Hi there,

Not sure if i have the right place for this question but I have several video files each roughly the same file size and bit rate. They all play well with vlc/quicktime however with ffplay the audio dropouts every consistently throughout playback.

This is mediainfo report on one of the files . Is there an incompatible codec I am working with? I also transcoded the file to a h.264 file with ffmpeg and it played well in ffplay.

Any advice? Thanks in advance

General
Complete name                            : /Users/EQ2317/Desktop/harry.mov
Format                                   : MPEG-4
Format profile                           : QuickTime
Codec ID                                 : qt   2005.03 (qt  )
File size                                : 26.4 GiB
Duration                                 : 17mn 1s
Overall bit rate                         : 222 Mbps
Encoded date                             : UTC 2016-07-21 16:14:58
Tagged date                              : UTC 2016-07-21 16:26:17
Writing library                          : Apple QuickTime
com.apple.proapps.logNote                : harry_new
com.apple.proapps.reel                   : 001
com.apple.proapps.scene                  : Scene 01
Media/UUID                               : 8CE6BF11-1ABF-4D4B-978D-9D841DD8F66F

Video
ID                                       : 1
Format                                   : YUV
Codec ID                                 : v210
Codec ID/Hint                            : AJA Video Systems Xena
Duration                                 : 17mn 1s
Bit rate mode                            : Constant
Bit rate                                 : 221 Mbps
Width                                    : 720 pixels
Height                                   : 480 pixels
Display aspect ratio                     : 4:3
Frame rate mode                          : Constant
Frame rate                               : 29.970 (30000/1001) fps
Standard                                 : NTSC
Color space                              : YUV
Chroma subsampling                       : 4:2:2
Bit depth                                : 10 bits
Scan type                                : Interlaced
Scan type, store method                  : Interleaved fields
Scan order                               : Bottom Field First
Compression mode                         : Lossless
Bits/(Pixel*Frame)                       : 21.333
Stream size                              : 26.3 GiB (99%)
Language                                 : English
Encoded date                             : UTC 2016-07-21 16:14:58
Tagged date                              : UTC 2016-07-21 16:26:17
Color primaries                          : BT.601 NTSC
Transfer characteristics                 : BT.709
Matrix coefficients                      : BT.601

Audio
ID                                       : 3
Format                                   : PCM
Format settings, Endianness              : Little
Format settings, Sign                    : Signed
Codec ID                                 : sowt
Duration                                 : 17mn 1s
Bit rate mode                            : Constant
Bit rate                                 : 1 536 Kbps
Channel(s)                               : 2 channels
Channel positions                        : Front: L R
Sampling rate                            : 48.0 KHz
Bit depth                                : 16 bits
Stream size                              : 187 MiB (1%)
Language                                 : English
Encoded date                             : UTC 2016-03-21 16:26:01
Tagged date                              : UTC 2016-03-21 16:26:17

Other
ID                                       : 2
Type                                     : Time code
Format                                   : QuickTime TC
Duration                                 : 17mn 1s
Time code of first frame                 : 04:44:44;04
Time code, striped                       : Yes
Title                                    : time
Language                                 : English
Encoded date                             : UTC 2016-03-21 16:26:01
Tagged date                              : UTC 2016-03-21 16:26:17

Transcode into a deinterlaced Apple ProRes LT

• What about a button just with transcoding into ProRes and
  an other (= this) button with additional options, e.g.
  de-interlacing?
  (c/o Reto)

Got the thumbs up from @edsu to add this script in. I'll do this when I have time but making an issue in case someone finds time first! https://gist.github.com/edsu/36beb5576896013179356bd8ed95c34c

FFmpeg dropped some knowledge into my timeline when I was complaining about the weather and I need to drop that knowledge into this! https://twitter.com/FFmpeg/status/695689188205731846

the @privatezero tutorial last week got me thinking about deinterlacing. the only yadif flag in ffmprovisr is in the "Transcode to Prores" command.

Wondering if there should be something separate for clarity, and wondering if anyone has any exp with deinterlacing filters other than yadif (I count 4 others: bwdif, kerndeint, nnedi, w3fdif).

Alternatively, does anyone have a method for controlled deinterlacing (or denoising) testing?

thanks.

think it would be beneficial to add ffmpeg installation instructions (or possibly, a link to @retokromer 's excellent FFmpeg cookbook, https://avpres.net/FFmpeg/).

I really like the "Testing Your Installation" section; I find it super helpful.

Hi all, as archives may wish to create preservation clips in a lossless fashion from digital masters, maybe instructions on lossless segmentation and and concatenation would be beneficial?

As there is already in commands for clipping and concatenate (Join files together/Trim a video) without re-encoding, maybe they could be added to the preservation section as well with an emphases on the stream copy function:
-c copy
use stream copy mode to re-mux instead of re-encode

thanks and what a great resource.

regards,
Andrew

@kfrn has agreed to do this. Making an issue so we know it's a work in progress.

