Is it possible to transcode .MOV with animation codec and Alpha Chanel to 24 bit PNG sequence using Google Cloud Transcoder?
The documentation leads me to believe this is impossible.
https://cloud.google.com/transcoder/docs/concepts/supported-input-and-output-formats
If it is possible, does anyone have a thought on how to make it work?
Related
I have a video of several cv::Mat RGB frames that I need to store as videos with lossless encoding on Jetson Xavier NX. In this process, I need to convert the RGB frames to YUV444 and then use Jetson's inbuilt lossless encoding tools to encode to h264/h265 and then save in an mp4 file format.
I can read cv::Mat frame and then convert it to the YUV444 pixel format (either manually element-wise or using OpenCV's inbuilt converter). Currently, I am writing the YUV444 frames to a binary file per frame using std::ofstream(filename, std::ios::binary). However, when I encode the yuv file to h264 or h265, I get very bad results. Is there something I am missing?
Is there a better method to save a video into a yuv file (such that I can use the encoder on the entire video).
-->> Original image in yuv444 space
-->> Image in the encoded video
Here is the link (https://imgplay.zendesk.com/hc/en-us/articles/360029411991-What-is-GIF-Dithering-Option-) where it says When you save the file as GIF with dithering, it can make your GIF more natural.
How to implement Dithering for creating more natural GIF from UIImages or video frames using Objective-C or Swift?
Assuming your source image is 8-bit per channel RGB, you could use vImage. vImage doesn't have a vImageConvert_RGB88toIndexed8, but you can split your interleaved image into three 8-bit planar buffers for RGB. I don't know exactly how well this would work, but you could convert two of the three channels to Indexed2 with vImageConvert_Planar8toIndexed2 and the other channel to Indexed4 with vImageConvert_Planar8toIndexed4. That would give you the required 8-bit lookup table.
Apple have loads of Accelerate sample code projects here. Optimising Image-Processing Performance discusses converting interleaved images to planar format. If you have a known palette, Applying Color Transforms to Images with a Multidimensional Lookup Table may be a solution to quantising your image to 256 colors.
For a project I'm currently working on, I'm trying to convert a bunch of PNG images to HEIF/HEIC. These images will be used in Xcode's .xcassets, which will then be "compiled" into a .car file.
Compiling the PNGs (~150 total files) results in ~40 MB of Assets.car, which is why I'm trying to convert them to HEIF/HEIC in the first place. I've tried various solutions, such as ImageMagick, "Export as" in GIMP, biodranik/HEIF, libheif's heif-enc, exporting a PNG as 8-bit or 16-bit in Photoshop and doing everything all over again. But everything results in the .heic file being "broken" on iOS. The first image shows the best output I've got so far, but still fringes around the edges. The white rounded rectangle on the right is iOS' Face ID padlock.
The second image is (I think) a 16-bit PNG converted to HEIC using libheif#1.8.0, upgraded through Homebrew. Lossless quality preset, 10-bit output. heif-enc complained about the color space being converted from RGB to YCbCr, stating even though you specified lossless compression, there will be differences because of the color conversion
Is there any way to properly convert PNG files to HEIF/HEIC without such quality loss? Please don't suggest online services to convert files, as I'd like to keep total control of my files.
Note: To get lossless encoding, you need this set of options. Try :-
-L switch encoder to lossless mode
-p chroma=444 switch off color subsampling
--matrix_coefficients=0 encode in RGB color-space
I'm working with lots of camera's which capture in BG bayer pattern natively.
Now, every time I record some data, I save it to the disk in the raw bayer pattern, in an avi container. The problem is, that this really adds up after a while. After one year of research, I have close to 4TB of data...
So I'm looking for a lossless codec to compress this data. I know I could use libx264 (with --qp 0), or huffYUV, dirac or jpeg2000, but they all assume you have RGB or YUV data. It's easy enough to convert the bayered data to RGB, and then compress it, but it kind of defeats the purpose of compression if you first triple the data. This would also mean that the demoasicing artefacts introduced by debayering would also be in my source data, which is also not too great. It would be nice to have a codec that can work on the bayered data directly.
Even more nice would be that the solution would involve a codec that is already supported by gstreamer (or ffmpeg), since that's what I am already using.
A rather late suggestion, maybe useful for others..
It helps to deinterleave the Bayer pattern into four quadrants and then treat that image as grayscale. The sub-images (e.g. all red pixels in top left) have half the spatial resolution, but their pixels are more highly correlated. This leads to lower residuals from predictors using nearby pixels and therefore to better compression ratios.
I've seen this reach 2-3x lossless compression on 12-bit raw camera data.
If a commercial solution is ok, check out Cineform. I've used their sdk for a custom video compressor and it works great plus they have some great tools for processing the raw video.
Or if you prefer the open source route check out Elphel JP4.
All I know about Bayer Patterns I learned from Wikipedia, but isn't conversion to RGB more of a deinterlacing than a tripling? Doesn't the resolution for red and blue go down by a factor of 4 and green by a factor of 2? If so, a lossless image compression scheme like lossless jpeg might be just the thing.
I recently wrote C programs for image processing of BMP images, I had to read the pixel values and process them, it was very simple I followed the BMP header contents and I could get most of the information of the BMP image.
Now the challenge is to process videos (Frame by Frame), how can I do it? How will I be able to read the headers of continuous streams of image frames in a video clip? Or else, is it like, for example, the mpeg format will also have universal header, upon reading which I can get the information about the entire video and after the header, all the data are only pixels.
I hope I could convey.
Has anyone got experience with processing videos?
Any books or links to tutorials will be very helpful.
A video stream, like MPEG, is composed by a number of frames dependent from (obviously) its duration and its frame-rate. To read a pixel you must start from what is called an Intra frame, which is not dependent from a previous frame into the stream. Any successive frame is a frame which is temporally dependent from its previous frame, so to obtain its pixel, you have to decode the stream from the Intra to the frame you want.
Note that, tipically, an intra frame is inserted periodically to give to the decoder a way to synchronize with the stream. This is very useful in a context where errors can occur.
What you want to do isn't an easy work. You have to use an MPEG decoder and then modify the frame before diplaying it, if you want to do post processing, like a filter or other.
I suggest you to study video coding, and you can find a lot of material on that, starting from the standard MPEG.
I would recommend looking into FFMpeg. It has a command line utility that can grab frames from a movie and dump them to an image like JPG. You can then modify your existing reader to handle JPG files (just use something like libjpeg to decode the JPG to a raw pixel buffer).
Alternatively, you can use the FFMpeg APIs (C, Python, other), and do the frame capture programatically and look at the pixels as you move through the video. Video formats are complex, so unless you want to start understanding all of the different codecs, you might want to grab a library to do the decode to raw pixel buffer for you.
MPEG 1/2/4 videos are much more difficult to handle than bitmaps because they are compressed. With the bitmap data, you have actual color values stored directly to the file. In MPEG, or JPEG for that matter, the color information goes through a number transformations before being written to the file. These include
RGB -> YUV 420P (sub-sampled chrominance)
Discrete Cosine Transform
Weighted quantization
zig-zag ordering
differential encoding
variable-length encoding (Huffman-like)
All of this means that there is no simple way to parse pixel data out of the file. You either have to study every minute detail of the standard and write your own decoder, or use some video decoding library like ffmpeg to do the work for you. ffmpeg can convert your video to still images (see answers this recent question). Also, you could interface directly to the ffmpeg libraries (libavformat and libavcodec). See the answers to this question for good tutorials.