Search:

But why? This is poor (full artifacts) machine learning crap...

VLC may work for you then... ffmpeg has ffserver but it is not officially available for Windows - however you may try to search for Windows ffserver - if i recall there are some unofficial builds in Internet.

https://stackoverflow.com/questions/2679569/how-to-do-simple-video-streaming-with-ffserver-on-windows

https://trac.ffmpeg.org/wiki/ffserver

Check Avalpa software - currently i can't find details but i recall they offered free software able to properly broadcast so perhaps it will work for you.
I was able: https://github.com/aventuri/opencaster so perhaps this will suit you.

On my Win 7 ffmpeg build(s) located on site (search for "Automated nonfree builds of FFmpeg for Windows using GitHub Actions " for WIP and for stable ffmpeg search for "Automated stable nonfree builds of FFmpeg for Windows using GitHub Actions") works fine.

It should be lossless same as FLAC or WAV - i would try to convert files anyway to FLAC as it is more popular and less problematic than WMA.

provide report - add "-report" to command line - log will be created so error can be analyzed.

I like Dune universum - this is great book about time perspective but my point was rather about Villeneuve movie going opposite to book - this is like feminist bullshit - manifesto of fight with white male patriarchy - two different stories - currently very inaccurate but still best Dune movie is David Lynch movie - also there is completely different league of actors there - long story short - plenty to write and this is not place for it.

Btw i know lot of people prefer instead of reading books listening - there are audiobooks - not my favorite way but still better than movie going against novel.

ffmpeg mapping x264 parameters automatically.


SET x264opts="crf=%vq%:vbv_maxrate=9999:vbv_bufsize=9999:level=4.0:keyint=%key%:ref=7:bframes=4:cabac=1:tff=0:interlaced=0:no_psnr=1:no_ssim=1:open-gop=0:pic_struct=1:aud=1:nal_hrd=vbr:force_cfr=1:bluray_compat=1:fullrange=off:overscan=show:colorprim=smpte170m:transfer=smpte170m:colormatrix=smpte170m:stitchable=0:sliced_threads=0"

@ffmpeg -hide_banner -v 32 -stats -y -i %filename% -vf %vproc% -c:v libx264 -x264opts %x264opts% -x264-params %x264opts% -preset medium -tune film -profile:v high -level:v 4.0 -af %aproc% -c:a libmp3lame -q:a %aq% -f mp4 -movflags faststart "%~n1_.mp4"

Get a book - this movie is Dune abomination.

delay OTHER channels then - also it should be possible to perform negative delay (i.e. advance audio) but it will be performed in similar fashion i.e. by introducing delay on other channels.

This can be useful too https://ffmpeg.org/ffmpeg-filters.html#setpts_002c-asetpts

Use different - straight (i hope this is on Videohelp forum politically correct term) plug or use sockets providing sufficient distance.

Use https://ffmpeg.org/ffmpeg-filters.html#adelay

If your goal is to transcode on the fly (i.e. some movie library to render device such as TV) then you can set highest possible bitrate - If your intention is to store (long term?) transcoded movies then i would consider using CPU transcode to get highest possible quality.

To be honest i can't tell - for sure NVEnc is pure HW solution where Intel offer also hybrid mode (HW+CPU) - in theory this should lead to better quality. Recent ffmpeg introduced if i recall correctly Ultra HQ setting so it may be important to use latest possible drivers and software tools.

Based on https://en.wikipedia.org/wiki/Intel_Quick_Sync_Video?useskin=vector seem CoffeLake using newer gen than SkyLake so there is chance that it will offer higher quality at the same bitrate.

You can try to use rigaya https://github.com/rigaya tools to squeeze absolute maximum possible from both HW encoders.
Also recent Intel GPU's (ARC series) may offer better quality (so instead replacing CPU's perhaps it could be easier to simply use Intel ARC graphic card - they offer also AV1 encoding - not so common).

My personal impression is that newer hardware offer usually higher quality of video encoding - this is natural - more complex HW is able to deliver better quality.
So if your goal is quality then or you use newest HW encoder possible (with moderate bitrate increase when compared to CPU driven encoder) or you are accepting higher bitrate and to deliver same quality on older HW encoder.
This is always tradeoff between quality and bitrate.

