Handbrake Question: Lower quality 1080p, or higher quality 720p?

[natebetween]

Hi all,

I don't have a huge BD collection, but I've decided to start ripping and saving my collection.

I'm still messing around with the quality settings in Handbrake, and it brought up an interesting question. Would you rather have a lower quality 1080p, or higher quality 720p file?

For comparison sake, I did some test runs, and here are the size differences I came up with? Thoughts? Should I be using other settings instead? It would seem that for the most part, the increase in file size to the 1080p is still within reasonable size limits. Thanks.

Below are my results encoding several movies.
The first number is the GB size of a 720p encode at quality setting 19 (Smaller size then 1080p, higher quality)
The second number is the GB size of a 1080p encode at quality setting 20 (Lower quality than 720p, larger size)
The third number is the GB size of a 1080p encode at quality setting 19 (For comparison only)

A: 1.75___2.87___3.45
B: 2.38___4.12___5.17
C: 6.45___18.1___18
D: 2.69___4.83___5.84
E: 2.09___3.32___3.96
F: 3.42___6.16___6.01
G: 2.72___4.42___5.18
H: 2.29___4.47___5.75
I: 2.61___3.85___4.37
J: 3.58___6.5___7.73

[_Al_]

Difficult to advice how you should encode, this is entirely personal choice.
720p, CRF18, tune film, that would hold up up to 50" screen, watching from regular 8 feet distance. But as soon you increase your screen or sit closer to screen, you might start to see artifacts. Thing is, most do not care at all, at least I don't, it is just movies. To future-proof it, you'd need to leave resolution as is and be generous with bitrate and therefore more hardisk space is needed, about 3x.

[_Al_]

double post

[RazorBurn]

That really depends upon your free storage space you have left.. If you have little space remaining? Use the lowest disk space available. Save up money to buy larger drives or wait till 8TB drives comes at $150. Rip and reencoding again using the the 1080p higher quality if you have the bigger drives.. Just make sure to safely store your original BD disk..

[newpball]

Would you rather have a lower quality 1080p, or higher quality 720p file?

Honestly I do not understand how you get into this false dilemma.

You are ripping and saving your BDs to disk.
Great!

So could you give me one single sane reason why you simply don't make good quality 1080p versions?

[natebetween]

Would you rather have a lower quality 1080p, or higher quality 720p file?

Honestly I do not understand how you get into this false dilemma.

You are ripping and saving your BDs to disk.
Great!

So could you give me one single sane reason why you simply don't make good quality 1080p versions?

Ummm...space. Storage space is not infinite.

While I'm not at a loss for space right now (9TB available on my NAS), I'm saving both the ORIGINAL rip (30 to 50 GB) plus the encoded version for easy watching and portability. If I do many movies, there could, concievably, come a point at which I am out of space. I could have stored ~twice the amount of movies @ 720p vs 1080p.

What I decide now could affect me in the future. And at that point it would take FOREVER to re-encode however many movies that I would have at that point.

[LMotlow]

Maybe because 1080p isn't valid for BluRay?

Unless, of course:
- the 1080s are film speed
- O.P. doesn't care about regulation BD disc and/or will resize and re-encode as many times as possible to get a small file size.

[poisondeathray]

An important factor to consider are the source characterics .

Many retail BD's can be downscaled and upscaled with little visible loss. The so called "upscale downscale test" . These are "low actual resolution" BD's. For example , upscaled sources, older scans and prints. Those don't "deserve" 1080p treatment and shouldn't have been released as 1080p on BD in the first place

[newpball]

upscaled sources, older scans and prints.

Agree with upscaled sources and old scans.
Disagree with old prints.

Old prints may have damage but age does not significantly impact the resolution.

[Kerry56]

Would you rather have a lower quality 1080p, or higher quality 720p file?

Honestly I do not understand how you get into this false dilemma.

You are ripping and saving your BDs to disk.
Great!

So could you give me one single sane reason why you simply don't make good quality 1080p versions?

Ummm...space. Storage space is not infinite.

