VideoHelp Forum

Try DVDFab and download streaming video, copy, convert or make Blu-rays,DVDs! Download free trial !
+ Reply to Thread
Results 1 to 21 of 21
  1. Member
    Join Date
    Jun 2008
    United Kingdom
    Search Comp PM
    Dear all

    I have got about 200 MP4 holiday clips (totalling about 2 hours) from my Nokia 1020 phone. They are 1920 by 1080, 24 Mbps and 30 fps encoded in AVC.

    I simply wanted to put a movie title before the first clip and finishing title after the last clip without any editing at all. The movie will be watched on TV through a stand alone player which doesn't support soft subtitles.

    I tried Avidemux by making a 5 seconds MP4 clip (30 fps) from a photo and adding a 5 seconds AAC audio track.

    The resulting title clip worked fine but I couldn't get Avidemux to append it to the other clips and all I saw was the title clip and a green screen.

    On my intel core i7 920 pc running windows 7, it took about 7 minutes to re encode the title clip and a 30 seconds clip using MPEG4 AVC x264, constant rate factor single pass and the result was perfect.

    I tried many free video converting software and they all work same way through re encoding.

    Is there any way to handle these 200 clips without spending hours of re encoding?

    Many thanks.
    Quote Quote  
  2. Member
    Join Date
    Jun 2008
    United Kingdom
    Search Comp PM
    MediaInfo: clip_01.txt

    Format : MPEG-4
    Format profile : Base Media / Version 2
    Codec ID : mp42
    File size : 113 MiB
    Duration : 46s 829ms
    Overall bit rate : 20.2 Mbps
    Encoded date : UTC 2015-10-11 20:22:12
    Tagged date : UTC 2015-10-11 20:22:12
    Šxyz : +50.9005-1.4132/

    ID : 1
    Format : AVC
    Format/Info : Advanced Video Codec
    Format profile : Main@L4
    Format settings, CABAC : Yes
    Format settings, ReFrames : 1 frame
    Format settings, GOP : M=1, N=30
    Codec ID : avc1
    Codec ID/Info : Advanced Video Coding
    Duration : 46s 829ms
    Source duration : 46s 817ms
    Bit rate : 20.0 Mbps
    Width : 1 920 pixels
    Height : 1 080 pixels
    Display aspect ratio : 16:9
    Frame rate mode : Variable
    Frame rate : 30.032 fps
    Minimum frame rate : 29.910 fps
    Maximum frame rate : 30.151 fps
    Color space : YUV
    Chroma subsampling : 4:2:0
    Bit depth : 8 bits
    Scan type : Progressive
    Bits/(Pixel*Frame) : 0.321
    Stream size : 111 MiB (99%)
    Source stream size : 111 MiB (99%)
    Encoded date : UTC 2015-10-11 19:22:12
    Tagged date : UTC 2015-10-11 19:22:12

    ID : 2
    Format : AAC
    Format/Info : Advanced Audio Codec
    Format profile : LC
    Codec ID : 40
    Duration : 46s 741ms
    Bit rate mode : Constant
    Bit rate : 256 Kbps
    Channel(s) : 2 channels
    Channel positions : Front: L R
    Sampling rate : 48.0 KHz
    Compression mode : Lossy
    Stream size : 1.41 MiB (1%)
    Encoded date : UTC 2015-10-11 19:22:12
    Tagged date : UTC 2015-10-11 19:22:12
    Quote Quote  
  3. .

    If I understood right (?), may be — & thanks to PoisonDeathRay — you should check this: (step by step: post #19).

    Note ˇ not for lazy people, though (I AM lazy myself, but not in this case, since I've been waiting since... I don't know... ages!, for such a solution). And, in my case anyway, it just... works!

    Of course, don't delete any of your original material before fully testing the result, or a short test result, in your a stand alone equipment, or better: more than one device...
    Last edited by bulgom; 11th Nov 2015 at 23:02.
    Quote Quote  
  4. His case is much more tricky, because of the VFR

    Frame rate mode : Variable
    Frame rate : 30.032 fps
    Minimum frame rate : 29.910 fps
    Maximum frame rate : 30.151 fps

    For clips that are minimally variable, you can probably treat it as 30.0 . But some clips that have lots of deviation, they will have sync problems unless you adjust the timecodes . You'd have to extract timecodes, edit them , then remux with timecodes. The edit isn't trivial - for example if you appended a 5sec "intro", all the timecodes would have to be shifted by the same amount
    Quote Quote  
  5. Member
    Join Date
    Jun 2008
    United Kingdom
    Search Comp PM
    Dear bulgom and pdr

    Thanks for your comments. the details in post #19 were so complicated and unfathomable to me that I decided to spend more time trying to walk through this journey into the complexities of x264 encoding, so I have been reading about frames, gops and various bits and bobs.

