions tab? Yes, that works!
  • John WalkerJohn Walker
    I''ve been having issues with punctuation not being included in the markers after an analysis is complete. Is there something I can do to correct this since it then throws off my subtitles punctuation?
  • Joachim HollerJoachim Holler
    Hey John, 
    normally this should work as you expect, is there anything 'unusual' about the specific subtitles (eg. a space in front of every period)? What specific punctuation characters are being deleted by the script?
  • David CrowtherDavid Crowther
    Hi Joachim,

    I installed the full product, it says v2.2 is available, but when I check AEScripts I can only download V1.0. Help appreciated! 
  • Joachim HollerJoachim Holler
    Hey David, v1.0 is the current one, so you have the right one! I will check soon what is wrong in the backend that tells you there would be a newer version.
  • John WalkerJohn Walker
    Hey John, 
    normally this should work as you expect, is there anything 'unusual' about the specific subtitles (eg. a space in front of every period)? What specific punctuation characters are being deleted by the script?
    I took my script and pasted into text edit, then opened as a plain text document with no formatting and removed all of the extra characters. I think the issue is happening because it sometimes creates extra markers when analyzing, and usually around 15 seconds into the video. It sets markers one after the other and I've been able to manually delete them, but it's time consuming and, like I said in a previous post, it throws off my punctuation. image
  • Shawn HerbertShawn Herbert
    I keep getting this message. I'm using a 32bit mono wav file. I'm running Yosemite with Adobe AE CC 2015.

    There was an error executing the Aligner. Error details:
    Exception in thread "main" java.lang.RuntimeException: used file encoding is not supported
    at edu.cmu.sphinx.frontend.util.AudioFileDataSource.setInputStream(AudioFileDataSource.java:197)
    at edu.cmu.sphinx.frontend.util.AudioFileDataSource.setAudioFile(AudioFileDataSource.java:162)
    at edu.cmu.sphinx.frontend.util.AudioFileDataSource.setAudioFile(AudioFileDataSource.java:118)
    at edu.cmu.sphinx.demo.aligner.Aligner.align(Aligner.java:162)
    at edu.cmu.sphinx.demo.aligner.AlignerDemo.main(AlignerDemo.java:70)

    Thanks,
    Shawn
  • Joachim HollerJoachim Holler
    Hello Shawn,

    I just got a support ticket in yesterday from a user getting the exact same error. I will post a solution as soon as I have it!

    Joachim
  • Mark ChadwickMark Chadwick
    I am doing an evaluation of textsyncr. It seems to be working, but I have a couple issues:

    1) When it completes (successfully) a dialog box comes up to import a Premiere file. Though I simply cancel out of this with no problem, do you know the cause?

    2) It seems that the markers appear approximately 2 frames early for most words. (Of course my sample is a bit small due to the trial restriction of under 15 sec). Is to a way to adjust this? FYI, my audio is quite clean with good dynamic range.
  • Joachim HollerJoachim Holler
    Hello Mark,

    1) When the analysis is complete, a custom popup should come up. Does this come up instead or in addition to the other one? What OS and AE version are you running?

    2) The longer the audio file, the more exact the analysis usually is. There is no automatic way to move all the markers together, but you could maybe just drag it 2 frames on the timeline, mute it, and add the clip again without changing its in and out?

    Joachim
  • Mark ChadwickMark Chadwick
    Thanks for the support Joachim. I just converted a 5 minute VO. This will work nicely as a framework for upcoming project. I verified it works well with TypeMonkey and the other Monkey Tool called "Scene Maker."
  • jehonel portejehonel porte
    There was an error in the installation, now get this:

    imageimage

    help!!!
  • Joachim HollerJoachim Holler
    I will write you a PM!

    -J
  • Ricardo MendozaRicardo Mendoza
    Hey Joachim,
    I got this weird error: see screenshot
    I did everything as instructed... the audio file is about 3.5 minutes

    Any help would be greatly appreciated!!
    Screen Shot 2015-12-06 at 3.10.12 PM.jpg
    928 x 1350 - 249K
  • Joachim HollerJoachim Holler
    This is indeed a weird error. Could you try cutting the audio file into two pieces?