Spectral subtraction and your ear are enough to remove most of the problems from your recording but proper methodology is key to success.

Key points are:

- use relatively small FFT (somewhere around 1024..4096 points), higher FFT can lead to more spectral leaks and as such various chirps
- choose proper sample for noise profile acquisition - it should be free from complex tonal signals - search for relatively short time frame where only noise is - present, no other sound especially things like vocals or music is present
- perform processing in chunks as in time noise profile may change in time so one acquired at the beginning of the recording may be different than noise profile in middle or at the end
- it is useful to search for optimal threshold level to select inverted processing i.e. removing signal while keeping noise - search for such threshold where noise is unmodulated by useful signal
- try to perform relatively mild noise removal strength - not go over 20dB - you can always repeat process multiple time to increase noise removal efficiency.

http://forum.doom9.org/showthread.php?t=174620

Also HC Enc may work for you.

btw -dc 5 is not allowed - 8 bit is minimum and 10 maximum - higher = better

there was quite recently similar thread - can be interesting for you https://forum.videohelp.com/threads/410677-ffmpeg-non-linear-quantizer?highlight=174620

Nowadays thermal runout can be easily verified with FLIR camera - i bet you can borrow such thing nearby or even try to buy something inexpensive.

If this is thermal problem then freezer in spray will help to detect issue...

Well... just use nonfree ffmpeg version with libfdk compiled in ( search for "Automated nonfree builds of FFmpeg for Windows" ). AC3 is definitely way more popular and widely supported by almost all consumer decoders. I have impression that with 320kbps even generic ffmpeg aac encoder should be OK.

Yes, check this:
https://trac.ffmpeg.org/wiki/AudioChannelManipulation#a2monostereo

Bellow example how it may look in ffmpeg - not tested as i have no similar source to yours - ffmpeg can be tricky in such things but definitely it is doable.


ffmpeg -i "D:\input.mxf" -filter_complex "[0:a:0][0:a:1]join=inputs=2:channel_layout=stereo,asplit=2[audio_1][audio_2];[0:v]bwdif=mode=1[video]" -map "[video]" -c:v h264_nvenc -r 50 -cq:v 18 -aspect 16:9 -pix_fmt yuv420p -map "[audio_1]" -c:a:1 libfdk_aac -b:a:1 320k -map "[audio_2]" -c:a:2 ac3 -b:a:2 640k "D:\output.mp4"

You need to use -map functionality
for example: https://stackoverflow.com/questions/62122503/can-ffmpeg-transcode-an-audio-track-and-add-it-as-a-second-audio-track-at-the-sa

I would check camera schematics to see if sensors require some bias current - described symptoms looks like some drift in time/temperature (my assumption is that you performed thermal grease replacement correctly).

Search for dilation - morphological operator - as johnmeyer suggested perhaps simple kernel operand will be sufficient.

You can use ffmpeg for this.

btw it is not clear what are you mean by applying DRC - are you interested about signal dynamics processing to get similar loudness or to copy DRC metadata embedded in AC-3?
Also it may be important to get more details - for example if AC3 source is multichannel so downmixing must be performed to for example stereo or there is no downmixing required.

I overlooked information about Grayscale() behavior in case of the YUV - it works in such case similarly to ShowY() - my fault, you was OK. Sorry.

Sure, sure - you are simply jealous and obviously being looser - ever tried to eat 3 lunches and 4 dinners? In between have a gym, jogging and try get some yachting or golf - this is hard work and not many people can do such things for years or even whole life with smile on face...

[/sarcasm]

Thanks! This worked perfectly! I increased the brightness to 25, and that looked best.
[/QUOTE]

Conversion from YCbCr to Y may introduce issues - it is more safe to extract Y channel from source and later convert it to RGB to apply proposed by hello_hello colorization. My goal would be to keep Y unaltered but if you are happy by modifying it then fine.

It was many years ago since i've used Avisynth so i can only suggest how i would perform processing:

- split your input source on 3 planes, Y, U, V, ( https://avisynthplus.readthedocs.io/en/latest/avisynthdoc/corefilters/extract.html ?)
- apply separately to U and V plane Limiter so signal values will be clamped to defined range ( https://avisynthplus.readthedocs.io/en/latest/avisynthdoc/corefilters/limiter.html )
- combine modified U and V planes with unaltered Y plane ( https://avisynthplus.readthedocs.io/en/latest/avisynthdoc/corefilters/merge.html ?)

