For those of you that torrent video files this question is geared toward you. I’m looking for a sweet spot between quality, size & speed for HEVC encoding. I’m using FastFlix and seem to be getting really wide and varying speeds.
I’m not really literate on all this video lingo but I can, at least, get it going. Most files take anywhere from 5-17 mins for a 30-40 mins clip. I have a AMD Radeon RX 470 graphics card but when I try and use the VCEEnc it won’t let me use CRF which I’ve heard it the best way.
Anyway, if you’re willing to share knowledge or what settings you use when you convert video to HEVC that might help me speed up my processing, I would be eternally grateful.
In most cases, most release groups already fine tune encoding settings towards various balances of file size and quality, so the best option is to decide on a set of release group whose standards meet your needs and just use the files as they come without further modification.
Applying lossy compression to a video that’s already had lossy compression applied to it degrades it unnecessarily, so if you’re going to compress it yourself, it’s best to start with the remux, aka the original media file.
I’d personally recommend releases from members of the qxr group and Vyndros.
I understand about reencoding and I’ve gotten a lot of flak about that from a lot of people but as I mentioned above it’s more of a space issue with me also. I appreciate the mention of groups to look to as that helps much. Thanks for your input!
IDK I’d assume anything uploaded more than 10 years ago needs to be re-encoded (but you should learn more about the old and new encodings before generalizing that blindly).
I’ve also had success removing embedded language audio tracks from a file that had 5+ languages from the original Blu-ray. Each language was over 1GB/per movie for a specific offending collection.
I think that’s part of my problems I have new stuff and stuff waaayyy over 10 years old and I’m trying to find a “sweet spot” for everything and I don’t think that’s possible. Yeah, I know what you mean about the additional language, audio, etc. tracks especially in stuff I get from Nyaa. I tend to batch demux a lot of that out before I even start messing around with re-encoding the file. Thanks for your input!
You need to do 2 pass encoding. You should also not use CRF. You should pick a bitrate for the file size you want. Do a first pass which analyzes the video to see which sections require more data, and then run a second pass which will give high bitrate to more action scenes and lower bitrate to the credits and slower talking scenes.
Some action scenes require 5 times more data to look as good as a talking dinner scene, you couldn’t even notice the quality difference but the bitrate requirement is literally 5 times more.
You also need to use the slow preset and use x265 if you’re doing this to archive the stuff forever. Do it once and do it right.
Thank you, this is VERY helpfuly information! As I mentioned in another reply the ffmpeg front-end I use FastFlix only allows bitrate or CQ encoding when using the VCEEnc Hardware encoder for AMD so your information is noted and very appreciated!
2 pass encoding is only to get the benefits of variable bit rate when targeting a specific file size. If you don’t have a specific file size in mind, that’s what CRF is for.
No, the specific file size is irrelevant, he’s wanting smaller file sizes. CRF is a waste of data on more than 70% of scenes in hollywood movies. You set a bitrate and let it go. This is also why virtually all music now is VBR
Are you confusing CRF with CBR?
CRF is the video equivalent of VBR music. The music equivalent of two-pass video encoding is ABR music.
When tuned for a specific file, CRF and two-pass video will give similar results. They both result in a variable bitrate encoding.
When using the same config on different files, you might find that two-pass encoding produces unnecessarily large files for something with little movement like anime, or has quality issues for something with a lot of movement like a lot of shaky camera or film grain. Meanwhile the same CRF setting will work well in just about any scenario, using more bitrate for files that need it, and less bitrate for files that don’t.
Don’t encode lossy to lossy. The encoding will take forever and the image quality will suffer. If you want file size savings redownload your media in h265 or whatever. Or temporarily download a lossless copy, encode, and delete.
There are no lossless copies of any movie that have ever been released to the public
Delivery formats (h265, h264, h263 etc) are compressed and lossy.
Are you short on disk space? Personally I’d just buy enough storage that I don’t even need to care
Yeah, you hit the nail on the head. I only have two 2Tb drives and they are about full so I’m looking to downsize my files. Someone recommended buying a USB hub and several 16Tb Seagate drives to accomodate but I’m not sure if that’s the best option.
Different people have different needs.
If someone has a lot of time and not a lot of money re encoding video is a decent answer.
I’ve been there and done that before.
Replacing (or adding a 10TB USB to your ) single 2 tb drive isn’t a horrible idea. It’ll take you quite a while to go through that 10 tb. In the meantime you look toward getting an old case and some kind of modest motherboard and setting up an Unraid. It’s a journey, and unless you are made of cash you’re not going to get to your endpoint all in one jump.
Unraid is budget friendly because you can add whatever size disc you want to do it, It supports a parity drive so you have some support against failure. The only truly difficult part is that the parity drive must be as big as the largest drive in the box.
