Page 1 of 1 |
|
Posted: Sun, 11th Sep 2016 04:51 Post subject: Bitrate vs size in movie releases |
|
 |
|
|
Back to top |
|
 |
Morphineus
VIP Member
Posts: 24883
Location: Sweden
|
Posted: Sun, 11th Sep 2016 04:57 Post subject: |
|
 |
|
|
Back to top |
|
 |
NFOAC
Posts: 6015
Location: India
|
Posted: Sun, 11th Sep 2016 05:00 Post subject: |
|
 |
Higher bit rate is in higher quality unless poorly encoded
|
|
Back to top |
|
 |
|
Posted: Sun, 11th Sep 2016 05:45 Post subject: |
|
 |
Yes in fact it's even better to have a low res 720p release at a decent bitrate than a 1080p one with shitty bitrate since in motion it can look even worse than 720p thx to ghosting and artifacts.
As a simple rule, always check the bitrate if you can (atleast 8000 kbit/s for 1080p), otherwise just dont go below 8 GB for a 1080p action movie of up to 2 hours length - unless it's an x265 encode.
|
|
Back to top |
|
 |
|
Posted: Sun, 11th Sep 2016 08:32 Post subject: |
|
 |
You should have included encoding settings. Perhaps this film doesn't really need 9k bitrate. If it gets ~6K bitrate when encoding at crf18 then I see no reason quality-wise for a higher bitrate. Actually, in some cases even crf20 might be enough, it depends. But crf18 is a safer bet.
1) Lenovo Legion 7 (AMD Ryzen 7 5800H, RTX 3080 16Gb, 32Gb DDR4, SSD 1TB +2TB
2) SFFPC (streaming via Moonlight+ Sunshine)
Last edited by h0rnyfavn on Sun, 11th Sep 2016 08:36; edited 2 times in total
|
|
Back to top |
|
 |
|
Posted: Sun, 11th Sep 2016 08:34 Post subject: |
|
 |
For me it's 10 mbits at least which is why I prefer p2p over scene.
|
|
Back to top |
|
 |
|
Posted: Sun, 11th Sep 2016 08:44 Post subject: |
|
 |
OK, I've checked it FGT version was encoded using crf20\mbtree as I suspected.That explains it. There might be nothing wrong with it. However, Their other encoding settings need some improvement.
If it was me I would increase some settings and then compare 2 sample files featuring different scenes (dark and light,action) @ crf18\mbtree VS crf18 no mbtree . If mbtree version looks fine (sometimes it screws up dark scenes) then I would go for it. crf20 for some 1080p encodes might be a solid choice too though but you'd better check what you get coz sometimes films require more than this (especially dark scenes\ heavy action) That's why I would chose a "universal" crf18 mode instead.
After x264 got improved I've never understood this obsession with 2 pass encoding unless you wanna go for a specific size. Actually, the best encoding groups from hdbits mostly use crf but their other encoding settings are considerably more demanding (CPU-wise) than that of FGT releases.
1) Lenovo Legion 7 (AMD Ryzen 7 5800H, RTX 3080 16Gb, 32Gb DDR4, SSD 1TB +2TB
2) SFFPC (streaming via Moonlight+ Sunshine)
|
|
Back to top |
|
 |
|
Posted: Sun, 11th Sep 2016 11:17 Post subject: Re: Bitrate vs size in movie releases |
|
 |
rax369 wrote: | So since I'm not pro on this, came here wondering if anyone could explain if it's worthy or not to have a movie with higher bitrate but bigger in size? |
I mean, like.. this is the most important thing to look for (when there's a significant bitrate difference - if the difference is small then the differences in picture quality should depend more on the various encoding settings rather than the bitrate).
So yes.
TWIN PEAKS is "something of a miracle."
"...like nothing else on television."
"a phenomenon."
"A tangled tale of sex, violence, power, junk food..."
"Like Nothing On Earth"
~ WHAT THEY'RE TRYING TO SAY CAN ONLY BE SEEN ~
http://www.youtube.com/watch?v=CHTUOgYNRzY
|
|
Back to top |
|
 |
|
Posted: Sun, 11th Sep 2016 11:46 Post subject: |
|
 |
I don't think the human eye can perceive any difference in quality above 2400Kbit/s
|
|
Back to top |
|
 |
|
Posted: Sun, 11th Sep 2016 11:50 Post subject: |
|
 |
What do we think is an average file size for a good quality 4K movie?
I'm seeing ranges of between 60 - 120 GB for a film?
|
|
Back to top |
|
 |
|
Posted: Sun, 11th Sep 2016 12:09 Post subject: |
|
 |
@nik_nak
It depends on a film,scene complexity etc.
If we are talking about rips you should stop judging rips quality based on just bitrates\sizes.
How is the "best" quality\size-wise bitrate determined?
You encode a sample file that takes different scenes from the whole film(so you have still scenes, dynamic scenes, dark\light and all)
For instance
selectTotal1=framecount()/100
selectTotal2=selectTotal1*2
selectrangeevery(selectTotal2,50)
(that's part of avisynth script)
Of course, you can increase your sample file duration to have a more precise results.