While I'm not at a loss for space right now (9TB available on my NAS), I'm saving both the ORIGINAL rip (30 to 50 GB) plus the encoded version for easy watching and portability. If I do many movies, there could, concievably, come a point at which I am out of space. I could have stored ~twice the amount of movies @ 720p vs 1080p.

What I decide now could affect me in the future. And at that point it would take FOREVER to re-encode however many movies that I would have at that point.

I have original discs, but the copies I make for my HTPC storage drives are 23gb (roughly) made with BD Rebuilder. I make movie-only copies and don't resize down to 720p. The visual quality is good enough I can't tell the difference from the originals when playing them. This tactic saves a lot of space compared to keeping full rips of the originals.

If I need a smaller copy for a portable device of some sort, I make those when needed with VidCoder, but I don't keep the smaller copies permanently on hard drives. When I get several made, I'll burn to a DVD/Blu-ray disc and delete them from the hdd. Basically they are just throwaway versions only used when traveling, and I rarely watch them twice on my tablet, so they get deleted and replaced on the microSD card for my tablet on a regular basis.

[poisondeathray]

upscaled sources, older scans and prints.

Agree with upscaled sources and old scans.
Disagree with old prints.

Old prints may have damage but age does not significantly impact the resolution.

I was just giving examples (you know the part you forgot to quote, "for example"...). I never said it impacts the resolution.

The point is to look at the source and what you have in your hands before deciding what to do. Just because you have a high resolution print, doesn't mean you will have a high resolution or detailed BD (effective lines of resolution)

For example, old scans can have high or low effective resolution. Old prints can have high or low effective resolution. New digital acquisition can have high or low effective resolution. Botched transfers and poor production and lead to low effective resolution

[natebetween]

Ha ha....I see this wasn't as easy a question as I originally thought.

I guess I'll just have to go with the old "personal preference" thing. I agree...at this point I would likely give Planet Earth, Hobbit, etc. the 1080p treatment, while 22 Jump Street, for example, doesn't deserve it I think.

As for quality settings 19 vs. 20 vs 21...looks like I'll have to go with personal preference on that as well. Looks like I'm going to have to start (re)watching some of these movies to see what level of artifacts I can stand.

Thank you all for your input.

[_Al_]

I'm going to have to start (re)watching some of these movies to see what level of artifacts I can stand.

Not talking now about poisondeathray's explanation quality of encoding depending on original, but artifacts that can strike your perception while watching ...

Those artifacts would show up out of the blue, where you can watch the whole movie not really noticing anything , but it could just "show up" quite noticeably in a particular shot. CRF 19 or even 18 is not enough to be artifact free. Those low gradient scenes etc. Just heads up, that you cannot eliminate scenes like that.

For example , Elysium, 720p, CRF 18, tune film, you can watch it without any visual negative feedback, but just one particular shot of that spaceship in orbit and it would look like DivX standard definition movie, but other than that you might not notice anything. Except that particular shot, where nothing is going on, but there is only one color gradient and there is nothing to look at as well that would distract those visible artifacts. You might eliminate that by watching the movie first, marking those scenes and encoding it again with -- zones, giving it more bitrate for those scenes. But that is a practical nonsense of course.

So one can try one movie , movie #1, thinking , now I know it all, only realizing with movie #2, that it is not the case. So any encoding with bitrate saving on mind gets you to that scenario.

[natebetween]

I'm going to have to start (re)watching some of these movies to see what level of artifacts I can stand.

Not talking now about poisondeathray's explanation quality of encoding depending on original, but artifacts that can strike your perception while watching ...

Those artifacts would show up out of the blue, where you can watch the whole movie not really noticing anything , but it could just "show up" quite noticeably in a particular shot. CRF 19 or even 18 is not enough to be artifact free. Those low gradient scenes etc. Just heads up, that you cannot eliminate scenes like that.

For example , Elysium, 720p, CRF 18, tune film, you can watch it without any visual negative feedback, but just one particular shot of that spaceship in orbit and it would look like DivX standard definition movie, but other than that you might not notice anything. Except that particular shot, where nothing is going on, but there is only one color gradient and there is nothing to look at as well that would distract those visible artifacts. You might eliminate that by watching the movie first, marking those scenes and encoding it again with -- zones, giving it more bitrate for those scenes. But that is a practical nonsense of course.