In the end only you can decide what works for you. If you want to re-encode your stuff, 2 pass is best. You are going to lose quality, that’s unavoidable, But if you’re watching it on a TV 12 ft away, You’re going to forget about any quality as soon as you get in grossed in anything you’re watching.
Thank you so much for your input. I’m so glad someone understands my point of view and doesn’t keep telling me DON’T REENCODE. I know that but we do what we have to when we’re not made of $$$. So, you’re saying in re-encoding to let it select it’s bitrate and use 2-pass and be done with it. That’s actually helpful because FastFlix will only allow CQP or Bitrate encoding when using the AMD hardware acceleration VCEEnc. Again, thanks for the advice!
It’s your media you do what you want with it.
Pick a busy movie with a bunch of stuff going on, and then pick a really dark movie.
Try different encodings with each one of those. You’re playing a game of time versus quality. And you keep in mind, the electricity for those encodes isn’t free either.
Try them with a fixed bit rate, try them with the two pass. If the fixed rate doesn’t look good try bumping the rate up. You’ll get a feel for it eventually.
Back when I was hard up for disc, I made everything 1080p HEVC single pass constant rate. I don’t even remember exactly what the bit rate was but I would just encode everything and then watch a sample out of it. If one of them turned out bad I would reincode it with better settings.
Dual pass will get you a little smaller and better output, But it takes forever, and you’re sitting there burning watts all night long.
In the end you just need to fiddle with it, and weigh the output versus your resources.
Thank you for your input and understanding my situation but I’ve been fiddling around so much I’ve haven’t gotten even one series or movie encoded. I’m about ready to just pirate a commercial product and see if that will do what I want it to but I probably still wouldn’t find a product that would do everything I want, I’m afraid. Thanks again for your time & input!
https://trac.ffmpeg.org/wiki/Encode/H.265
Most of the “commercial apps” are just wrappers for ffmpeg.
I’d start with FFmpeg. Once you see what your output is like you can then decide to move over to something more fancy.
I’m confused… Are you grabbing pirated video files from the net and re-encoding them… If you’re attempting to further compress already compressed video, you’re just zipping a zip file. It’s crazy and you’ll do nothing but bloat the file size (versus a properly compressed video file) and further reduce the quality of the video via artifacts. I’ll call the police and have you committed right now.
If you’re grabbing 8/4k or UHD BD movies and re-encoding them into lets say, 1080p HEVC 10bit, I could see that being worth it if you really love the movie (and have 5 days with nothing to do), but only if you’re going from an inferior compression to better (h264 to h265), otherwise like I said, you’re zipping a zip file.
you’ll do nothing but bloat the file size
That’s very wrong. Going from h.264 to h.265 cuts file size down to 25-50% of the original. That’s what the HE is for in HEVC.
- I’ve been encoding HEVC for a long time and I’ve not once seen a file-size drop that dramatically. You’re outrageously overestimating the file-size savings here.
- If a video file is already compressed you’ll see diminished and even negative returns by attempting to compress it further. OP seems to be taking pre compressed video files from the internet and attempting to compress them again (lossy to lossy) which is very very very stupid.
If a file is 5000kbps and you use 3000kbps you now have 25-50% savings like he just said. Nothing is overestimated, you can encode to w/e you want. This is how lossy encoding works, for everything.
Yes, but the point is keeping the quality the same. If you reencode a 5000 kbps x264 Blu-ray encode to 3000kbps x265 you will have visibly worse quality.
If you encode the corresponding Blu-ray remux with x265 to 3000 kbps the result will likely be nearly indistinguishable from the 5000 kbps x264 encode.For OP: I also prefer smaller releases so I download mostly h265 WEB-DLs. They are usually around 3000-5000 kbps (1.3-2.3GB/hour) and look fine (especially as they usually come with HDR).
Redownloading WEB-DLs in the right size will give you the best quality for the small size (and saves energy, depending on where you live).just fyi x264 and x265 are programs, written by VideoLAN organization. h264 and h265 are the codecs
And no doing that is no guarantee in visibly worse quality. Depends entirely on the video in that scenario. Plenty of them will look almost the same (though h265 is a lot blurrier than 264, I’d say h264 to h264 you’re likely to barely notice)
- This number is taken from the Wikipedia page, and represents a lab setting from totally uncompressed to full HEVC encoding. It’s not representative of data savings that you get from one video codec vs another, which is likely less than 4-6% between something like VP9 and HEVC… The 25-50% of completely fucking laughable in a realistic setting and you look like and idiot for bringing it up.
- It’s absolutely an overestimation by every conceivable metric available.
- You can encode to whatever you want, but if you take a VP9 encoded video and re-encode it, you may only save 1% and it will take several hours to encode. There are even situations where you will save nothing, or the resultant video file is actually larger even with HEVC.