You set encoding settings and then set crf, say, 18 (all around good choice).
Then you check I P B QP values in your x264 log. The more dynamic video sequence is the higher the difference between them and the higher B QP is.
As a rule of thumb if all frames do not exceed 18 QP then bitrate you got can be reduced by ~25%( if you make a 2 pass encoding) or CRF you used can be reduced a bit and maintain the 'same"quality. If all frames QP are higher than 22 -23 then bitrate is too low and you need to raise it. For a highly dynamic video an average QP (B only) at ~23+ might be acceptable.(visual check needed)
+ different encoding settings affect what you get. Say, mbtree can drastically reduce your rip size maintaining visually the same quality as without it. But it would increase I P B values (which might not matter much IF visually it looks fine). + it may screw up "black" in you rips.(more blocky).
You should always check what you get as in see how it looks like regardless of QPs values. Especially if a film has really complex scenes where really small details matter, like ,say, five thousands ants mating simultaneously under a palm's shadow. (here crf18 might be insufficient)
What I'm trying to say is if a rip is made properly you don't need to freak out that bitrate\size seems to be low. It all depends. And here a reputation of encoders comes into play.. Otherwise just stick to to bdrips with bitrate numbers that you feel should be fine quality wise.
PS FGT guys seem to be the ones who want to make a fast rip and release it first.So, their rips quality can differ greatly.
1) Lenovo Legion 7 (AMD Ryzen 7 5800H, RTX 3080 16Gb, 32Gb DDR4, SSD 1TB +2TB
2) SFFPC (streaming via Moonlight+ Sunshine)
Last edited by h0rnyfavn on Sun, 11th Sep 2016 12:53; edited 1 time in total
|
|
Back to top |
|
 |
|
|
Back to top |
|
 |
|
Posted: Sun, 11th Sep 2016 15:38 Post subject: |
|
 |
And don't judge the picture quality (just) by freezing a frame. That's not how you watch a movie. Compare two clips.
TWIN PEAKS is "something of a miracle."
"...like nothing else on television."
"a phenomenon."
"A tangled tale of sex, violence, power, junk food..."
"Like Nothing On Earth"
~ WHAT THEY'RE TRYING TO SAY CAN ONLY BE SEEN ~
http://www.youtube.com/watch?v=CHTUOgYNRzY
|
|
Back to top |
|
 |
|
Posted: Sun, 11th Sep 2016 18:02 Post subject: |
|
 |
h0rnyfavn wrote: | You should have included encoding settings. |
Oh ok, hopefully I didn't miss nothing else now, here are the encoding settings for both:
SECTOR7:
Writing library : x264 core 148 r2705 3f5ed56
Encoding settings : cabac=1 / ref=5 / deblock=1:0:0 / analyse=0x3:0x113 / me=umh / subme=8 / psy=1 / psy_rd=1.00:0.00 / mixed_ref=1 / me_range=16 / chroma_me=1 / trellis=1 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=-2 / threads=12 / lookahead_threads=2 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / bluray_compat=0 / constrained_intra=0 / bframes=3 / b_pyramid=2 / b_adapt=2 / b_bias=0 / direct=3 / weightb=1 / open_gop=0 / weightp=2 / keyint=250 / keyint_min=23 / scenecut=40 / intra_refresh=0 / rc_lookahead=50 / rc=2pass / mbtree=1 / bitrate=9361 / ratetol=1.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / cplxblur=20.0 / qblur=0.5 / ip_ratio=1.40 / aq=1:1.00
Default : Yes
Forced
FGT:
Writing library : x264 core 148 r2705 3f5ed56
Encoding settings : cabac=1 / ref=5 / deblock=1:-1:-1 / analyse=0x3:0x133 / me=umh / subme=7 / psy=1 / psy_rd=1.00:0.15 / mixed_ref=1 / me_range=16 / chroma_me=1 / trellis=1 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=0 / chroma_qp_offset=-3 / threads=24 / lookahead_threads=4 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / bluray_compat=0 / constrained_intra=0 / bframes=3 / b_pyramid=2 / b_adapt=1 / b_bias=0 / direct=1 / weightb=1 / open_gop=0 / weightp=2 / keyint=250 / keyint_min=23 / scenecut=40 / intra_refresh=0 / rc_lookahead=40 / rc=crf / mbtree=1 / crf=20.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40 / aq=1:1.00
Watched last night SECTOR7's encode, cann't say I could complain about something.
Watched it with my local streaming player (NMT A-400) and the video quality specially in the details was superb. You know these kind of movies uses lots of high quality FX, so that's why I like to get these kind of encodes @ 1080p.
"tR@Nce is nOT JusT mUsIK, iTs A wAy oF lIFe... ITWT"
________
R@x
|
|
Back to top |
|
 |
Nalo
nothing
Posts: 13515
|
Posted: Sun, 11th Sep 2016 21:02 Post subject: |
|
 |
Last edited by Nalo on Wed, 3rd Jul 2024 06:02; edited 2 times in total
|
|
Back to top |
|
 |
|
|
Back to top |
|
 |
|
Posted: Mon, 12th Sep 2016 07:34 Post subject: |
|
 |
|
|
Back to top |
|
 |
|
|
Back to top |
|
 |
|
Posted: Thu, 15th Sep 2016 20:25 Post subject: |
|
 |
One man's garbage is another's "perfect HD quality!!!1".
