Anything like DivX vs XviD vs x264 2012 tests?

[zmejce]

Im interested which one is better in 2012,

Back in the days of DivX 3.11 and nanDub there wasnt XviD.
Later XviD came out, and became (probably?) better than DivX 3.
How are they compared in 2012 with the latest releases?
I saw few screenshots, DivX vs XviD, the results from what i saw was that XviD has sharper image (better), DivX has smoother image (worse), but DivX has slightly better colors than XviD.

Also, what about the filesize, encoding and decoding speed (which one can be played on weaker pc)?
If latest DivX is smoother than XviD, than maybe it can be compared with x264 for HD content?

[Asmodian]

I don't think much has changed in the MPEG-4 ASP world.

Also, what about the filesize, encoding and decoding speed (which one can be played on weaker pc)?
If latest DivX is smoother than XviD, than maybe it can be compared with x264 for HD content?

File size cannot be separated from quality. Both are very fast to decode or encode in software (decode is the same for both) but now there is a lot of hardware available to decode mpeg-4 AVC which is not used for mpeg-4 ASP.

Anything can be compared with x264 but I know DivX doesn't compare well to x264, HD or not.

[Atak_Snajpera]

it should be written as
mpeg-4 asp (divx,xvid...)
mpeg-4 avc or h.264 (x264,atheme...)

[Asmodian]

it should be written as
mpeg-4 asp (divx,xvid...)
mpeg-4 avc or h.264 (x264,atheme...)

oops! Thanks :o

[zmejce]

Anything can be compared with x264 but I know DivX doesn't compare well to x264, HD or not.

I've did some dvd encodes to x264, they were smooth... XviD was better.

[LoRd_MuldeR]

I've did some dvd encodes to x264, they were smooth... XviD was better.

If, at the same bitrate, XviD looks better than x264, you did something seriously wrong... maybe with your pre-processing, maybe with your encoder settings or maybe with your comparing method.

See also:
http://x264.nl/developers/Dark_Shikari/website/compare.html

[zmejce]

http://x264.nl/developers/Dark_Shikari/website/compare.html

Interesting site, maybe that's what I'm looking for. Do you know from which date are those screenshots?

According to them, x264 is best, then DivX, then others (maybe XviD, & a good mp2 encoder should be similar to -original source- screenshots).

Which birate do you recommend for HQ rip at 720x480 res. from dvd?

[Atak_Snajpera]

[Asmodian]

I don't understand what you what to know. Are you really thinking Xvid might be better quality compared to x264?

Currently x264 is accepted as the highest quality video encoder available. I think this is true even counting the many thousands of dollars encoder packages (they offer other advantages of course) but it certainly true compared to the cheap or free options.

MPEG-4 ASP (Xvid, Divx) is a simpler standard that's easier to decode, as such it cannot achieve the same quality per size. This is even more true of MPEG2.

Which bit rate to use is up to you just make sure it is the same for all comparison encodes.

[zmejce]

I wanted to know which one of DivX and XviD is better in 2012.
I think to go with 1500-2000kbps birate DivX rip for video with 720x480 resolution from DVD source.

[zmejce]

If, at the same bitrate, XviD looks better than x264, you did something seriously wrong... maybe with your pre-processing, maybe with your encoder settings or maybe with your comparing method.


Oh no, Ive just used Handbrake 0.9.5 with High x264 preset, I didnt do any processing besides deinterlacing, still image is smoother with those rips, mp4 asp (divx or xvid) is sharper and clearer.

But on 1080p, x264 is absolutely best.

[hello_hello]

I've did some dvd encodes to x264, they were smooth... XviD was better.

I go out on a limb here, maybe set myself up for ridicule...

If someone were to ask me, I'd say x264 is better than Xvid. No doubt. No room for discussion.
Very occasionally though, and I can't quite put my finger on it... sometimes when encoding low quality sources, Xvid seems a little more pleasing to the eye to me. Probably because it's not as good at retaining detail so where normally you'd lose some fine detail (compared to x264) you lose some of the unwanted noise or compression artifacts instead. Just a theory anyway.... but for 99% of the video I encode, I'd use x264, and I don't think I'd ever say Xvid was sharper or clearer.