    J
  • Ricardo MendozaRicardo Mendoza
    I separated them into short 1 minute pieces and it sometimes it'll give me that weird error and other times it just crashes.
    Would really want it to work... any suggestions?
    Screen Shot 2015-12-10 at 3.14.55 PM.jpg
    834 x 476 - 72K
  • yasser alimyasser alim
    I installed trial version after i click on button , Erorr appears

    1.jpg
    1111 x 555 - 54K
    2.jpg
    844 x 603 - 43K
  • Joachim HollerJoachim Holler
    I will write you a PM!

    -J
  • Jim ZarbaughJim Zarbaugh
    Hey, Joachim! Thanks for the incredible script! I have a question about creating subtitles for a music video. The parameters seem to be working, but I would like the phrases to be divided naturally the way they occur in the song. Sometimes a short phrase can take a long time to ing, while other times the song can move quickly. The script breaks the phrases up evenly and doesn't allow me to change this easily. Is there some character I could add at the end of each phrase to tell the script where to end one phrase and begin another? That would make this script so much more helpful for subtitles.

    I'm in a bit of a hurry to get about twenty-five videos subtitled, so I understand if you can't get back to me in time to help. Any help you can get me would be great. This will not be my last subtitling gig.

    Thanks so much for your help!

     Jim Zarbaugh
    Plano, Texas
  • Joachim HollerJoachim Holler
    Hi Jim,
    do you mean the breaks between the segments or the syllables?

    Because the analysis can only recognise words as a whole and not syllables, there is simply no data to break the syllables automatically.

    If you mean the breaking of segments, the script tries to break after specific characters (.,)] and some others), but does not include time into the equation. Depending on how important this is for you, one could introduce a sequence of characters (like "<break>" for example) and write a script that goes through the markers, checks for that <break> and then adjusts the breaks if it finds one…?
  • Jim ZarbaughJim Zarbaugh
    Joachim,

    I really don't need a break in between syllables. I talking more about the natural phrasing you find in a song. For instance, here are the first lines in Billy Idol's "Dancing With Myself":

    On the floors of Tokyo

    Down in London Town’s a go-go

    With the record selection

    and the mirror's reflection

    I'm dancin' with myself


    The script will combine the first and part of the second line. But if you are singing along with the song, you don't sing half a phrase and then wait for the rest to pop up. You divide true subtitles with natural divisions like you see above. In reality, I would prefer to have lines three and four together since they are sung quickly in one breath. But if I set the script to do that, then I can't choose to have the shorter verses for line 1 & 2.


    I think a script that allows you to choose the breaks (either by typing something such as "<break>" or simply by pressing "return" or "enter" to create the next line as you see above) would make this script much more useful. The more control, the better. Right now, I can only use the script to create the markers and then hand-place each subtitle where the markers indicate. I'm not sure it saves me much time yet, but it is pretty amazing.


    Thanks for responding so quickly.


    Jim





  • Joachim HollerJoachim Holler
    To all of you guys who are interested in doing something like Jim - we implemented this feature to version 1.2, which is now available for download. You can now tick a checkbox if you want to control the phrase breaks via the line breaks in your text!
  • Martine KerrMartine Kerr
    Hi, I am a total newbie with all of this kinetic text and audio tracking but am excited to learn. My first attempt howerver gave me the same error I've had the same error
    There was an error executing the Aligner. Error details:
    Exception in thread "main" java.lang.RuntimeException: used file encoding is not supported
    at edu.cmu.sphinx.frontend.util.AudioFileDataSource.setInputStream(AudioFileDataSource.java:197)
    at edu.cmu.sphinx.frontend.util.AudioFileDataSource.setAudioFile(AudioFileDataSource.java:162)
    at edu.cmu.sphinx.frontend.util.AudioFileDataSource.setAudioFile(AudioFileDataSource.java:118)
    at edu.cmu.sphinx.demo.aligner.Aligner.align(Aligner.java:162)
    at edu.cmu.sphinx.demo.aligner.AlignerDemo.main(AlignerDemo.java:70)

    So I saved a 23s segment, went down to 41HZ from 48HZ, 16 bits instead of 32...and it seems to be working (no more error and found 100% of my words). Unfortunately, the script seems to be cramming words past the 15s mark all together.

  • Joachim HollerJoachim Holler
    So first it says that it found 100% and then on the same audio layer it crams the markers?
  • Tony StechTony Stech
    Hello!