So one can try one movie , movie #1, thinking , now I know it all, only realizing with movie #2, that it is not the case. So any encoding with bitrate saving on mind gets you to that scenario.

Your point is well taken, and I definitely agree. I'll be looking for a good quality vs. size. So far, 19 seems to be the magic number from what I've seen today. File sizes are very acceptable.

[hello_hello]

natebetween,
You might like to take a look at the screenshots in this post. It's an example of the "upscale downscale test" poisondeathray referred to. Only one frame, but I tried 720p and 540p. You can see the difference in that particular frame, which I why I used it, but I think 90% of that video could be resized down to 540p and back and I wouldn't notice. Playing the video normally, and back at normal viewing distance, it gets even harder to pick.

There's two 720p and 540p upscaled images. The first pair with the same resizer in both directions. Then again with a softer resizer doing the "up-scaling". I think that'd probably be worst case scenario up-scaling these days, but if nothing else it shows it is out of your hands much of the time. The player/TV's up-scaling isn't something you can change..
On the odd occasion, I've resized video down to 720p and it's actually looked a bit sharper than the original 1080p version when they're both running fullscreen.... but on the (very) odd occasion.

On the subject of encoding quality, it could be argued the same CRF value gives you slightly different quality at different resolutions if the video is always being displayed on the same TV/monitor. If upscaling can make compression artefacts more noticeable, then it follows slightly lower CRF values might negate it. Think of an average Xvid/AVI as an example. Something like 656x368 etc. View it on a large monitor at it's native resolution and it can look really good. Upscale it to 1080p and it might be a different story.
I tend to use CRF18 for 720p or lower and 19 or 20 at 1080p. There's no rule.

For me it also depends how clean the source is. Noise filtering tends to blur a little so for a noisy or grainy source I'd probably denoise first, then resize down to compare 1080p denoised to 720p denoised etc. and while I dislike noise it can increase the amount of perceived picture detail at times.....

Unless you're encoding to create a Bluray/DVD compliant video disc etc, you can do whatever you like. For standalone formats such as MKV and MP4 all bets are off. I've gone with 900p several times (1600x900). It's the HD resolution for indecisive people.

I agree with _Al_.
CRF aims for the same perceived quality throughout, but nothing's perfect, and sometimes there'll be a small section of video that just doesn't look as good. I've split encodes into parts a few times in order to encode them at different CRF values, then appended the encoded video with MKVMergeGUI.
One of the main "it doesn't look as good" culprits would be colour banding, which might only show up in a scene or two, although even your original Bluray video won't be immune, so then maybe you might want to try to fix it....
Sooner or later we'll all be using the 10bit HEVC encoder (x265) and that won't be much of an issue, but for the moment it's 8 bit x264 (there is a 10 bit version of x264 but virtually no current hardware players support 10 bit).

[sophisticles]

Ummm...space. Storage space is not infinite.

Ummm...why rip and re-encode in the first place? I have hundreds of Blu-Rays, I just keep them as they are and pop which ever one I want to watch into my BD player. If you want to "back up" then buy a BD burner and some good BD media and do a real "back up".

[hello_hello]

Ummm...why rip and re-encode in the first place? I have hundreds of Blu-Rays, I just keep them as they are and pop which ever one I want to watch into my BD player. If you want to "back up" then buy a BD burner and some good BD media and do a real "back up".

How have you managed to go this long without requiring medication or heavy sedation. Waiting for DVD/Bluray discs to load and menus to open drives me nuts, and I'm not particularly impatient.

You can't beat the convenience of a hard drive connected to a TV's USB media player. Or even to a USB Bluray Player. These days if I decide to watch a different program and it necessitates swapping hard drives in the USB dock it irritates me a little, and I'm not particularly lazy. But discs!!

I tried inserting a Bluray disc into my Android tablet recently but I didn't have much luck.