[poisondeathray]

Oh no, Ive just used Handbrake 0.9.5 with High x264 preset, I didnt do any processing besides deinterlacing, still image is smoother with those rips, mp4 asp (divx or xvid) is sharper and clearer.

But on 1080p, x264 is absolutely best.


Maybe that was your problem?

If it was a theatrical source (e.g. a movie), you wouldn't deinterlace it

[CWR03]

Very occasionally though, and I can't quite put my finger on it... sometimes when encoding low quality sources, Xvid seems a little more pleasing to the eye to me.
I know exactly what you mean. I had a Creative DVR card that captured in half-height resolution even at its best settings - I exported a lot of its files and saved them encoded in XviD, some of which look nearly as good (if not a bit better) as the DVD release of the same show that I encoded to x264. I never tried to figure out why, just always thought it was interesting.

[datauser]

QUOTE by hello_hello
If someone were to ask me, I'd say x264 is better than Xvid. No doubt. No room for discussion.
Very occasionally though, and I can't quite put my finger on it... sometimes when encoding low quality sources, Xvid seems a little more pleasing to the eye to me. Probably because it's not as good at retaining detail so where normally you'd lose some fine detail (compared to x264) you lose some of the unwanted noise or compression artifacts instead. Just a theory anyway.... but for 99% of the video I encode, I'd use x264, and I don't think I'd ever say Xvid was sharper or clearer.

Have to agree too! Might also explain why if you use a high bitrate/high resolution x264 source and recompress it to avi, to find that it mostly comes out looking better than if you just encoded to avi/Xvid in the first place from your direct source to same specifications in my subjective opinion.

The x264 has done all the hard work preserving the source detail etc, so all you are really doing is some fine tuning with Xvid and losing unwanted noise artifacts in the compression process rather like a fine tooth comb filtering process/sieve. Ironic since there are no denoise options built in Xvid unlike the option in DivX to remove noise during the encoding process.

[zmejce]

It might be strange, but DivX 6.9 1080p rip @ 7500kbps (2pass encoding and Highest quality setting) is better than x264 rip (with Handbrake's High Profile @ ~13.3Mbps).

Also, for DivX 2pass Insane Quality encodes, it takes half the time of x264 High Profile 1pass with Handbrake.

Also DivX it's very light on the pc for decoding.

TESTED ON 37" 1080 LED TV with vivid color setting and crystal clear image.

BTW I was surprised when I saw a lot of dirty artifacts, from 2m distance from the TV on the x264 High Profile encodes that were ripped from Blu-Ray disc from ~40GB. DivX @ 7500kbps has dirty artifacts that can be seen only from being few inches away from the TV.

x264 High Profile's dirty artifacts can't be seen on my 24" monitor, neither the ones from DivX (as they would probably be very hard to be noticed on 32" TV).

[poisondeathray]

It might be strange, but DivX 6.9 1080p rip @ 7500kbps (2pass encoding and Highest quality setting) is better than x264 rip (with Handbrake's High Profile @ ~13.3Mbps).

Also, for DivX 2pass Insane Quality encodes, it takes half the time of x264 High Profile 1pass with Handbrake.

Also DivX it's very light on the pc for decoding.

TESTED ON 37" 1080 LED TV with vivid color setting and crystal clear image.

BTW I was surprised when I saw a lot of dirty artifacts, from 2m distance from the TV on the x264 High Profile encodes that were ripped from Blu-Ray disc from ~40GB. DivX @ 7500kbps has dirty artifacts that can be seen only from being few inches away from the TV.

x264 High Profile's dirty artifacts can't be seen on my 24" monitor, neither the ones from DivX (as they would probably be very hard to be noticed on 32" TV).


Are you sure you didn't mix them up? DivX @ 7.5Mbps vs. x264 @ 13.3Mbps

There is no contest here . There must be a mistake somewhere

Maybe you're misinterpreting the "dirty artifacts" - are you referring to grain in the original? Because x264 will do a better job at preserving the original quality (this includes grain)

[zmejce]

No, I'm not messing the stuff.

The videos that I've encoded were different for x264 and DivX, however the DivX is better in that case.

I'll probably do tests to the same video when I'll have time and post screen shots.

[Bloax]

The best way to compare codecs is to test them on different sources and compare the results.
Yes, logic is good definitely here absolutely quite yes.