To give some perspective, Netflix's 1080p streams up until last year were 5800kbps (nowadays they are "per title"). So if you had seen some movies on Netflix and thought they looked fine, you don't necessarily have to go after 8000 encodes.
TWIN PEAKS is "something of a miracle."
"...like nothing else on television."
"a phenomenon."
"A tangled tale of sex, violence, power, junk food..."
"Like Nothing On Earth"
~ WHAT THEY'RE TRYING TO SAY CAN ONLY BE SEEN ~
http://www.youtube.com/watch?v=CHTUOgYNRzY
|
|
Back to top |
|
 |
|
Posted: Sun, 25th Sep 2016 22:20 Post subject: |
|
 |
In plain terms for the not know in encoding
SECTOR7: they used a 2pass average bit encoding with max bitrate of 9361 for fast movement = encoding takes a lot more time but the results are pristine and files are smaller .... it does two passes , one calculates where is the movement, and in what bit rates it encodes ...the other does the work
FGT: did the easy fast encode of crf=20.0 at fast pass (normally for 1080p you do crf=18 , for 720p crf=19)... crf=20 means 4000-5000 max , washed up color scheme , all compressing done at the fly ... with look ahead that rapes the CPU ... only good if you use with CUDA ... it compresses faster but the results are sometimes the same in quality , BUT the file sizes of that done project is larger
......
if you care about file size you go for 2pass encoding with set bit rate, but you have to know if its fast or slow moving project , to set the correct the bitrate ... if you dont care about the file size and care about fast results of encoding you go constant rate factor , but when the PC hiccups you can get marcoblocking arctifacts

|
|
Back to top |
|
 |
|
Posted: Sun, 25th Sep 2016 22:22 Post subject: |
|
 |
No comment.
TWIN PEAKS is "something of a miracle."
"...like nothing else on television."
"a phenomenon."
"A tangled tale of sex, violence, power, junk food..."
"Like Nothing On Earth"
~ WHAT THEY'RE TRYING TO SAY CAN ONLY BE SEEN ~
http://www.youtube.com/watch?v=CHTUOgYNRzY
|
|
Back to top |
|
 |
|
Posted: Sun, 25th Sep 2016 22:32 Post subject: |
|
 |
2 pass encoding is for old farts who are stuck in the past when x264 was still in the middle of active development
A properly made CRF rip is the way to go. 
1) Lenovo Legion 7 (AMD Ryzen 7 5800H, RTX 3080 16Gb, 32Gb DDR4, SSD 1TB +2TB
2) SFFPC (streaming via Moonlight+ Sunshine)
|
|
Back to top |
|
 |
|
Posted: Mon, 26th Sep 2016 02:57 Post subject: |
|
 |
depends on your eye sight as my eye sight bad I can watch DVD quality smaller downloads.
|
|
Back to top |
|
 |
|
Posted: Mon, 26th Sep 2016 04:17 Post subject: |
|
 |
h0rnyfavn wrote: | 2 pass encoding is for old farts who are stuck in the past when x264 was still in the middle of active development
A properly made CRF rip is the way to go.  |
if you dont care about space , but if space is the point here 2 pass is the only way to go when you can control the quality ... crf is a roll of the dice how much a file will be
consolitis wrote: | No comment. |
whats so funny , troll
|
|
Back to top |
|
 |
|
Posted: Mon, 26th Sep 2016 08:53 Post subject: |
|
 |
@Hfric
Well, yes. The point is if you want a certain quality (which tbh you should), not just blindly setting a certain bitrate\ size, your 2 pas encode will be practically the same size-wise as crf encode. 2 pass was popular when we were stuck with DVDs but now that we have lots of HDD space even on mobiles it has become irrelevant. Actually it was scene with their stupid rules that influenced the most this nonsense. They used to make 4.5 rips even for 720p encodes even when films didn't require much bitrate or required more bitrate.. Such a waste of time
and,sometimes, space.
Still,
Quote: | only good if you use with CUDA |
huh?
CUDA is shit for encoding when it comes to quality unless you are talking about DGDecNV but that's different and tbh it doesn't help much except for indexing them huge blurays much faster. AFAIK It can help with lookaheads though but not much.
As for "raping CPU"... encoding "rapes" CPUs in general. Btw it's the best way to test OC stability. Not ridiculous benchmarks. I used to pass benchmarks just fine for hours and then crash in minutes on encoding stability test. And tbh I didn't see any difference in raping when it comes to 2pass and CRF. Same load and same temps.
1) Lenovo Legion 7 (AMD Ryzen 7 5800H, RTX 3080 16Gb, 32Gb DDR4, SSD 1TB +2TB
2) SFFPC (streaming via Moonlight+ Sunshine)
|
|
Back to top |
|
 |
Page 1 of 1 |
All times are GMT + 1 Hour |