What's the difference between the different versions of a movie when I torrent it...

Please do.

Alright, do you have a preference for where you want it uploaded? Cause if not I'll just use a google drive. I keep a few clips handy to use as tests for encoding stuff, and one I like to use in particular is from Captain America - Civil War because the bluray was on clearance at walmart and the way they mastered it was great. I'll upload the bluray source that I demuxed and split (no transcoding), and my encode for comparison, although I'll be leaving in a bit so I may have to send you some other way to contact me if the upload isn't finished by the time I have to leave.

It's uploading now, it'll be about 15 minutes. My upload speed is slow lole

I would be ok with just a few selected still image comparisons, user.

Maybe he is troll

stop posting that faggot
also pirates torrents contains viruses
and movies and tv shows are for man children.

Oh, right... well it's almost done anyway.

14 year olds actually believe this shit lmao

Here you go:
drive.google.com/open?id=1s0oYL0YP-TY_YBjv0925jO_CkNGncL7X

You will of course notice the sensor noise is gone, but I don't consider noise to be meaningful detail, as long as it isn't useful in hiding banding.

this, legally obtained malware, spyware and bloatware are much better! Dont steal, but be stolen from.

ROFL

>a file not much bigger than YIFY sizes look nearly the same as the blu-ray source
literally impossible

Check out the bitrate on my upload if you don't believe me. It's about 1600kbps, but the entire movie is about 2400kbps.That's about 2.6 GB for a 147 minute movie, whereas YIFY usually makes his uploads about ~2GB

pretty good for 1/20th the size but it's night and day from the source. noise might not be "information" but sometimes fabrics or skin texture isn't too far from that resolution.
old films shot on actual film also lose a lot when the grain is filtered out.
i understand internet isn't great for everyone but storage is dirt cheap nowadays. i normally download 8-20GB 1080p recodes if nothing else is available. i would be disappointed with your encode quality if i downloaded a film worth watching more than once.

How can the bitrate be 1600 when its 2400 quality? Also is it an HEVC file?

>slow movement and still frames
There is nothing impressive about it. Now do the same on a scene with lots of changes between the frames with the same bitrate constraint (1600-2000Kbps) for a 3:11 clip and watch it fall apart.

Oh, well if that's your opinion after downloading and watching it, then I suppose that's fair. I just wanted to show what I could do with low bitrates, and that not everything that's small in size is as terrible as YIFY. I agree about storage nowadays, and with a slightly CRF I could probably reach a good compromise for you.

The bitrate of this clip is 1600kbps, but the entire movie is 2400. the bitrate is variable over time (Variable Bitrate). It's not HEVC, just good old x264.

You're right of course, this is an ideal sample, but if you want an action heavy, worst case scenario, the movie has a lot of it, and as I said it's 2400kbps for the entire movie.

Sure it looks acceptable for that size, but I don't call that "look nearly the same as the blu-ray source". Also if you're going for file size savings then use HEVC. AVC's strength are high bitrates.

10-bit AVC

Attached: comparison_01.png (922x719, 489K)

Personally I find 10-bit AVC is actually pretty damn close to HEVC, 8 or 10-bit, at the same bitrate. I've done a few bitrate normalized tests, and when you use something like the very slow preset they wind up looking nearly identical, and HEVC takes a ton more time to encode. For slightly lower bitrates it has better control of artifacting, but at that point it's not really watchable for me. The encode I uploaded is right above the edge where I stop nitpicking the details and enjoy the movie. I imagine that's a somewhat personal preference thing, and until now I assumed it was good enough, but I suppose in the future I'll try using CRF 19 or maybe even 18 instead.

A:10 V:10

THANKS YIFY!

You are welcome sirs.

Plz like and subscribe to our scene group x(0)x-RaheshRIPz for more amazing films.

Not really what I saw in my tests (both subjectively and with metrics like PSNR, SSIM and VMAF), but I get the speed argument.
As for the quality, use whatever is good enough for you. No reason to waste space on details you don't care about. Just be careful with those claims. You can tweak the encoder for optimal results, but no encoder can produce miracles.

Dude, I just dump my source video file into handbrake, use H.265 10-bit, decomb filter (I don't even know what that shit does, I don't understand the explanation), have everything on "same as source", the quality slider at CRF 22, and I end up with three to seven times less size than the original file with the same results as this shit you uploaded.

How the everyliving fuck do you get away with almost twenty times less size?!