[natebetween]

@hello_hello : Thanks for all of the great information. This is more information than I was looking for...in a good way. Too funny about 900p...I never even thought of that. I guess when all is said and done I'm looking for a good compromise between size, quality, and time taken away from my real life. I doubt that I'd ever come across a situation when I'd encode different sections of a movie at different values and splice them back together. I just want convenience that I wouldn't mind paying a little visual price for...but I wouldn't want it to distract from the movie in any crazy ways. To be honest, the worse I've ever seen is when a completely black scene is just full of blocks and it is completely distracting. A few artifacts or slight fuzziness here and there is a small price to pay. Still playing around....but so far 720p at 18 or 19, and 1080p @ 20 seem to be working well for me. Still testing...

@sophisticles : Exactly what hello_hello said. Each of our 3 bedrooms and the living room all have a WDTV hooked to my NAS. Historically I've used it to play TV shows (not available on Netflix and Hulu) and music from my collection. Playing my BD movies CENTRALLY from my NAS has been one of the best decisions that I've made and is extremely liberating...especially since I don't own 4 BD players. With a little work with Thumbgen I've generated artwork and am on my way to a true movie jukebox.

There's also an added bonus. I often take my Nexus 10 with me on vacation (great screen) and having a portable file format would be fantastic. I've already played with it, and the transfer rate to my SSD USB drive is fantastic. I can load several movies in a matter of minutes, and play with my tablet via the OTG connector.

[_Al_]

if you are into this home cloud thing, future etc, ..., those things progress fast

NAS will be "tomorrow" capable of live transcoding for not that much money, so beside home cloud, that could be extended over web as well ...... in this case - transcoding live - maybe backing up originals would be better solution, but for sure that is still costly ...

[natebetween]

Wow...transcoding on the fly...didn't know that was a thing that was happening. Probably will be little while before it makes sense for me. Currently rocking 4 x 3TB hard drives with 1 drive in parity for a total of 9TB of storage. 2 bays makes this kind of hard if you are storing full rips on the NAS. Still requires a lot of physical space.

I keep testing this and it looks like 19 is my magic number. I give up a little bit in scenes with very little action, or very grey/black scenes...but so far it's nothing I can't live with. I'm going to play with 18 to see if the additional space is worth it.

Overall, my BD collection isn't near big enough to worry about this now...but I can foresee it getting big enough in the future. Especially with extended edition LOTR, Hobbit Trilogy, and Back to the Future on my wish list...and the Harry Potter complete on my Wife's wish list. Keep going like this and the space requirements will eat up my NAS (along with my TV shows and lossless FLAC music collection).

[jagabo]

I keep testing this and it looks like 19 is my magic number. I give up a little bit in scenes with very little action, or very grey/black scenes...

Try using aq-mode=3. That gives a little more bitrate to dark scenes, keeping the grain that prevents posterization artifacts. If that's not enough try adding aq-strength=1.3 or more.

[natebetween]

Oh...thanks for the advice. I haven't even begun to dive into the advanced settings. I suppose that's my next bit of homework, to learn about what all the settings do, and which are worth using for me. Thanks.

[jagabo]

By the way aq-mode 3 was only recently added to x264. I don't know if Handbrake has been updated yet. You'll probably need a nightly build, not the stable release. You can use aq-strength without aq-mode=3. You have to use the Extra Options box in Handbrake for those.

[sophisticles]

Ummm...why rip and re-encode in the first place? I have hundreds of Blu-Rays, I just keep them as they are and pop which ever one I want to watch into my BD player. If you want to "back up" then buy a BD burner and some good BD media and do a real "back up".

How have you managed to go this long without requiring medication or heavy sedation. Waiting for DVD/Bluray discs to load and menus to open drives me nuts, and I'm not particularly impatient.

You can't beat the convenience of a hard drive connected to a TV's USB media player. Or even to a USB Bluray Player. These days if I decide to watch a different program and it necessitates swapping hard drives in the USB dock it irritates me a little, and I'm not particularly lazy. But discs!!

I tried inserting a Bluray disc into my Android tablet recently but I didn't have much luck.