    I tried MKVmerge to append the title clip, it didn't work.

    I couldn't find helpful info about sps_id like 2 so any explanation is most welcome.

    Perhaps one of you experienced guys would be able to advise me how to encode a 5 seconds dummy title clip which could be appended to my clips?

    Starting with ReFrames settings, the default is 1 according to Avidemux. I failed miserabley to reproduce this setting value in a sample clip encoded with mp4 avc x264 encoder. Does this setting affect appending of clips? [in my title clip ReFrames value was 3]

    Thanks to all.
    Quote Quote  
  6. Originally Posted by poisondeathray View Post
    His case is much more tricky, because of the VFR.
    Darn!.. :

    edit ˇ May be you could attach to a post a short (few seconds) excerpt, of one of the original clips. So I'd TRY to join a "dummy" to it (exactly what I did to mine): just to see if THAT, at least, works easily enough — & knowing:

    that it'd only be a start, since x seconds of a clip aren't comparable to the whole job, in the case of... VFR!

    Originally Posted by shaema View Post
    the details in post #19 were so complicated (...)

    you experienced guys
    No... there is... ONE, experienced guy, not two (guess which one...).

    Yes, "post #19" is somewhat like "room 237 or so, in Kubrick's The shining" (oops! autocomplimenting me... since I understood wrong.... again!)...

    I just edited that #19 post: replaced the attached "routine" + grouped most of the — video — processing commands anyway, into one. So it's less complicated (I didn't say simpler...).

    But, and that's what counts: it works. I guess that with PoisonDeathRay's help (let's don't talk about mine), you should be able to fix your problem, or find the least tedious solution

    knowing however that the "VFR situation" is or has been one of my nightmares... Against that, all I could ever do was to recode. Now, if you only observe slight FPS shifts, in your files (?), MAY BE you could avoid recoding: ?
    Last edited by bulgom; 13th Nov 2015 at 06:53.
    Quote Quote  
  7. Member
    Join Date
    Jun 2008
    United Kingdom
    Search Comp PM
    Hi bulgom

    Thanks for your response.

    See attachment in #9 please.
    Last edited by shaema; 14th Nov 2015 at 02:22.
    Quote Quote  
  8. .
    "Well"... no, to me anyway, it looks like it didn't upload right (?).

    I downloaded "ten_sec.mp4"; but it weighs 189 KB only, and it's black... May be I did something wrong?

    Easier: instead of "Google Drive", attach it to a post, here:

    click "Go advanced" button, then the paper clip icon. Then, browse, and send the clip: click "Upload' button. Don't close the upload window before it finishes uploading (only when the yellow stripe turns off, download's complete). Then you can click "Close this window", and the "Submit reply" button, to send the post, including the attachment.

    Also, try to open the test I attached to this post. If it works for you (?) like it did for PoisonDeathRay, you should see a 1st part, labeled "Test title" (in red), then fade to black, then the original video sample: tree leaves..., not recoded. It's silent: I didn't mux any sound to it (some players don't like that: the old "U-Mplayer" forces me to click the play button: twice...).
    Image Attached Files
    Last edited by bulgom; 13th Nov 2015 at 22:37.
    Quote Quote  
  9. Member
    Join Date
    Jun 2008
    United Kingdom
    Search Comp PM
    Thanks bulgom.
    Image Attached Files
    Quote Quote  
  10. .
    please take note:

    1. appending a title to your sample worked for me (see attachment) — ALTHOUGH:

    2. I did the WORST (test) job you, or anybody could think of!

    3. Don't forget PoisonDeathRay's warning: the VFR (variable frame rate) mode of your videos might very "well" prevent correct appending — I mean: you COULD get more or less if not totally desynchronized audio!

    But, to begin with, I just wanted to check some "plain / raw" appending (joining). Nothing more.