[hello_hello]

TESTED ON 37" 1080 LED TV with vivid color setting and crystal clear image.

BTW I was surprised when I saw a lot of dirty artifacts, from 2m distance from the TV on the x264 High Profile encodes that were ripped from Blu-Ray disc from ~40GB. DivX @ 7500kbps has dirty artifacts that can be seen only from being few inches away from the TV.

x264 High Profile's dirty artifacts can't be seen on my 24" monitor, neither the ones from DivX (as they would probably be very hard to be noticed on 32" TV).

Maybe that's part of the problem.... your TV is trying to do clever things to the picture. If you really want to compare the two using your TV, do so with all the picture enhancing junk disabled. "Dirty artifacts" isn't really descriptive enough to know what it is you're seeing. Ideally you'd want to calibrate the TV properly first too." Vivid color" sounds particularly nasty to me.
I've never seen image enhancements in a TV which I thought improved the picture. I can't imagine why people use them. The first thing I do when setting up a TV is to go into it's settings and disable all of it completely.

Also you didn't say what you're using to play the encodes. ie a media player built into the TV, or a Bluray player etc. I've had fairly limited experience using different types of media players but even so it's fairly obvious they're not created equally. Even the image produced when playing the same encode using the two different Bluray players in this house (same TV) looks slightly different. Neither is bad, just one's a little more pleasing to the eye. If you're comparing encodes using a TV then ideally you'd probably want to be playing them on a PC connected to it, as chances are it'll be the best media player you own anyway (assuming you're using a decent software player).

The fact you say you can't see these "dirty artifacts" when playing the encodes and viewing them on your PC monitor would seem to indicate you need to find out what's messing with the picture when you're viewing them on the TV. Artifacts don't magically disappear and if they're encoded into the video you should be able to see them using your PC monitor too. Often when people post regarding visual problems playing encodes it's a problem with sending the TV the wrong levels or something similar.

[zmejce]

hello_hello vivid color setting isn't an enhancment, its a color profile, it doesn't enhances the picture (besides setting colors).

I have all filters (such as deblocking) disabled, hardware decoding is enabled for both codecs, and it's a very clean image.

But it looks like the more options you enable to x264 the slower the encode and more damaged (processed) the image and using Normal profile wirh cRF:20 isn't enough.

[hello_hello]

hello_hello vivid color setting isn't an enhancment, its a color profile, it doesn't enhances the picture (besides setting colors).

If it's changing the colors then it's changing the way the video displays on the TV, whether it be a color profile or an "enhancement". Chances are it also uses different brightness, contrast and/or gamma settings to make the color "vivid", not by simply adjusting the colors.
To my way of thinking a TV should be calibrated properly (generally the most boring factory preset is probably going to be fairly close) and display the video as it was intended to be displayed, but each to their own.....
My nephew has some vivid color setting enabled on his TV and he's probably played with various other settings too. He thinks it looks great. I can barely stand to be in the same room when his TV is on.

The bottom line is if the video looks good on your PC monitor but bad on your TV then they can't be displaying the video the same way. That doesn't mean the video must be perfect, but they both can't be displaying it "correctly" either.
Mind you it does seem odd that you're only having quality issues when encoding with x264 but not DivX. I don't have a good explanation for that one.

I have all filters (such as deblocking) disabled, hardware decoding is enabled for both codecs, and it's a very clean image.

Using what as your playback device? If it's a PC how is it connected to the TV? Which media player are you using? Probably the number one cause of complaints when playing video on a PC connected to a TV would be using the wrong levels and depending on the player it may correct the levels for some types of video by default, but not others.... the same can even apply to viewing video on a PC monitor. There's lots of variables.

But it looks like the more options you enable to x264 the slower the encode and more damaged (processed) the image and using Normal profile wirh cRF:20 isn't enough.

I don't know exactly what "options" you're referring to and I'm not a Handrake user. Generally the idea would be not to play with any x264 options unless you really know what you're doing. x264 has it's own built in "speed" presets which have been tuned by the x264 developers. Changing the speed preset automatically changes the appropriate advanced x264 options accordingly. For 99.9% of us, using the built in x264 speed presets is the best option.
Unfortunately though, for reasons I don't understand Handbrake is one of the few encoder GUIs which doesn't give you access to the x264 speed presets (at least as far as I remember). If you're fond of Handbrake, maybe try VidCoder as an alternative. It uses the Handbrake "engine" but it's a different GUI. It's very much like Handbrake, but it does also give you access to the x264 speed presets.