Decomb is supposed to counter interlacing (like in pic related). If you don't have interlaced footage and the interlace detection is on, then it doesn't do anything.
If you really want to cut down the file size then denoise and use a slower preset than the default.

Attached: 1515459197687.jpg (250x189, 7K)

Well turn off the decomb filter if your source isn't interlaced, and probably the most important thing would be to turn the preset in the video tab to "Very Slow"

Ah, right, a lot of the content I recode is 1080i not 1080p.

>denoise
>slower preset

Got it, thank you!

Which denoise preset do you use? NLmeans ultralight? I really don't want to lose much detail.

>I cant enjoy the movie because of partial quality loss
The movie is shit, you know it.

why encode it, why not just dump the raw bluray disc as bytes and mount it later?

people that watch or care about movies are dumb af, that's probably the reason

grab files with x265 encoding for most efficiency

I use them via ffmpeg not Handbrake, so I'm not familiar with any presets. What I can say is
>prefer temporal to spatial denoising
>keep both to a minimum by doing a few test encodes
>NLMeans produces superior results but is also a lot slower than hqdn3d
>keep in mind that NLMeans was designed to denoise old VHS footage, so be careful about smoothing the video too much
Take a look at mattgadient.com/2013/06/29/in-depth-look-at-de-noising-in-handbrake-with-imagevideo-examples/ for more info.

Not everyone has a fast internet connection or access to unlimited storage. Disk space may be cheap, but Blu-Ray remuxes can still be considered a waste of space.

Thanks for the help.

OP here, I use an program called Hybrid, made by some german guy a couple years ago. It's not really beginner friendly and has a spartan GUI, but it has every feature you could ask for and a ton of built in Avisynth/Vapoursynth support. The amount of denoisers available with the click of a button might give you whiplash.

Er, not OP, I'm a fucking retard

>10kb
You are like a little child

Attached: 5c2c1167d894a.jpg (1024x576, 12K)

Wait. It clearly says 5kb on my phone

Attached: croppedImage_1546392086935.png (1282x1184, 65K)

Well, my coding time doubled, I hope the result will be better. I'll know in 5 hours, kek!

>5 hours
Wew lad, what CPU?

BLCK clocked [email protected] GHz

Get you an R7 2700 and OC it. An old duo core isn't gonna cut it for video encoding. The 2700 would be almost 3 times faster than that.

I'll make do with what I have, I'm a poorfag. Me coding movies for a smaller file size is really just another attempt at saving money by creating more free space so I don't have to buy a new hard drive.

>more than 10KB
>not even trying

Attached: out.png (1024x576, 8K)

Ah, in that case I'd advise you to just learn how to pick out good encodes. Generally speaking the sweet spot is around 3500kbps for 1080p, which translates to about 26MB per minute pf video. You'll also wanna look for 10-bit encodes, because although the source may not be 10-bit, and 10-bit sounds like a bigger number than 8-bit, because of the way quantization works you can eliminate banding better with 10-bit using the same bitrate compared to 8-bit. Also if you don't have surround sound avoid things with 5.1+ audio, or multiple audio tracks, and make sure it's not high-bitrate audio, as it would be completely wasted bandwidth. 128kbps AAC is perfectly acceptable for movie audio. x265/HEVC is also a pretty safe bet, as HEVC is good with banding at low bitrates, even at 8-bit. (Somewhat)

Thanks, I'll pick my sources more carefully whenever I can.

blurays have some ridiculous DRM and crap on them. Older bluray players can't even play new blurays because they changed the encryption keys. Also few people can or will seed files that huge. If you can even download it in a reasonable time to begin with.

It would be nice if pirates included some of the extra footage that comes with DVDs on their rips. I haven't seen the special features on a movie in a very long time.

5 hours really ain't anything.

Attached: Untitled.png (328x333, 9K)

Wintoddlers need not apply.

Well, mine crept up to 11 hours too, still not as bad as sixteen, kek!

>Handbrake
>encoding

if you view videos at a slow enough playback speed to notice details, you are wasting precious hours of your life

I didn't say either of those words. I think you need to re-read my post.

bitrate

Jesus christ why even bother encoding at that length of time. I was considering buying two 8-core xeon e5-2650 V2s just to speed up my encodes

Do you put out more than 500 releases a year?

How do I become a magic encoder human

All of my encodes are for myself at the moment, although I did have an idea for my own group. Not sure if I wanna take on that level of risk for no reward though.

The reason I wanted them is cause they're extremely cheap used ($60ish) and I was gonna make my own media server.
Lots of learning over a long period of time. If you want a crash course another user suggested Matt Gradient's site, and he has some good content.
mattgadient.com/category/encoding/

I'll see what I can find in the way of literature and get back to you on that.