Here's a start. I think it would need to be in two passes, one to get the scene change times and one to split, so first pass with signalstats and 2nd with the segment muxer.

This command outputs a csv of the time of scene change and the YDIF value. I'm using 10 below as a scene change threshold, adjust as needed.

ffprobe -f lavfi movie=HEY.mkv,signalstats -show_entries frame=pkt_pts_time:frame_tags=lavfi.signalstats.YDIF -of compact=p=0:nk=1:s=',' | awk -F, '$2 > 10'

The csv could then be processed to make it an input to the segment_times option of the segment muxer, see: http://ffmpeg.org/ffmpeg-formats.html#segment_002c-stream_005fsegment_002c-ssegment

Move README to gh-pages branch, make gh-pages branch the default view, and then destroy the master branch

As the number of recipes increases, I think it could be useful to be able to link directly to a recipe.
I don't know how to implement this, but if there's any repetitive retrospective data entry required to make those links, I'll volunteer to do it, once I know roughly what to do.

Thanks!

I am thinking about SD to HD using
colormatrix=bt601:bt709
and HD to SD using
colormatrix=bt709:bt601
@kieranjol @ablwr Any thoughts?

Is it possible to add a changelog or a 'most recent recipes' page to the site? Someone on AMIA-L recently asked for something somewhat similar, though I think he wanted updates posted on the listserv instead. It could be useful for visitors who may not notice that a new recipe has been added, and it could also be a useful promotional tool as we could link to the new scripts on various public forums.

I'm not sure if linking to the commit history would be the best way to go about it as it could confuse people not familar with github.

At a some point, I guess, we should also consider an example of interlaced ProRes.

there's info on the ffmpeg doc pages (http://ffmpeg.org/ffmpeg-filters.html#readeia608), but think could use the ffmprovisr treatment (clear examples, description, etc.)

also wondering about hex-to-text possibilities. anyone have any ideas?

thanks!

From @kellyhaydon (BAVC):

«Is it possible to divise an ffmpeg script that will open and play two files in a directory at the same time, and with the temporal difference filter? It would be beneficial to add to ffmprovisr.»

See: bavc/qctools#221

Pull specs from video file

• I would indicate that json and flat are other output
  formats.
  (c/o Reto)

Will be provided "tomorrow".

@kieranjol Maybe you could add here as recipe the information you posted on FFmpeg-user? The general command with the examples ffmpeg -h encoder=ffv1 and ffmpeg -h muxer=matroska would be very useful, in my opinion. (If you don’t have the time, just let me know, and I’ll do.)

Hello @ablwr! I have the impression that the "aria-labelledby" parameter is not used. If so, I can delete it. Please let me know. Thank you!

Hi all, I have a batch script for Windows (PowerShell) that I could add as a separate command.
Maybe we have some Windows users that would like to automate tasks. What do you all think?

We should find something nice that is improved by the use of the -report parameter. Suggestions welcome!

Sort the code blocks into categories with subheadings for better organization of information. This ticket entails thought towards how to best categorize these scripts and also implementation of that categorization. Ideas for categories are something like... access copy creation, format transformation, exporting, image extraction, snippet extraction, digital preservation methods.

Right now there's a JS dependency called by a CDN. It'd be good if this was integrated into the app so it could be downloaded and run even without an internet connection. Also the cute fonts drop out. ;)

Additionally it would be good to reduce the overall weight of the application with consideration to slow connections.

I'm currently working on scripts that automate the addition of head and tail plates to a .mov, while adding a transparent png watermark that kicks in after 3.5 seconds.
ffmpeg -safe 0 -f concat -i mylist.txt -i "/Users/bla/playertests/logo.png" -c:v prores -filter_complex "overlay=:enable='between(t,3.5,2000)':main_w-overlay_w-5:5,yadif" outputrigh12345.mov
I'll add something like this over the coming days.

Please see the conversation in this thread: #139

Boost up documentation on contribution guides and ways people can contribute / make it as easy as possible for contributions to happen.

I'd like that!?????

We need MOORRRRRRREEEEE!!!

Ref tips from @kieranjol and @retokromer: #36

Create frame md5s

• We should also indicate in an additional button how to check if the MD5 are correct or not.
  (c/o Reto)

What about a little different structure?
One can chose between

• different input formats,
• different actions to apply, and
• different output formats.

Then the app generates the corresponding window with the full command and the related explanations.
(As it's three-dimensional, that would be a kind Rubik's Cube for audio-visual archivists.)

Use Case: with exported thumbnails from video, want to make 'preview' GIF that's still low-res enough to be used as a thumbnail supplement.

Possible starting point: Create animated gif from a set of jpeg images stackoverflow question

We link to CRALS and Sourcecaster in the README but I think we should link to it on the site itself, underneath the link to the CLI learning resource. I can do this but not at this time, so filing an Issue for the middle period (or if someone else wants to do it).

Heya,

Has anyone successfully concat files using this method? https://amiaopensource.github.io/ffmprovisr/#join_files

The language is a little unclear and I have not had success.

Attached is how I interpreted how to create the join list, but I keep getting the also attached error.

If there are any other possibly more simple methods for successfully joining DV files I'd be in love.

BAVC1004511_2385M.txt

QCTools examples just leave the big ol lump of signalstats to be mysterious and undeciphered, would be good to break it down better for understanding

Inside the tag the zeroes look like zeroes, but in the

Note so self: Write a recipe using an image sequence and the -start_number 086400 parameter.

Saw someone asking about this on the internet so a mental reminder to add script that extracts audio without loss from a video file: ffmpeg -i video.[mp4] -vn -acodec copy audio-only-file.[aac]

In Firefox, when viewing https://amiaopensource.github.io/ffmprovisr/ (note the HTTPS), I get a mixed content error that stops the individual items from expanding when clicked on.

Reduce the overall weight of the application with consideration to slow connections. This can be done by cutting out lazy dependencies on Bootstrap and jQuery and using only what is necessary.

Adapted from existing Excerpt to end

Excerpt from end

ffmpeg -i input_file -sseof -5 -c copy output_file

This command copies a video file starting from a specified time before the end of the file, removing everything before from the output. This can be used to create an excerpt, or extract content from the end of a video file.

ffmpeg
starts the command
-i input_file
path, name and extension of the input file
-sseof -5
Tells ffmpeg what timecode in the file to look for to start copying, and specifies the number of seconds from the end of the video that ffmpeg should start copying. The end of the file has index 0 and the minus sign is needed to reference earlier portions. To be more specific, you can use timecode such as 00:00:05.
-c copy
use stream copy mode to re-mux instead of re-encode
output_file
path, name and extension of the output file

Will be provided "tomorrow".

Continued from here, if that's OK? #36

Im not sure how ye see longer scripts within ffmprovisr.. Go with bash or also include other languages that also allows Window support? Or do ye want to keep the site to single ffmpeg commands?

It could be handy to knock up a simple watermark script, where the font size is calculated in proportion to video height?
Or possibly check what the codec and wrapper is of each file in a directory, and transcode if certain conditions are met?

Here's a python one that (last time I checked) works in windows and OSX - haven't tried Ubuntu yet but it should work.
https://github.com/kieranjol/pyh264bitc/blob/master/py_h264_bitc.py
I need to change a good few things in that script, mostly based on things I've learned from ffmprovisr! Also there's some things in there that are specific to where I work which should be removed.
The main difference in terms of OS adjustments is getting python to escape the colons in the timecode/fontpath in different ways depending on which OS it detects.

There seem to be some overlaping/broken commands in this section.

https://amiaopensource.github.io/ffmprovisr/#thumbnails
This command is called "generate thumbnails" in the plural, but actually seems closer to the script for generating only one. The explanation within this command also refers only to making a thumbnail "at a specific time" so there is discrepancy between the title and explanation.

Additionally, in its current form with an extra input tied to a variable, the script is broken. Is this a typo, or was it trying to do something else that accidentally didn't make it into the final script? (Since the annotations don't mention anything else I suspect the first explanation).

Since there already are commands for creating both single and multiple thumbnails, is this command redundant?

Is it worth showing how to just decode a file in order to check fixity against internal crcs in an mkv? I'm not sure if this is too obvious/simple?I'd like to get feedback on this before I create an entry.
Something like ffmpeg -i input_file -f null -

I would find this useful within a preservation context. I didn't know how to harness fixity info within a container/codec until @dericed showed me on twitter. Perhaps there's other benefits to this beyond fixity?

If it is worthy of inclusion, should it be called "verify embedded checksums" or should it just be called something general like "decode a file"?
There's also the err_detect bitstream and crccheck tools but I don't see any benefit to these over just decoding to -f null - , though I'm probably using them incorrectly.

Hey, as discovered at the Tate workshops, the change display aspect ratio recipe (that i wrote)is misleading,as it also transcodes to prores.As stream copy doesn't work with a filter, specifying a codec is required .
@dericed wrote a better one in his 'change stream properties' script,which uses field_order,but from my limited tests, this doesn't work with Matroska,only mov(probably mp4 too?).

I think the setfield script should stay in but a better explanation is required so that the re-encode is made more obvious.

In the WAV to .mp3 script ffmprovisr gives instructions for a constant bit rate which will result in access files that are significantly larger than if a constant quality setting was used.
http://amiaopensource.github.io/ffmprovisr/#wav_to_mp3

FFMpeg's documentation suggests that anything set between -qscale:a 0 and -qscale:a 3 should in most cases be indistinguishable from files encoded with the max constant bitrate in terms of sound quality:
https://trac.ffmpeg.org/wiki/Encode/MP3.

Is this worth revising, or making a note about constant quality vs bit rate?

I guess we should add some comments on http://amiaopensource.github.io/ffmprovisr/#batch_processing explaining how to customise the path. (This is a recurring question.)

Someone just asked if they could use this template for another project and I realized we don't have a license! I'm happy with the most open CC license. Can I get a +1?