If I remember correctly, Handbrake's own "Normal" preset isn't "normal" for x264. It's more like using x264 and a faster x264 speed preset if memory serves me correctly (default is medium). Even most Handbrake devotees will probably tell you not to use it. Handbrake's "High Profile" preset is probably close to x264's default settings with maybe a couple of advanced options tweaked a little. It should give you decent quality. If not, try another encoder GUI. I'd be interested to know if your results then change.

All I can tell you is your "experience" seems contradictory to general opinion, so either you're correct and the rest of the world is wrong, or maybe it's the other way around......
I can tell you I've encoded plenty of video, both HD and SD, using a CRF value of 20 and the default "medium" x264 speed preset (although I tend to use CRF 18 these days) and the majority of the time the encoded version is pretty much indistinguishable from the original when displayed on my 51" plasma (unless I deliberately compare them frame by frame). My TV is literally sitting a couple of feet from my desk, so I've done lots of comparing using the TV when encoding.
Why your experience seems different to the "norm" I don't really know, all I can do is offer possible reasons, but maybe if you do post some screenshots it might help to clear things up.

[zmejce]

I use mpc-hc latest x86 with latest coreavc and divx.

The TV is connected with PC via HDMI2HDMI 1.3b gold plated cable.

Chances are that I'll post screen shots of same rip as soon as possible, but atm I'm not able to make rips for comparison.

In meanwhile here are the screenshots of different videos for comparison. And intentionally I had took screenshots of the places where I can see dirty artifacts at peak.

x264 High Profile (@~13.3Mbps, I haven't touched the settings except mkv instead mp4) encoded with Handbrake 0.9.5 from Blu-Ray 40GB disc:
http://1.0.1.img98.net/out.php/i451124_1999-blue-streak.mkv-snap6.17-.jpg
http://1.0.1.img98.net/out.php/i451125_1999-blue-streak.mkv-snap6.54-.jpg


DivX @ 7500Kbps 1080p 2pass insane quality, encoded from H264 10.2Mbps source:
http://1.0.1.img98.net/out.php/i451123_britney-spears-3-1080i.4.10-.jpg

*notice for the DivX screenshot: The screenshot of the original H264 source looks identical or almost identical (It's already dirty a bit).

And as I wrote previously, You'll probably need around 40" or bigger TV/Monitor to notice the dirty artifacs especially for the DivX screenshot. They are very hard to be noticed on my 24" 1080p monitor (especially the DivX ones). Probably you won't be able to notice the dirty artifacts at around 30" monitor or TV too.

[zmejce]

If it's changing the colors then it's changing the way the video displays on the TV, whether it be a color profile or an "enhancement". Chances are it also uses different brightness, contrast and/or gamma settings to make the color "vivid", not by simply adjusting the colors.
To my way of thinking a TV should be calibrated properly (generally the most boring factory preset is probably going to be fairly close) and display the video as it was intended to be displayed, but each to their own.....


If I play the Blu-Ray disc on the TV with vivid color profile, definitely it looks better than if I play it with Normal color profile.

Shouldn't the rips do the same?

[kypec]

Rule #1: compare the screenshots of different codecs with the same scene/frame and make sure you encode with each codec from the same source material - be it Bluray disc or HDTV recording, otherwise it's totally useless.
Your "dirty artifacts" are in fact the grain & noise that is present in the original BD source. x264 has done excellent job in keeping it there in the encoded H264 stream. I can understand that DivX encode is more pleasing to your eyes, because it's been heavily denoised so everything looks so clean, smooth... simply too unnatural. If you want to achieve the same look also for x264 encodes then (over)denoise your source too, you can use MCTemporalDenoise or just --nr 100 noise reduction built-in filter.

[detmek]

1. Post the same frame from source, Handbrake encode and Divx encode. You/we can not compare different images.
2. Do not use JPEG to take screenshots as JPEG is lossy compression with its own artifacts. Use PNG.