Funny. Yeah those 20 seconds that it takes to load a BD or DVD is really long compared to the hours it takes to re-encode a full length Blu-Ray, ROTFLMAO!!!

And you watch movies on an Android tablet? WHY? Given a chance between watching a movie on a TV while sitting in my recliner and taking my medication or watching on an Android while being heavily sedated I would put I Wanna Be Sedated by The Ramones on loop on my IPOD and become Comfortably Numb.

[hello_hello]

Funny. Yeah those 20 seconds that it takes to load a BD or DVD is really long compared to the hours it takes to re-encode a full length Blu-Ray, ROTFLMAO!!!

Re-encoding is also partly a hobby for some, although I think you might find the computer generally does the bulk of the work without user input whether it takes 2 hours or 2 days to re-encode a movie.

And you watch movies on an Android tablet? WHY? Given a chance between watching a movie on a TV while sitting in my recliner and taking my medication or watching on an Android while being heavily sedated I would put I Wanna Be Sedated by The Ramones on loop on my IPOD and become Comfortably Numb.

I'm happy for you if you're able to take your TV and recliner with you everywhere you go, but the rest of us live in a different world. In that world portable devices also have HDMI or MHL outputs so "watching movies on a portable device" wouldn't be a wise assumption.

Once the disc is ripped and re-encoded you're no longer limited to using a Bluray player so your definition of a "real backup" doesn't apply.

You might consider ripping and re-encoding to be unnecessary or a waste of time, but from my perspective it's not in the same ballpark as posting in a forum dedicated to Bluray ripping just to inform everyone you don't rip your own discs. Did you have some time to kill waiting for the Bluray player to start up?

[hello_hello]

I keep testing this and it looks like 19 is my magic number. I give up a little bit in scenes with very little action, or very grey/black scenes...

Try using aq-mode=3. That gives a little more bitrate to dark scenes, keeping the grain that prevents posterization artifacts. If that's not enough try adding aq-strength=1.3 or more.

I generally don't seem to have many problems with dark scenes being of lower quality. Not that I've noticed. But thinking about it, I don't use --aq-mode 2 at all, only --aq-mode 1. Not for any reason other than --aq-mode 1 is the default and I've never really had a need to think about it. Maybe Tune Film helps, as I generally use it.
Out of curiosity though I ran three test encodes. One hour of video with lots of dark scenes. CRF18, Tune Film and Preset Slow each time.

--aq-mode 1, 723 MB
--aq-mode 2, 634 MB
--aq-mode 3, 871 MB

The large difference in bitrate surprised me a little, and I'm still trying to get my head around why --aq-mode 3 increased it so much, but visually I'm not seeing any difference between any of them. Sure, when looking at individual frames it's obvious they're not always encoded exactly the same way, but the difference isn't a quality one as such. I deliberately put the encoded video through a PC to TV levels conversion to make what's going on in the dark regions easier to see and I'm not seeing any real difference.

I'll try again later with a different video and a higher CRF value.

[natebetween]

Wow....that's very interesting...a decrease...and then an increase. Seems unexpected...

I have to admit that I didn't think that I'd be diving this deep into the additional options. I figured that I'd just "set it and forget it." I've been doing all the reading I can to see what I can change, and what's worth changing given the tradeoffs with file size and encoding time.

I found the article below, which I've been toying with some of the settings. It's over 2 years old though....so I'm looking for more recent articles, besides the Wiki, which is very dry and not user friendly...haha.

http://textonlytech.blogspot.com/2013/02/tech-tip-handbrake-part-2.html

[hello_hello]

I don't play around with x264's settings much at all myself. At least not for my usual encoding. About the only time I experiment is when a thread like this prompts me to do so, but I invariably go back to picking a tuning a CRF value most of the time.