Thank you! It's only the first day of 2019 but you've been more helpful than all of Jow Forums in 2018.

Bitrate, bits per pixel, codec used, resolution. These are your four biggest factors that determine file size and quality.

For my own rips I knocked the file size down from over 1GB per (40 min tv show) to around 360 - 400 mb. The results look damn good on my 65" HD set.

When doing rips you need to factor in the viewing device your gonna be watching them on. Two small a rip (low res, low quality) will look like shit on a large screen. Two large a size and your wasting drive space for no gain. You may as well just keep the source vobs and be done with it.

You'll hear a lot of people bash Xvid/H.264 codecs. Don't listen to them. Both work perfectly fine. Both are widely supported by devices. Both deliver nice output quality if used properly. Yes new H.265 may be better, uses less space,etc. But drives are cheap and as I said Xvid/H.264 still work perfect as is.

x264 is still great but XVID is severely dated and has fairly nasty blocking artifacts. Sure it looks OK with a high enough bitrate but that kinda defeats the point.

At this point x264/h264 is so widely supported and common with encodes that there's really no reason to keep using XVID.

aaaaaaaaaaaa, you're welcome user

I'm watching this video atm to comfirm that it's actually good, but I'll link it because it's 40 minutes long and I don't really want to make you wait that long. Finding good sources is harder than I thought it would be.

player.vimeo.com/video/104554788

Yea, that's why there are at least two people lurking the thread waiting for someone worth their salt i.e. you, finding one that is in your professional opinion good teaching material.

So it looks like the first video, the one I linked, is good for the basics and I recommend it, but his export settings video is based on Adobe Media Encoder which is quite frankly terrible, and although this guy knows what he's doing, you should avoid AME like the plague.

Handbrake is very easy to use for the begginer and has fairly decent documentation on it's site. If you get a bit more advanced I'd recomment using tools like MeGUI, Hybrid, FFMPEG, MKVToolNix, etc, but by then I trust you'll know enough that you can learn on your own. here's the handbrake documentation:

handbrake.fr/docs/en/1.0.0/table-of-contents.html

Thanks again!

is it worth getting a used GTX 960 so I can take advantage of hardware encoding on ffmpeg? Will it be any faster than encoding the a CPU?

NVENC is very fast but there is a slight efficiency loss. If you want the best possible compression efficiency then CPU encoding is the way to go, but if you don't want to wait a year then NVENC is absolutely fantastic, and much faster.

Also, I watched the second video and 90% of what he talked about applies to other encoders so I would actually recommend that video too:

vimeo.com/115534648

There is one thing I wanna add though, once you've watched through the video you'll notice he talked about CBR and VBR, but there's a third option called CRF, Constant Rate Factor, which is generally much better than either CBR or VBR as long as you don't need to hit a target bitrate, but it works much differently than CBR or VBR.

CRF works by using a psychovisual model to target a given quality level, and then assigns as much bitrate as needed to achieve that level of quality. This means the size of your encodes can vary significantly, but the quality level stays the same. It also has the advantage of not needing a second encoding pass in order to achieve optimal results. I highly recommend using CRF unless you have a reason not too. Find the lowest CRF that gives you great results and then use that. Caveat: CRF values differ slightly in the actual quality depending on the preset/settings you use, so technically CRF 20 for the Fast preset is not always gonna be the same as CRF 20 for Slow, although they will be fairly close.

Yea, I'm coding at crf 20 + very slow right now with 3:3:7:7 NLMeans, in another 5 hours I'll know how it turned out. fml

Until then I'll have plenty of time to go through all the great information you've provided. I really appreciate it!

Hmmm, i might give it a go then. I'm doing encodes on an old octocore xeon workstation and that shit takes all night (doing a crf of 18 and a preset of slower).