This is probably fastest method.

Bellow example shows YUV statistics, as input i've used one of your pictures.


ffprobe -f lavfi movie=sz15pq.png,signalstats -show_entries frame_tags=lavfi.signalstats.YMAX,lavfi.signalstats.YHIGH,lavfi.signalstats.YAVG,lavfi.signalstats.YLOW,lavfi.signalstats.YMIN,lavfi.signalstats.UMAX,lavfi.signalstats.UHIGH,lavfi.signalstats.UAVG,lavfi.signalstats.ULOW,lavfi.signalstats.UMIN,lavfi.signalstats.VMAX,lavfi.signalstats.VHIGH,lavfi.signalstats.VAVG,lavfi.signalstats.VLOW,lavfi.signalstats.VMIN


and stats looks like this:


TAG:lavfi.signalstats.YMIN=18

Probably easiest way will be read Cb and Cr values of tint (should be constant) from red tint source and later set fixed value for whole Cb and Cr correspondingly to red tint source values - my assumption is - source is black&white (grayscale) so only Y channel carry information, Cb and Cr normally for grayscale video should be equal to Cb=128, Cr=128 - in case of red tint Cb and Cr should have constant value different than 128.

This is very wide question - depends on your source, sometimes removing PID 0x1FFF may reduce significantly size of your stream.

It is obvious - unknown format and codec used to perform capture - recently more and more newly opened threads imply use of some lossy, interframe video codec - in such case to get linear access to every video frame is to export whole sequence as series of pictures - also this depends on editor possibilities - for NLE it is probably easier to navigate and export and later substitute single pictures but for linear editors with limited functionality (mostly limited to split, cut, combine) whole series of the separate pictures may be easier to control.

Perhaps some machine learning (AI) based solution - something for automatic speaker recognition and separation .

For working solution search for Logitech programmable remotes - they are expensive but at least fully programmable also you can download pre-configured setup for your device (brand) over USB. Perhaps something listed here suit your expectations:
https://www.tomsguide.com/us/best-universal-remotes,review-4464.html

You can export captured video as images, edit few of them and later import whole sequence.

Works for me - i suggest check your Chrome configuration - perhaps popups are disabled by default or there is some other protection active (maybe AI driven to release you from pain of talking decisions and protect you from your own actions - yes, this was intentionally a bit sarcastic)

FFV1 is lossless but conversion from YUV 4:4:4 10b to YUV 4:2:0 8b is lossy - colors can be the same or similar (within 8 bit accuracy) - depends on your YUV 4:4:4 10b source - if your source is 8 bit converted to 10 bit and after back to 8 bit then there is chance that in overall conversion will be lossless.

Console or you can create batch and use it from console and/or GUI.

GIF can't be lossless, also it is highly questionable if 256 colors can provide "visually lossless" quality - of course you can create more than 256 colors (using GIF syntax or for example performing temporal dithering) but this is highly dependent on your GIF viewer capabilities.

Quality of GIF depends mostly on color quantization quality - error diffusion can help but color selection is most important - there is not many good color quantization algorithms - my favorite are NeuQuant and SColorQ - also color space can be important - technically you need algorithm matched with human vision - this is similar area to automatic subjective image/video quality measurement - all methods are still very simplistic when compared to human vision.

https://bisqwit.iki.fi/jutut/colorquant/
https://www.imagemagick.org/Usage/quantize/

This may be also interesting for you https://bisqwit.iki.fi/story/howto/dither/jy/ - more about dithering topic .

And short reply - IMHO GIF is not even close to be visually lossless.

Well, there is example with looping source:


ffmpeg -loop 1 -framerate 2 -i input.png -i audio.m4a -c:v libx264 -preset medium -tune stillimage -crf 18 -c:a copy -shortest -pix_fmt yuv420p output.mkv

so to modify it for looping 3 seconds into 10 hours you need to modify it in such way:


ffmpeg -loop 12000 -i inputfile -c:v libx264 -preset ultrafast -crf 18 -c:a aac -b:a 160k outputfile.mkv