If you adjust the x264 speed preset, all it does it automatically adjust some of x264's advanced options in a pre-defined way. As does adjusting the tuning. For example preset medium uses --me hex and --subme 7, while the slow preset uses --me umh and --subme 8.
In a perfect world when you select a speed preset using HandBrake, you'd switch to the advanced tab and it'd show you the new defaults, but it doesn't. I think that's why at some stage it was configured so enabling the advanced tab disabled the presets. Many GUI's let you use both. If you change a speed preset, the relevant advanced settings are changed to their new defaults and you can fine tune/fiddle with the advanced settings from there.

I haven't found an easy to read list of all x264's default settings, but this shows you which settings are changed by a speed preset and what they're changed to. http://forum.doom9.org/showthread.php?t=148149
Edit: Actually no it doesn't. I just realised it's old and the very slow speed preset isn't included, so it's outdated, but it'll give you an idea. Anyone know of an up to date list of settings?

I've run a a few more encodes. Half an hour of a movie this time, 1280x536, Level 4.1, Preset Slow, Tune Film, no fast P-Skip (I just realised the latter was enabled for one of my MeGUI encoder presets so I'll leave it enabled when running more test encodes).

CRF16 --aq-mode 1, 788 MB
CRF16 --aq-mode 2, 718 MB
CRF16 --aq-mode 3, 913 MB

CRF18 --aq-mode 1, 527MB
CRF18 --aq-mode 2, 490 MB
CRF18 --aq-mode 3, 634 MB

CRF20 --aq-mode 1, 363 MB
CRF20 --aq-mode 2, 347 MB
CRF20 --aq-mode 3, 427 MB

CRF23 --aq-mode 1, 226 MB
CRF23 --aq-mode 2, 220 MB
CRF23 --aq-mode 3, 259 MB

I've only looked at the CRF 16 and CRF 18 encodes so far. There's visual differences when comparing individual frames but quality-wise it all looks pretty much the same to me (comparing the CRF16 encodes to each other and the CRF18 encodes to each other). If anything --aq-mode 2 and --aq-mode 3 seem a tad more prone to blocking in some areas, but I'll keep looking.
I'm fairly sceptical about --aq-mode 3. If --aq-mode 1 is standard adaptive quantisation and --aq-mode 2 is supposed to be some sort of auto-variance, then so far it appears it does it's job (less bits for the same quality at given CRF value). If --aq-mode 3 is supposed to be a more clever version of --aq-mode 2, I don't see how, given compared to --aq-mode 1 it increases the bitrate by a fair amount.

Anyway, I haven't looked closely at the CRF 20 or CRF 23 encodes yet, but I will later today. Maybe I'll start to see quality differences. I'll also run some more test encodes using the same bitrate, rather than the same CRF value.

I ran a quick search for other recent --aq-mode 3 discussions here. I only found this one where it was claimed --aq-mode 3 makes a "significant difference" to quality, but as tends to happen, the "did you compare at the same bitrate" question seemed to be a thread stopper. https://forum.videohelp.com/threads/369199-Anyone-try-x264-s-new-AQ-mode

[jagabo]

I haven't found an easy to read list of all x264's default settings, but this shows you which settings are changed by a speed preset and what they're changed to. http://forum.doom9.org/showthread.php?t=148149
Edit: Actually no it doesn't. I just realised it's old and the very slow speed preset isn't included, so it's outdated, but it'll give you an idea. Anyone know of an up to date list of settings?

I made a table a few years ago (using mediainfo to extract the list of settings):

ultrafast: cabac=0 / ref=1  / deblock=0:0:0 / analyse=0:0       / me=dia  / subme=0  / psy=1 / psy_rd=1.00:0.00 / mixed_ref=0 / me_range=16 / chroma_me=1 / trellis=0 / 8x8dct=0 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=0  / threads=6 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / bluray_compat=0 / constrained_intra=0 / bframes=0                                                                           / weightp=0 / keyint=250 / keyint_min=23 / scenecut=0  / intra_refresh=0                   / rc=crf / mbtree=0 / crf=16.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40                 / aq=0
superfast: cabac=1 / ref=1  / deblock=1:0:0 / analyse=0x3:0x3   / me=dia  / subme=1  / psy=1 / psy_rd=1.00:0.00 / mixed_ref=0 / me_range=16 / chroma_me=1 / trellis=0 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=0  / threads=6 / 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=1 / keyint=250 / keyint_min=23 / scenecut=40 / intra_refresh=0                   / rc=crf / mbtree=0 / crf=16.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40 / pb_ratio=1.30 / aq=1:1.00
veryfast:  cabac=1 / ref=1  / deblock=1:0:0 / analyse=0x3:0x113 / me=hex  / subme=2  / psy=1 / psy_rd=1.00:0.00 / mixed_ref=0 / me_range=16 / chroma_me=1 / trellis=0 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=0  / threads=6 / 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=1 / keyint=250 / keyint_min=23 / scenecut=40 / intra_refresh=0 / rc_lookahead=10 / rc=crf / mbtree=1 / crf=16.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40                 / aq=1:1.00
faster:    cabac=1 / ref=2  / deblock=1:0:0 / analyse=0x3:0x113 / me=hex  / subme=4  / psy=1 / psy_rd=1.00:0.00 / mixed_ref=0 / me_range=16 / chroma_me=1 / trellis=1 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=0  / threads=6 / 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=1 / keyint=250 / keyint_min=23 / scenecut=40 / intra_refresh=0 / rc_lookahead=20 / rc=crf / mbtree=1 / crf=16.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40                 / aq=1:1.00
fast:      cabac=1 / ref=2  / deblock=1:0:0 / analyse=0x3:0x113 / me=hex  / subme=6  / 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=6 / 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=1 / keyint=250 / keyint_min=23 / scenecut=40 / intra_refresh=0 / rc_lookahead=30 / rc=crf / mbtree=1 / crf=16.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40                 / aq=1:1.00
medium:    cabac=1 / ref=3  / deblock=1:0:0 / analyse=0x3:0x113 / me=hex  / subme=7  / 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=6 / 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=16.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40                 / aq=1:1.00
slow:      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=6 / 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=crf / mbtree=1 / crf=16.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40                 / aq=1:1.00
slower     cabac=1 / ref=8  / deblock=1:0:0 / analyse=0x3:0x133 / me=umh  / subme=9  / psy=1 / psy_rd=1.00:0.00 / mixed_ref=1 / me_range=16 / chroma_me=1 / trellis=2 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=-2 / threads=6 / 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=60 / rc=crf / mbtree=1 / crf=18.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40                 / aq=1:1.00
veryslow:  cabac=1 / ref=16 / deblock=1:0:0 / analyse=0x3:0x133 / me=umh  / subme=10 / psy=1 / psy_rd=1.00:0.00 / mixed_ref=1 / me_range=24 / chroma_me=1 / trellis=2 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=-2 / threads=6 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / bluray_compat=0 / constrained_intra=0 / bframes=8  / 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=60 / rc=crf / mbtree=1 / crf=16.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40                 / aq=1:1.00
placebo:   cabac=1 / ref=16 / deblock=1:0:0 / analyse=0x3:0x133 / me=tesa / subme=11 / psy=1 / psy_rd=1.00:0.00 / mixed_ref=1 / me_range=24 / chroma_me=1 / trellis=2 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=0 / chroma_qp_offset=-2 / threads=6 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / bluray_compat=0 / constrained_intra=0 / bframes=16 / 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=60 / rc=crf / mbtree=1 / crf=16.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40                 / aq=1:1.00

I don't know if any have changed since I made the table. All that is without specifying a tuning option.

[sophisticles]

I ran a quick search for other recent --aq-mode 3 discussions here. I only found this one where it was claimed --aq-mode 3 makes a "significant difference" to quality, but as tends to happen, the "did you compare at the same bitrate" question seemed to be a thread stopper. https://forum.videohelp.com/threads/369199-Anyone-try-x264-s-new-AQ-mode

Based on your tests with different CRF values and the resulting bit rate differences a prudent person would conclude that aq-mode 3 takes the old "no replacement to displacement" approach to increasing quality in dark areas, by specifically using more bit rate in those areas.

It's as I've always said: the 2 most important factors in the quality of an encode is the quality of the source and how much bit rate is used, everything else is just details.