As a guy with slow FUP internet and limited storage I will download the smallest filesize. Bitrates be damned.

it sounds like he things h.264 and xvid are the same thing

fyi, 'xvid' is an encoder for MPEG-4 Part 2 video, and h.264 is MPEG-4 Part 10 video
they're different standards, h.264 is newer and more efficient than xvid, and unless you're working with a "divx" dvd player or other early 2000's hardware such as xbmc on an original xbox (which can play SD h.264 video, or 720p xvid video), there's no reason to use xvid anymore

I wish ~1Go sized torrents were better seeded than the bigger ones. The quality difference is marginal for me and I'd rather have faster downloads/less space taken on my HDD.

quality
that said, "lower quality" doesn't mean it will be noticed, and even if it is noticeable, it doesn't mean it will be annoying
take pic related as an example, what i do for tv shows is i download the highest quality ones i can get, then when i've seen them, i'll compress them further so i have an OK quality one to keep that doesn't take up much space
top: 3.3GiB, 1080p, H.264
bottom: 681MiB, 720p, H.264
(excuse the use of jpeg, png was too big)

Attached: a.jpg (1920x2160, 3M)

ahhh, well as that one user pointed out, with NLMeans you might wanna run some tests at like CRF12 Very Fast with different settings for NLMeans to see if you can tolerate what the denoiser is doing first. Very Fast will only take a few minutes at most, and with the CRF set to 12 it'll basically only be showing you what the denoiser is doing, it'll just make a very large file that you can safely delete when you're done testing.

And you're very welcome, I love talking about this subject and video in general but it's kindof a super niche and autistic thing so not many people are interested.

>Also few people can or will seed files that huge. If you can even download it in a reasonable time to begin with.
Not a problem on private trackers.

If this 10 hour coding will be a waste of my time I'll do 1 minute segments where I fuck around with a lot of settings one by one. It's certainly not something I should use an entire video to experiment with using an i3-6100, but I'm already 47% in.

Can someone explain to me why Netflix 4K doesn't take up 60gb per movie, but a UHD bluray remux does?

Is it framerate or something?

You can watch it as it's encoding if it's in mkv or raw .264. I don't think you can in .mp4 though.

Netflix is compressed to shit.

It's mp4, I don't think Handbrake does anything else.

Yes it does.

Handbrake supports mkv, I've used it several times. In the first tab under format it has a selection menu with mp4 and mkv.

Mine only has MP4 and M4V.

What version are you on? I'm on 1.1.2-and this is what my menu looks like.

Attached: HandBrake_2019-01-02_01-07-11.png (1018x680, 202K)

>grainy film
>some retard decides to compress it all to shit
>retards then download and watch this yify crap
It makes me sick. You need at least 2GiB/hour for 720p 264, and 4-6 for 1080p. Which is why a lot of the time 720p upscaled in video player has better quality than 1080p, because of the higher bitrate/pixel.

is u an appul friend
is u useing mac compuder sir?

Ah, there, I'm blind. 8:10 AM here, sleep deprivation is beginning to get the best of me.

Badly compressed grain is indeed terrible, but if you know how to deal with it you don't need even remotely as high bitrates as that. 12GB for a single 120 minute movie is outrageously overkill.

I can't blame you because you obviously only go by what you know of other people's work, but if you knew the actual capabilities of encoders, filters, denoisers, etc, I'm positive you would change your mind.

I have a short clip from the movie Labyrinth that had a ton of grain, I denoised it with MCDegrainSharp and compressed it down to 3400kbps, and it still looks pretty decent, although obviously denoising a heavily grainy source isn't going to give you perfectly pristine results I would certainly consider it medium to medium high quality.

if you're targeting a bitrate low enough to turn grain into artifacts, you need to denoise
denoising obviously gets in the way of an accurate encode, but if size is a bigger concern than accuracy, denoising will give you a better result

Short answer, lossy compression.

Here's a slightly longer answer. This image is 4K, look at the filesize. You can describe this image with very, very little information. "Rectangle 3840 pixels wide by 2160 pixels wide, all black" is enough. I don't need to write the exact value of each individual pixel, that description is enough to perfectly recreate the image on your screen if you want. Now suppose I add (truly) randomly placed dark grey dots on the screen, almost black. You wouldn't notice them at a glance. But all the sudden I need to convey a shit ton more of information for you to recreate said image perfectly. But, like I said, you can't notice them. So why don't I just approximate the image by that solid black rectangle and save a ton of fucking bits? Now, the degree to which I allow those "approximations", which is a lot fucking more complicated in reality, determines how lossy and worse looking (source information discarded and thus permanently lost on the output) my encode will be, but it'll also almost always make for much "easier to describe" frames and thus a much smaller file size, since it's less information to convey. Again, this is an extreme oversimplification, but it's basically what encodes are all about at it's core. You set a bar for fidelity and then try to get that file size as low as possible while staying above that bar, usually. Yifishit and such is the other way around, trying to make the best out of a given filesize, since turd worlder internet and impatient people who don't care about quality anyways.

Attached: Untitled.png (3840x2160, 9K)

Thanks.