It’s zipping a zip file. Endlessly re-compressing things doesn’t yield positive results in the way you describe.
It’s zipping a zip file. E
no it isn’t, zipping is lossless. encoding is lossy.
You’re correct that it will reduce file size but encoding lossy to lossy is foolish. You will introduce compression artifacts and have an objectively worse quality image, the encode will take much longer than if you used a proper lossless source, and if you don’t set your configs right you’ll strip out subtitles, tags, chapters, etc
Additionally if the h264 was already compressed by a lot h265 won’t save all that much space, giving you all the downsides with basically no upside
Only dummies encode lossy>lossy. The debate about lossy>h265 is one thing (h265 is not for archival) but h264>h265 will result in visible distortion
I also went through the phase where I thought I was an audiophile/videophile and everything I collected needed to be in ultra high quality. Eventually I realized it was stupid and now I spend a third as much money on storage and still have perfectly fine media that I have no issues with in practice despite the flaws I’m supposed to see in theory.
I don’t understand what you are arguing?
If you’re arguing that downloading remuxes and only flac is foolish then yeah, 99.8% of the time h264 and 320 mp3 are going to be indistinguishable on most setups with most content. H265 will be the same on like 99.5% of setups with slightly less content and will save tons of space. Sure. But this assumes the lossy encodes were done properly from a lossless master
if you encode lossy to lossy it will result in visible and audible distortion of the image and audio. Sometimes it’s minimal, sometimes it’s quite bad, sometimes it’s masked by your equipment, but it’s always there. Further, you’d spend more money on electricity running your cpu on full blast encoding terabytes of video files when you could simply just redownload your library in whatever format by someone who knows what they’re doing (if you’re so concerned about space and don’t care about quality go av1)
But you do you
it will result in visible and audible distortion
This is completely wrong, it “might” be visible is accurate. The actual answer is “that depends”
That’s correct, I misspoke, I clarified that with the rest of my post though
Lossy to lossy generally doesn’t matter, that is why people transcode over the web on their media players and the video seems fine, they are doing lossy to lossy on the fly there. What is actually stupid is saving media that is 100gb that you ALWAYS have to transcode to play, so no matter what you’re ramping up your wattage use to play a file. This is also why 99.9999% of consumed media is compressed, you can’t play it otherwise. Internet would explode too.
Likewise can size a jpeg down from 4000kb to 1000kb and it’ll likely be almost identical and good enough for most cases. There are certainly zero handheld devices you’d be able to notice it on.** If you size the 4000kb to 40kb now there will be an actual noticeable difference. There are different levels to all of this. **
Similarly a 100mb wav file should be a 10mb mp3 and you can’t tell the difference. You couldn’t even tell the difference with a 3mb mp3 file.
Blu-rays are compressed. “Zipping a zip file” doesn’t apply here because zips are lossless. Video encoding is almost entirely lossy, and there’s a lot of tradeoffs to be made between file size and quality. The whole point of the more efficient codecs is to minimize the quality tradeoff. There’s also a bunch of parameters to tune the resulting bitrate which is the #1 factor in deciding the final filesize.
That being said, I’ll agree that the least quality loss will come from using a Blu-ray remux since those are very high quality.
Blu-rays are compressed.
All streaming data is compressed at some point. I clearly meant not over-compressed. 4K video or UHD BD can both be taken from their original states and processed through HEVC to get crisp 1080p h265 10bit at a steep data discount. But it’ll take a very long time to process. It’s simply not worth it.
“Zipping a zip file” doesn’t apply here because zips are lossless.
It’s a figurative expression and I feel that was pretty damn obvious…
you can re-encode something at literally any bitrate, this isn’t relatable at all to “zipping a zip file” this is “opening a zip file, opening the document inside, removing data you don’t need, resaving it and rezipping it”
In my opinion the ideal x265 size/speed/quality is using a tuned slow preset, perhaps with filtering if the source is grainy. A test encode or few should be done to determine an ideal CRF per source.
Since you don’t seem very familiar with x265, I would just stick with the defaults in slow preset, but consider using aq-mode 3 or 5 (only available in the patman mod). You can also adjust the aq-strength to help control the resulting size somewhat, I wouldn’t go lower than 0.5.
Thank you for the input! I’m unsure what aq-mode is but I’m sure it’s somewhere in FastFlix. There’s tons of settings in it that I need to look over. The default CRF for HEVC in that program is 22 and I don’t know if I should go higher, lower or what but I appreciate your insight.
CRF 18 is usually where I start at, but that is focusing on quality. I’d imagine you would want a higher starting point.
If you want to find out what different options are or do, the docs are a good place to start: https://x265.readthedocs.io/en/master/cli.html
Thank you for the input and reference material. I’ll give it a look-through and appreciate your time.