About images you posted. Frames from H.264 encode have a lot of film grain which is very hard to encode. Divx image is very smooth, blured. You don't have any fine details. Just compare hair and eyebrows from first H.264 image and Divx image. So, for me, H.264 images look much better.

Too slow.

[zmejce]

I had wrote that at the moment I'm not able to make DivX rip of the same movie, I'll probably make it as soon as possible.

Here's another DivX@ 13.3Mbps 2pass, from 35Mbps mpeg2 source
Divx@13.3Mbps (http://1.0.1.img98.net/out.php/i451411_13.3mbps.avi-snapshot-15.3.59-.jpg)

On my screen it looks dirty only from few inches away from the screen, unlike the x264 screenshots that look dirty from 2m away from the screen. Also nothing can be seen at 24" monitor.

[Groucho2004]

On my screen it looks dirty only from few inches away from the screen, unlike the x264 screenshots that look dirty from 2m away from the screen. Also nothing can be seen at 24" monitor.

The first x264 screen you posted is actually pretty good and has proper grain retention. The divx screen you just posted has atrocious artifacts.
You call them both "dirty". It makes no sense.

You were told to post lossless screen shots since we have no idea if the artefacts are from extreme jpg compression or really, really bad encoding or a really, really bad source.

Posting a screenshot is useless when there is no point of reference.

[zmejce]

No, they look acctually the same as their video source.

A screen shot from original source is same as those screen shots, just 100% clean.

[zmejce]

The first x264 screen you posted is actually pretty good and has proper grain retention. The divx screen you just posted has atrocious artifacts.

The bottom line is that for both codecs I'll have to raise the birate to get the result I want.

I'll go with DivX.

[kypec]

The bottom line is that for both codecs I'll have to raise the birate to get the result I want.

I'll go with DivX.
No, the true bottom line is that you don't follow the advices of many experienced users though you've been told them several times already::mad:

Post screenshots in lossless format PNG, not JPG which is lossy itself!
Always use the same sources for direct encoder comparisons, not different scenes for every screenshot you want to point out!
Post your exact settings you encoded your video with, input sample files & output log files are more than welcome!

I've personally switched from xvid to x264 more than 2 years ago and I would never go back, the visual quality of x264 beats xvid with much better compression for all my sources, especially cartoons that were somewhat denoised as well.:cool:

[detmek]

OP just likes blurry images so he prefers Divx.
x264/H.264-AVC is more efficient encoder/standard compared to Divx/MPEG4-ASP so it requires lower bitrate to produce video of the same quality. Here is a link to MSU Video Codec Comparisson (http://compression.ru/video/codec_comparison/h264_2012/) page that shows efficiency of Xvid, x264 and Divx H.264.
Note, it's Divx H.264, not Divx ASP. And Divx H.264 gives better quality compared to Divx ASP.

[zmejce]

Handbrake doesnt deinterlaces 1080i correctly, at least some videos have to be encoded in vdub.

Also Im quite sure that the manual way of encoding with vdub is better than auto-way of encoding (ex. handbrake).

I have troubles to encode to x264 with vdub, otherwise I would probably give it a shot. DivX with vdub has no problems.

I had found that DivX is faster & better than XviD at any birate Ive tried.

Also in MPC-HC Ive set shaders: denoise & shapren.

The video that is encoded with DivX 17Mbps (theres a screenshot without filters above in reply) looks almost like a lossless Blu-ray rip with them, its even better than the original mpeg2 36Mbps source which looks pretty good.

[detmek]

If you do not apply filters encoded video can't look better then original. And VirtualDub 1.10 has a support for external encoders so you can use any CLI encoder you want: x264, ffmpeg, neroaac, qaac, lame and merge streams with mkvmerge, ffmpeg, mp4box...

[zmejce]

I know how to merge streams, just Im not familiar with setting vdub to use commandline, neither to set x264 commands.

[Bloax]

x264 Is definitely better than DivX/XviD.

Here you go (http://www.mediafire.com/?vdce9f45979f8df), x264 vs. XviD on a quite tough source.

Unlike with x264, I couldn't tweak the settings to all hell to make them look good for this one source with XviD. But the speeds are 15 fps for x264, and 23 for XviD. (So it's not like I used "fast" settings for it.)