    Uncompress the ".7z" attachment — to check "z_RESULT_joined-video_silent__.mp4", in the folder — & tell me if your player(s) reads it OK; or garbled images? or... no play at all, may be?

    If you want to try it yourself:
    knowing how fond you are of "post #19" , the attachement includes the tools.

    Now, when I said WORST job, I really meant it, as I don't think I ever made such a little effort — in fact, not the least effort: I didn't even modify the batch a bit!

    ALL I did was: resize my "dummy" (= title / never mind its LOUSY quality: lightweight file wanted above all, here) to 1920 × 1080 pixels, to match YOUR resolution; and rename your sample to "_ORIG.mp4" ('cause too lazy to modify the file name in the batch )

    Then, I just clicked that batch (.bat file), and let it work. And that's... it...

    [ Note: couldn't test in my Win. 7 PC (busy recoding) / only under old 32-bit Win. XP. ]
    Image Attached Files
    Last edited by bulgom; 14th Nov 2015 at 10:49.
    Quote Quote  
  11. Member
    Join Date
    Jun 2008
    United Kingdom
    Search Comp PM
    Hi bulgom

    Something is shaping up!

    Your first , 640 by 480 clip ( z_RESULT_joined-video_silent__.mp4) worked ok on vlc and wmp. I felt there may be hope!

    And your next clip WORKED!! is it the magic of sps_id?

    Correct me if I am wrong:

    1 - you prepared TITLE-to-append.avi using virtualdub
    2 - you run x264.exe with variety of obscure options, giving it the avi as an input file [ I started reading, you know! ]
    3 - x264 produced T.264, a raw byte stream
    4 - you fed my original clip _ORIG.mp4 to MP4Box with -raw 1 option to extract the video stream
    5 - you did a binary copy to concatinate these two video streams into one
    6 - finally you fed the resulting video stream to MP4Box with an -add option to output the final MP4.

    Now I have to figure a way of making my equivalent of TITLE-to-append.avi. Then the audio!

    Again thanks to you and pdr for helping me in this fascinationg world of x264 encoding
    Quote Quote  
  12. ˇ
    Yes, the magic of "SPS-id"! I sure could hardly believe it works that easily!

    "1. to 6.": right: nothing to correct. Some remarks on the batch: below.

    In your place, I would first try to append about 10 clips, without adding a title for now, to check for possible audio desynch.!

    Then, I'd try to append the 200 clips (unless you don't want one only, out of the 200), still without adding anything: same audio check.

    If the appended result (or the 200 clips join) shows perfect, or at least good audio synchronisation, I would, then only, add a title clip — unless you want more titles?

    About "1." and "how to prepare a title clip": I had it already: 640 × 480 px., from my previous test. So I just blew up its resolution, using "VirtualDub" "Resize" filter, to 1920 × 1080, not even caring about proportion change (=> squished, in that case / for a simple test).

    You could use any other tool, of course: "AviDemux" (free), "AviUtl ExtraPack version" (free and very advanced, on titling), "Vegas" (commercial)... To "create" the initial title clip, I forgot if I used "AviUtl ExtraPack" or "Vegas", to — very simply — overlay some text onto an example video I had.

    I often prepare a "Photoshop" text on transparent background & save it to its ".psd" proprietary format: "Vegas" accepts it, dropped to a new video track

    — while "AviUtl" accepts ".png" files, with transparent background (and, to overlay, requires dropping them BELOW the main video track, instead of above = "Vegas").
    If needed, you could check this (post #11) attachment on "AviUtl" overlay: Very basic / simple (be warned) LITTLE step-by-step, to overlay a tansparent background logo., or any text.
    Then save that to "Title_something.avi", using "HuffYUV" codec, since it's lossless. Or even uncompressed, if not too long a clip.

    "2." That's when PoisonDeathRay suggested this:
    x264\x264 --crf 18 --profile baseline --level 3.0 --ref 2 --keyint 30 --bframes 0 --sar 1:1 --force-cfr --fps 30000/1001 -o T.264 TITLE-to-append.avi
    + explained that a method works in 100% of cases: "SPS-id + binary join". So, I just added the "--sps-id 2" option to the line:
    x264\x264 --crf 18 --profile baseline --level 3.0 --ref 2 --keyint 30 --bframes 0 --sar 1:1 --force-cfr --fps 30000/1001 --sps-id 2 -o T.264 TITLE-to-append.avi
    in which SOME options are still more or less obscure to me also, i.e. not all. You'll find documentation "googling" etc., + here. Anyway, an important one, "CRF (constant rate factor)" is one (re)coding mode among others, set here to 18 = good quality (but that I often set between 20 or 23 and 25, if not 25 to 28, in some cases: small resolution and/or no need to preserve quality).

    About "30000/1001", I guess that (like with aspect ratio commands) the encoder doesn't accept decimal value inputs, so, in order to get 29.976 images per second, 30000/1001 = 29,970... (I didn't test 30006/1001, that equals 29.976)

    "3. to 5.": right. And "6.": as you see, I used "add" without adding anything — here, it would have been an AAC audio file —, just to output to ".mp4" container.

    Usually, I rather feed a UI, with the two streams (video + audio): "Yamb/MP4box"... And I'm still wondering which one UI will give me less audio desynch — but, in fact, I might (?) have observed: "depending on the case!"...

    So, "AviDemux" SEEMS to works OK, quite often, but more in the "recode-everything" case — although I often use it on "copy", on video, and thorough recode on audio stream: resample to 48000 Hz + reset to stereo + aac (set to something between 160 & 384 kbs) conversion.

    Note ˇ in general or when possible, I avoid joining clips containing their individual audio track. I prefer reworking the whole joined audio file aside (with old "Cool Edit Pro", very handy, or "Audacity" if more than 2-channel audio), as it lets me correct at least easy (= constant) desynch. shifts;

    of course in uncompressed ".wav" mode, since lossless — that I'll feed to "AviDemux" AAC recode (= to avoid recoding audio more than once).
    Last edited by bulgom; 15th Nov 2015 at 16:48.
    Quote Quote  
  13. Member
    Join Date
    Jun 2008
    United Kingdom
    Search Comp PM
    Hi bulgom

    Thanks for your advice.

    I have managed to create my new_title.avi [15 sec long]

    Using Avidemux, I have successfully appended 10 clips into 10_clips.mp4 [9 minutes.39 sec long]. The audio was ok and no desync problems.

    With your .bat file, both x264 and mp4box managed to produce the silent_title_and 10_clips.mp4 which played fine on vlc and wmp. BUT it was [11 minutes 53 seconds]! I don't know how to explain this, I expected it to be 9.39+0.15=9.55 seconds. Woud it interfere with final video/audio muxing?

    With ffmpeg I managed to extract the audio stream from new_title.avi [15 seconds, as expected] as aac file

    With mp4box I managed to extract the audio stream from 10_clips.mp4 [9.39 seconds, as expected] as aac file

    If I am doing OK then next is to figure a way of concatenating those two audio streams.
    Quote Quote  
  14. You're not batch appending the title to each clip, only once - correct ?

    The sample you uploaded in post #9 is actually CFR, or very close to. But that doesn't necessarily indicate if other clips might be VFR or the degree of severity.

    If there are large deviations, your final result will be out of sync for certain unless you adjust the timecodes taking into account the appended intro

    The settings used in your camera are slightly different than the ones used in his camera sample. When you use the --sps-id method, many of the settings don't matter, but some actually do. It uses 1 reference frame instead of 2, and there is no fixed_frame_rate flag - you wouldn't have to "force" CFR in your encoded intro segment (it sort of makes sense since your camera uses VFR). It uses main profile at Level 4, and CABAC instead of CAVLC . I would try to emulate the settings as close as possible

    First, did you even try the "normal" method of appending? I would try that first with matching settings. The normal method copies timecodes over, so you wouldn't have adjust timecodes. I would only use that clunky sps-id method, if that failed, especially since your camera appears to use VFR
    Quote Quote  
  15. Member
    Join Date
    Jun 2008
    United Kingdom
    Search Comp PM
    Dear pdr

    Thanks for your comments

    As suggested by bulgom [I would first try to append about 10 clips, without adding a title for now, to check for possible audio desynch.!] from post #12, I simply opened the first clip in Avidemux, then using file>append, I appended the second, then the third until the tenth clip, then video copy, audio copy and saved in mp4v2 format. The result mp4 clip worked fine with no audio issues. The avi title clip was used once in the batch file.

    All of these 200 clips were consecutive mp4 files produced by my phone camera, shooting at 30 fps in 1920/1080 resolution. According to MediaInfo, all of them were [variable frame rate], but I did see frequent flactuations from 15 fps up to 30 fps.

    Could you expand on the "normal" method of appending files please? I thought I was doing the right thing with Avidemux appending function.

    Thanks again for your expert input.
    Quote Quote  
  16. The "normal" method would be trying to make an intro with similar characteristics, then appending that to your joined 10 or 200 clip video, by whatever method such as avidemux, mp4box, etc... ie. The "normal" method of not jumping through hoops , using commandline, sps-id, binary join etc... Sometimes you luck out and it works. To maximize the chances of that method, I would try to match the settings more closely.

    If you see fluctuations reported by mediainfo as low as 15 in some clips, chances are you will go out of sync if the timecodes aren't adjusted - but it's worth a try treating it as CFR

    A good predictor would be when you demux your big 10 or 200 clip joined video to elementary video and audio, then remux without timecodes, is that in sync or not? If that's already out of sync, then joining the intro will be out of sync as well, unless the timecodes are adjusted
    Quote Quote  
  17. It worked ok on your sample using "normal" method , appending with ffmpeg concat, at least for your 10sec sample. The timecodes are preserved so if you have large fluctuations in the frame rate in other sections, it should still be in sync. As mentioned in the other threads, the --sps-id method works, but is a last resort and not user friendly to use, also you'd have to adjust timecodes for a VFR scenario - which is no fun to do at all.

    I'll try to post instructions. I might have missed something, or assumed you knew something. If it's not clear, just ask.

    1) make your video intro using similar settings (notice no sps-id settins are used, only the reference frames, max keyframe interval, sar, b-frames, level and profile to match - these are the ones specific for your case; other cameras might need other settings)
    x264 --crf 18 --profile main --level 4 --ref 1 --keyint 30 --bframes 0 --sar 1:1 -o intro.264 input.ext
    2) make your audio with the same length . I used avisynth and audiodub() with blankclip so it makes the audio the same length as the video . I pipe the avs using ffmpeg to qaac, because qaac has a --no-delay option, which is better than somehting like nero which typically has a 40ms delay. As mentioned in the other thread, if you had lots of nero encoded segments, there is a larger possiblity of sync issues

    eg. using blankclip to generate "blank" audio. (Obviously there are other methods to generate blank audio a certain length, you can use an audio editor for example)
    aud=blankclip(vid, audio_rate=48000, stereo=true)
    And you don't have to use "blank" audio, it can be audio that is appropriate for whatever your intro was. The only setting to match is the CBR bitrate of your source (256kb/s CBR LC-AAC)

    eg. piping to qaac.
    ffmpeg32 -i "input.avs" -acodec pcm_f32le -f wav - | "qaac" --cbr 256 --no-delay -o intro_qaac_cbr256.m4a -
    3) mux the video & audio using mp4box to intro.mp4 . I assume you know how to do this or you can use a GUI like yamb beta 2 or mymp4boxgui

    4) join intro.mp4 to the other video(s) using ffmpeg concatentate

    a text file "mylist.txt" contains the names and paths , in order (change "PATH" to your real path)
    file 'PATH\intro.mp4'
    file 'PATH\ten_sec.mp4'
    ffmpeg -f concat -i mylist.txt -c copy ffmpeg_concat.mp4
    Image Attached Files
    Last edited by poisondeathray; 15th Nov 2015 at 14:07.
    Quote Quote  
  18. .
    To join two or more .aac audio files, I just tested OK:

    copy /b 001.aac + 002.aac + 003.aac Result.aac
    (copy that line, paste it to a text file, replace "001.aac" etc. with YOUR files names, save it to "Join.txt" and rename it to "Join.bat")

    About the 11'53" duration instead of 9'55", I would first let the whole thing play, timing it (using Windows clock or else), to check how long it ACTUALLY lasts. I'd also watch the video , to check for anything missing?

    I said... "first", since it's not because everything plays OK + nothing's missing or displaying wrong... that I'd garantee no problems (later)! Here, PoisonDeathRay knows — while I don't.

    About the normal or most usual method — which we always try in the 1st place of course —, I assume, and remind:

    (according to your Nov. 11th initial post) you had problems joining a title to an original video, either (previously) joined or not, yet.

    In my similar — i.e. not identical... — case, whatever I tried, no way! But one thing's sure: I'm not advanced, when it comes to encode a new video, matching very precisely an H.264 main video tech. specs!

    [ "A lot of time" has passed, since Nov. 11th... MAY BE you could TRY to recode the title video, matching as much as you can the 1st original clip specifications — & see if it accepts the usual append method. I know it does NOT work with mine; but with yours: ? ]

    Edit: while I was typing this post, PoisonDeathRay just posted the needed command line, to match your original video specs. That's the way to go — & to learn (note ˇ with "x264", I don't see any other way than concrete help from someone advanced enough; i.e.: actual & working example).

    For instance, adjusting time codes is new to me. I just don't know how (but haven't forgotten PoisonDeathRay's remark: it can be VERY tedious...). So far = two days ago, all I was able to do was: extract timecodes, using "mp4fpsmod.exe" and

    @echo on
    mp4fpsmod -p TC.txt _ORIG.mp4
    About the result — here: "TC.txt" —, I'll have to search: how to modify the lines without goofing (too much...), AND how to "inject" the result to a video — IF that's the correct routine?..
    Sorry but, as you see, I cannot help, on that point. I attached "mp4fpsmod.exe (+ the batch) to this post, though — since I had difficulties finding it (I copied it from "Hybrid" vid. converter folder).

    About audio (appendix) ˇ As I explained, I MY specific case, while I avoid recoding the video stream (but I'm still & often forced to recode!), I fuss somewhat less on the audio stream. After extracting it + converting to ".wav" (many tools work; I use "Pazera Free Audio extractor", among others) I just join ".wavs" onto some audio editor UI (see my previous post) — as it allows:
    — fine tuning of the internal audio volume(s) (+ music mix or other additional sounds) ;
    — and SIMPLE + "wisiwig" correction of desynch. shifts (simple... if constant!);
    — fast end processing, since audio recodes quickly.
    Now, I understand you may want to avoid that. In that case, to begin with — i.e. as a TEST —, I'd simply join the audio streams to one long track, using the bin. copy batch (see top of this post).

    About the WAY you join the original clips to one:

    since I didn't experiment with comparison in mind, I don't know (either...) if using "AviDemux" — and that's also how I quick append most of the time — might produce a different result than binay joining, technically.

    Therefore, I would try both — as long... as a rather quick method or tool helps me list the 200 (!) clips files names:

    under Win. 7 64-bit, (the very handy "TextGrabber" doesn't work, but) "Directory Lister 0.9" * does output a list of all files contained in a folder (1st modify default settings to avoid all other mentions than files names).

    Then, after copy + paste to "EditPad" — or even "Word": may be easier, I'd just replace all paragraph marks ( ^p ) with <space>+<space> in order to produce with the least effort, the required list: clip-001 + clip-002 + clip-003 etc. (or whatever names **).

    * Freeware version: (or )

    If your files names are long and/or contain spaces, you'd probably need to use quotes, so that the binary "DOS" command would (still) WORK... But: simpler = better! Just create a copy of the 200 clips and batch rename the whole stack to SIMPLE & SHORT names (such as Clip-001.mp4 etc.),

    using "Renamer" (for instance & freeware / faster: I use "ACDSee", set to Tools > Options > User interface > File list > check "Other file types"; then Ctrl+A on the File list, right click > Rename, ### as template, start from: 1).

    edit: One thing I've been wondering: why don't you recode the whole thing? since it's only 2 hours long...

    Of course, I'm VERY interested by any or all no-recode methods: 1. because I (badly) need to LEARN 2. & because I'm often asked to help with 10, 20, up to 100 hours of video (from camcorders & from cellphones, + "how to mix" both), and even more! It may sound crazy, but some people film like crazy, precisely...

    Needless to say it eats up considerable storage space and, before it's recoded (= sometimes a delay of several months, or years!), some of my friends want to title their appended video as is.

    But 2 hours, I would recode...
    Image Attached Files
    Last edited by bulgom; 15th Nov 2015 at 22:11.
    Quote Quote  
  19. Member
    Join Date
    Jun 2008
    United Kingdom
    Search Comp PM
    Dear bulgom

    Thanks very much for your detailed input, it would take me a while to read and digest this new info.

    Dear pdr

    You have again proved yourself to be a real giant in video editing. I raise my hat to you!!

    1 - I run your new batch file without sps_id, and successfully produced intro.264

    2 - I extracted the audio from title clip using:

    [ffmpeg -i new_title.avi -vn -ab 256 new_title.aac]

    3 - I merged intro.264 with new_title.aac using

    [mp4box\mp4box -add intro.264:fps=30 -add new_title.aac#audio -new res.mp4]

    4 - I appended the res.mp4 to my ten_clips.mp4 using

    ffmpeg -i res.mp4 -c copy -bsf:v h264_mp4toannexb t0.ts
    ffmpeg -i ten_clips.mp4 -c copy -bsf:v h264_mp4toannexb t1.ts
    ffmpeg -i "concat:t0.ts|t1.ts" -c copy -bsf:a aac_adtstoasc ten_final.mp4


    I now have a ten_final.mp4 with its introductory title and no audio sync issues.

    I managed to do the same with the next few clips and so far no problems.

    I have learned so much in this adventure, but this is life: we keep learning

    Thanks to all.
    Quote Quote  
  20. Originally Posted by shaema View Post
    —— let's dance a little (but good & fat) boogie-woogie! something HAPPENED! ——
    With the efficient + excellent help of PoisonDeathRay, titling a more or less reluctant * AVC H.264 clip or a bunch of joined clips — with no recoding except the title — becomes more accessible to... say, "intermediate ** newbies"! Using: a Windows based PC.

    So I've tried (again...) to simplify the manipulations; see attachment. Tested under Win. XP & 7.

    * This occurs (!) often "enough", not with cellphones clips only: at 1st sight, no way to append a title or end credits clips without recoding everything...

    ** Why "intermediate" (newbies / & that includes me)? Because the most important part might still be difficult to set. I.e., I'm not advanced enough *** to analyse & therefore apply in all cases the (precise & very) right "x264" encoding parameters — required, for the title clip to accept appending...

    *** Btw: I of course tried to apply the method explained here, to "my" stubbornely discouraging until few days ago, weird camcorder video:

    it works!
    So, the no less interesting "--sps-id" method is not even needed — in that (difficult) case either.

    Result: tested OK in Win. players, not in cellphones / standalone players, neither with "ffms + l-smash", yet.

    Now, since I'm often asked: "How can I title my STACK of clips, quickly + with no quality loss? ", I prepared the attached "kit" + example;

    it's a try, a start, an attempt, a draft, whatever one may call it. It's no "UI", nor anything definitive. And again: intended to help (as much as... I can, & knowing my ""level""...): newbies.

    Please note: comes with no... warranty ! / any improvement suggestion welcome.

    It uses free tools: "x264 + FFMpeg + MP4box + qaac": included ("kit"). Knowing "qaac" requires "Apple Application Support" installation, that I'd rather avoid, I made it portable / tested OK under Win. XP + 7. [ Setup file included, though. ]

    And "qaac" because, as explained earlier by PoisonDeathRay, the resulting .aac files aren't "padded", knowing that "padding" might desynchronize audio from video.
    Expected: create a title clip of same image size in pixels and framerate as the one(s) it will be appended to, in .avi container, with uncompressed .wav audio. Use preferably: "HuffYUV" or other lossless codec (not too exotic...) / though the 4 seconds included examples are "x264" .avis (to keep them light enough).


    Edit June 9th, 2019

    I know it's an ooold post: 2015!, but the new 2019 attachment belongs to it. = New version, Windows 10 (64-bit) compatible – on request – of the "video_Append-title_No-recode-to-orig" : "Win-10" "toolkit".

    Image Attached Files
    Last edited by bulgom; 8th Jun 2019 at 18:04. Reason: New 2019 attachment (update).
    Quote Quote  
  21. Member
    Join Date
    Jun 2008
    United Kingdom
    Search Comp PM
    Dear bulgom

    Thanks for posting these excellent tools, which I would certainly come back to them many times in future.

    In particular, I liked these detailed batch files, which bring to me memories of the lovely old MS DOS!

    I think it is remarkable achievment that you have prepared a batch file, which, with some modification, could do the whole job of producing working videos without re encoding. Thanks for spending time doing it.

    Long live MS DOS!!!
    Quote Quote  

Similar Threads