    I just purchased the script. I have my mono wav file selected and I pasted the text into the text area. When I click the analyze button, I'm getting the "There was an error executing the Aligner " error. What can I do to fix this? I'm using AE CC 2014 and I've reinstalled the script twice with no luck.
  • Joachim HollerJoachim Holler
    Hi Tony, I will write you a PM.
  • Atanga MungandiAtanga Mungandi
    Hello same issue as tony Stech!

    I just purchased the script. I have my mono wav file selected and I pasted the text into the text area. When I click the analyze button, I'm getting the "There was an error executing the Aligner " error. What can I do to fix this? I'm using AE CC 2015 and I've reinstalled the script twice with no luck.
  • Joachim HollerJoachim Holler
    What exactly does the error say? Does it appear right after you hit "analyse" or does it take a minute before it pops up?
  • Andrew GoodmanAndrew Goodman
    hi, i just downloaded this as its exactly what i need but installed java, mono wav file, etc hit 'Analyze' and then ae stalls (running latest version CC 2015.3 but also tried it on cc 2014). Any ideas? i even tried to cut down my audio and everything... then recorded me saying 'one two three four' and the markers go in completely the wrong place (all at the start no where near the audio), any help would be much much appreciated
  • Joachim HollerJoachim Holler
    Hi Andrew,

    there are a few reasons why this could happen. Most likely the file is either too long (cut in between words) or changing the compression might do the trick:

    Try 32 bit sample rate (NOT float) in the WAV settings. Also make sure that in your recording there is a little space (half a second) empty before someone is speaking.

    Remember that accents, funny intonation, singing, or more than one speaker will mess up the analysis.
  • Andrew GoodmanAndrew Goodman
    hi joachim, thank you very much for your reply... i have done all of the above but still no luck :( could it be my accent? here's a test clip i am trying to get working as the long one (2mins) stalls ae... this one completes the job but the makers appear in all first 2 secs way out of sync
    https://cl.ly/0W3T08391O1b
    thanks
  • Joachim HollerJoachim Holler
    Hi Andrew,

    so I downloaded your file and also made a few recordings myself. Mine worked, yours didn't (and I have a horrible German accent). Yours is 44.1 kHz and 32 bit float (which is a problem) but even when I convert it to 44.1 kHz and 16 bit, it still doesn't work. 

    I made a few recordings myself in Audition, speaking both faster as well as slowly like you did, and it got 100% every time.

    Apart from being 32 bit float, I can only imagine that your specific intonation/accent in that particular file is causing this. Maybe you could just make another one real quick!

    -J
  • Andrew GoodmanAndrew Goodman

    hi joachim,

    a MILLION thanks!!! this worked 100% i.e. i did my recording in audition and then brought it right into ae. 

    what i had been doing before (and this might help other people in future) was recording audio via my phone (samsung s6… bad idea i know but correct audio is music so i was just needing to speak the words etc) i transferred the audio onto computer and converted it in media encoder then bought that into ae. my guess was recording audio with my phone messed the whole thing up maybe? any way via audition works!

    again thanks so much for your help, much appreciated, this script is fantastic!

    andrew

  • Joachim HollerJoachim Holler
    Perfect! :)

    Not sure what exactly caused the problem, but it might have had something to do with the encoding.

    -J
  • Matthew PhelanMatthew Phelan
    This doesn't seem to work in latest version of AE. I have a clear Sound booth audio clip, and it simply doesn't match up. Even when exported as given preset output for encoder. Additionally Adobe Colour theme keeps opening up every time Analyze is pressed?????  ANY ideas?
  • Joachim HollerJoachim Holler
    (This issue was solved in a direct conversation)
  • John BrewsterJohn Brewster
    Besides the 16 bit depth and mono, are there any other audio tips or tricks? I am testing it with a short audio file and I keep stalling After Effects and am having to restart. At this point I am not sure if it will analyze the audio, but want to have it add all of the markers in spaced out evenly so I can move them into place myself. But AE can't seem to handle it. Is there way to use a dummy audio file to just get the markers in place?
  • Joachim HollerJoachim Holler
    Hi John,
    placing markers without analysis can't be done with textsyncr, but you could do that with lipsyncr and skip the analysis there. Also, several other scripts might do that for you.

    There is a timeout in the script which is set to 1 minute. If the analysis would take over 1 minute, it will stop, in order to prevent your computer from freezing completely. Have you waited for this long?

    If you're on Windows, don't worry if the progress bar doesn't display correctly. I have seen this happen as well.
  • Brandon HarrisonBrandon Harrison
    Hi Joachim,
    I am having a little trouble myself.  I am getting an error now myself.  I've used the Media Encoder preset you gave.  Even still - I keep running into the same issues.  Here is the error I keep getting: 


    There was an error executing the Aligner. Error details:


    Exception in thread "main" java.lang.OutOfMemoryError: Java heap space

    at java.util.HashMap.resize(HashMap.java:462)

    at java.util.HashMap.addEntry(HashMap.java:755)

    at java.util.HashMap.put(HashMap.java:385)

    at edu.cmu.sphinx.decoder.search.AlignerSearchManager.setBestToken(AlignerSearchManager.java:489)

    at edu.cmu.sphinx.decoder.search.AlignerSearchManager.collectSuccessorTokens(AlignerSearchManager.java:588)

    at edu.cmu.sphinx.decoder.search.AlignerSearchManager.collectSuccessorTokens(AlignerSearchManager.java:597)

    at edu.cmu.sphinx.decoder.search.AlignerSearchManager.collectSuccessorTokens(AlignerSearchManager.java:597)

    at edu.cmu.sphinx.decoder.search.AlignerSearchManager.collectSuccessorTokens(AlignerSearchManager.java:597)

    at edu.cmu.sphinx.decoder.search.AlignerSearchManager.collectSuccessorTokens(AlignerSearchManager.java:597)

    at edu.cmu.sphinx.decoder.search.AlignerSearchManager.growBranches(AlignerSearchManager.java:394)

    at edu.cmu.sphinx.decoder.search.AlignerSearchManager.recognize(AlignerSearchManager.java:348)

    at edu.cmu.sphinx.decoder.search.AlignerSearchManager.recognize(AlignerSearchManager.java:252)

    at edu.cmu.sphinx.decoder.Decoder.decode(Decoder.java:65)

    at edu.cmu.sphinx.recognizer.Recognizer.recognize(Recognizer.java:110)

    at edu.cmu.sphinx.recognizer.Recognizer.recognize(Recognizer.java:126)

    at edu.cmu.sphinx.demo.aligner.Aligner.start_align(Aligner.java:169)

    at edu.cmu.sphinx.demo.aligner.Aligner.align(Aligner.java:164)

    at edu.cmu.sphinx.demo.aligner.AlignerDemo.main(AlignerDemo.java:70)


    Any help?  I'm a day short of desperate.


    image
  • Joachim HollerJoachim Holler
    Hi David,

    try cutting your audio file into smaller parts if it's too long. 

    Also, encode with 16bit depth and 48kHz sample rate, that is usually best for performance.
  • Emilia MetzelEmilia Metzel
    Joachim, I just got this message today too: textsyncr There was an error executing the Aligner. Error details: null

    I've encoded the audio in 16bit and 48khz and got this message as well.


  • Emilia MetzelEmilia Metzel
    Joachim, I think I've found the solution.

    I've divided the audio into a few smaller chunks and it turned out that only one chunk was giving this error. I tried to find out what is different in this chunk, compared to the other ones... It was the word '3D', which the script apparently had a problem with. Once I changed 3 into 'three', the error didn't appear anymore.

    Maybe somebody could try this out too?
  • ruben cabreraruben cabrera
    Hi,

    After pressing "analyze", it took a while but then it gave me a pop up box with some errors...and then it shuts down the program.

    Thoughts??

    Need help ASAP!

    Ruben
  • Joachim HollerJoachim Holler
    Hi Mi La,

    yes, this may well have been the cause of the problem. You can compare the engine to a dictionary and some words are not in there. This is why it is always advisable to change the words causing trouble with similar sounding words that are more common, just like you did.

    (Ruben opened a support ticket and is getting help there)


  • Nick GardnerNick Gardner
    Same issue with the alignment error. Used every solution proposed in here to no avail. Professional voiceover no music. Creative Cloud 2015 error occurring on Mac and PC.
Add a Comment