[zmejce]

Yep, x264 is better than DivX even with lower birate.

x264 @ 13.3Mbps with the MPC-HC filters it looks almost like the Blu-Ray.

Which commands I should use for the command-line x264 and vdub?

I saw this from Handbrake 0.9.5 High Profile:
ref=2:bframes=2:subq=6:mixed-refs=0:weightb=0:8x8dct=0:trellis=0

I don't want to use Psychovisual or such effects, and I want the file to be at about 15Mbps.

Also when I set command-line encoder in vdub, is the procedure the same (File->Save as->avi)?

If not, then how?

[zmejce]

Bloax can you share your x264 command-line settings?

[Bloax]

I unfortunately have no idea how to use the external encoders in VDub, I find that using JEEB's builds (which include audio encoding, yay) in a command line is much easier. :p

My commandline is quite biased for the game I'm playing there (which is just a bit noisy as you might've noted).

But in a nutshell;
--preset Slow --tune Grain --bframes 16 --aq-strength x --nr y

--aq-strength Kind of determines what kind of content to focus the quality on. Values lower than 1 focus on noise (I actually used 0.4 on my clip there), higher focus more on flat areas. (Not so surprisingly all but the StillImage tuning tweak it downwards.)

And --nr is for the in-loop denoiser, which usually actually speeds up the encoding while reducing the final size. (The value you should use kind of depends on the noisiness of the source. Just mess around with values around 100-500 I guess.)

Of course I also did some rather pointless other things, but that's mostly because it's what it is.
Don't forget to use tunings though, they're good stuff. (--tune Grain is a great tuning, though it also increases the size rather dramatically.)

Disabling the psychovisual stuff going on isn't really a good idea.

I can't really say how you'd hit 15 Mbps, other than taking some time to mess with the --crf values.
You'll learn it eventually anyways.

[zmejce]

Well cRF:20 high preset produces ~13.3Mbps.

I don't know what would be for around 15Mbps?

[Bloax]

Crf 19.25 would probably do. ('Tis why you mess around with it.)

[Groucho2004]

Crf 19.25 would probably do.
Funny. I call your 19.25 and raise you 0.03. :cool:

[detmek]

Yep, x264 is better than DivX even with lower birate.

x264 @ 13.3Mbps with the MPC-HC filters it looks almost like the Blu-Ray.

Which commands I should use for the command-line x264 and vdub?

I saw this from Handbrake 0.9.5 High Profile:
ref=2:bframes=2:subq=6:mixed-refs=0:weightb=0:8x8dct=0:trellis=0

I don't want to use Psychovisual or such effects, and I want the file to be at about 15Mbps.

Also when I set command-line encoder in vdub, is the procedure the same (File->Save as->avi)?

If not, then how?
1. In VirtualDub folder create new folder "External encoders" and put all exe files and dependencies in that folder.
2. In Options - External encoders set an encoders, something like this:
For x264:
Command arguments: --crf 20 --tune film --b-adapt 2 --direct auto --demuxer raw --input-csp i420 --input-res %(width)x%(height) --fps %(fpsnum)/%(fpsden) -o "%(tempvideofile)" -
Output filename: %(outputname).264

For neroaac:
Command arguments: -q 0.4 -ignorelength -if - -of "%(tempaudiofile)"
Output filename: %(outputname).m4a

For mkvmerge:
Command arguments: -o "%(outputname)" --default-duration 0:%(fpsnum)/%(fpsden)fps "%(tempvideofile)" "%(tempaudiofile)"
and create encoder sets by combining video, audio and muxer presets.
3. You start encoding with File - Export - Using external encoder.

[zmejce]

Thank you, I'll try that :)

[zmejce]

detmek, do you recommend the same x264 settings for dvd-ripping?

or they are just for hd?

I'm asking because when I'm ripping DVD's with handbrake and High profile x264, it produces a very smooth image (DivX is better in that case).

[detmek]

For DVD rip I would go with CRF 18.
But, I don't like smooth images. I prefer to preserve fine details so Divx/Xvid is not my choice.
If you want smoother image remove --tune film. You could use some denoiser/smoother in VirtualDub or add --nr xxx (where xxx is in 100-1000 range) to use x264 internal denoiser. