Weird combing in progressive source

[embis2003]

Here you go:


https://drive.google.com/file/d/1b5ywDny6I2pVcd4Y0ZAXe4giOCLCUaV8/view?usp=sharing

That's exactly the same video, just in a TS container rather than MP4.

Oof yikes sorry about wasting your bandwidth DB83.

[embis2003]

It's not the same kind of what I said.Your TS file is as the same as the file ofembis2003's link.
It's easy to get this 4200 or 5200K vevo AVC TS.
The method to get higher bitrate file I showed is still a secret....

Well, it must be a BIG SECRET because I am part of many communities and forums who have only been able to rip < 5200kbps from VEVO. But if you ever find it out, I would love to know.

[DB83]

Here you go:


https://drive.google.com/file/d/1b5ywDny6I2pVcd4Y0ZAXe4giOCLCUaV8/view?usp=sharing

That's exactly the same video, just in a TS container rather than MP4.

How can it be the same ?


The OPs vid was 199mb. Thus one is 217mb. Don't tell me that the container takes 18mb. The bitrate is bigger. Go check.

[poisondeathray]

Here you go:


https://drive.google.com/file/d/1b5ywDny6I2pVcd4Y0ZAXe4giOCLCUaV8/view?usp=sharing

That's exactly the same video, just in a TS container rather than MP4.

How can it be the same ?


The OPs vid was 199mb. Thus one is 217mb. Don't tell me that the container takes 18mb. The bitrate is bigger. Go check.

Demux them both. Compare the size of the elementary video stream only

[hintlou]

Here you go:


https://drive.google.com/file/d/1b5ywDny6I2pVcd4Y0ZAXe4giOCLCUaV8/view?usp=sharing

That's exactly the same video, just in a TS container rather than MP4.

How can it be the same ?


The OPs vid was 199mb. Thus one is 217mb. Don't tell me that the container takes 18mb. The bitrate is bigger. Go check.

emm....It's the truth that the TS container took 18mb.You can check.

ffmpeg -i  yours.TS -c copy yours.mp4

your 216mb TS will become a 199mb mp4.
I have downloaded the 216mb TS when embis2003 showed the m3u8 link yesterday。。。earlier than you...

I think,links
"http://22-cf-168198817351468082649c6ba82cee92.vip1-dal1.dlvr1.net/ed3c085a-8a07-11ea-ab7b-bb28a3b566a9/v3/hls/2015/10/USUV70502343/f92ea866-1b8e-45f8-adff-109f01624580/5200/usuv70502343_5200k_1920x1080_h264_5200_aac_128.m3u 8"
or "http://22-fa-168198817351468082649c6ba82cee92.vip2-ams1.dlvr1.net/35593fc2-89ef-11ea-a...0_aac_128.m3u8"
is not the right link for a higher bitrate file.
Maybe we can change words in the links to get a big file.

[DB83]

Here you go:


https://drive.google.com/file/d/1b5ywDny6I2pVcd4Y0ZAXe4giOCLCUaV8/view?usp=sharing

That's exactly the same video, just in a TS container rather than MP4.

How can it be the same ?


The OPs vid was 199mb. Thus one is 217mb. Don't tell me that the container takes 18mb. The bitrate is bigger. Go check.

Demux them both. Compare the size of the elementary video stream only

Oh bugger.


I re-checked the video bitrate for both and ... (well you know the next bit)


It was the overall bitrate that threw me when the audio was identical.

[jagabo]

Here you go:


https://drive.google.com/file/d/1b5ywDny6I2pVcd4Y0ZAXe4giOCLCUaV8/view?usp=sharing

That's exactly the same video, just in a TS container rather than MP4.

How can it be the same ?


The OPs vid was 199mb. Thus one is 217mb. Don't tell me that the container takes 18mb. The bitrate is bigger. Go check.

I did check. The video streams are identical. Transport streams have lots of overhead because they are designed to be picked up any where along the stream. They contain many headers that describe the video throughout the stream. (Think TV broadcast where you might start watching a channel at any time -- the TV needs to know the resolution, frame rate, etc. before it can start playing. So all that information is broadcast every few seconds.) Containers like MP4 and MKV usually have only one header (at the beginning or end) that describe the entire video.

[DB83]

I have learnt the hard way.


But what I can not get my head around, which is what led me to offer this version, is why should a manifest (if that is the right word) offer mp4,mpg and ts yet different download programs only pick up one type. Or is there only one type and the download program wraps what it thinks it sees even altering the bitrate for the mpeg to suit.

[embis2003]

Some intresting updates regarding this.

I have purchased the real ProRes master of this video, that was retrieved straight from the label. Presumably transferred straight from the betacam tape, it is in 4:3, 480p and the original interlaced structure is intact.

Posessing this allowed me to reverse engineer the process which created the artifacts in the VEVO transcode.

IVTCing the master with tfm(pp=0).tdecimate() creates nearly identical artifacts in the same exact spots as in the VEVO rip. Leading me to believe, the artifacts I was actually seeing was simply unprocessed comb frames in the IVTC process.

On the master, resizing to 720 x 540, then cropping all the letterboxing off leaves me with a 1:85 video. Further cropping an additional 16 pixels off both sides leaves me with a 16:9 video just as the VEVO rip was. I then upscaled to 1920 x 1080. Comparing this to the VEVO rip, it is basically the same.

With this knowledge, upon reimporting the original vevo rip, I de-lanczos resized back down to the same cropped SD size. I added 16 black pixels to the left and right side to pad out the difference. I then added 77 pixels on the top and 75 pixels on the bottom to add back the letterboxing, leaving me with 750 x 540. Then, I de-bicubic resized to 720 x 480.

I applied tdecimate(cycle=25) to get back to 24. I then did ChangeFPS to get back to 29.97, then applied tfm().tdecimate() again, without pp=0 And yep, you probably could've guessed, it worked perfectly. The fields were 100% preserved. The compression did distort them to some extent, however, using clip2 with QTGMC it deinterlaced the unprocessed frames very well.

So theoretically, if you have a video like this, it is possible to de-interlace the frames if you can find out the original, interlaced size. For videos which don't have weird cropping, it would probably be a lot easier to guess the original size.

But I just thought I'd leave this here in case anyone find it useful when working with VEVO videos.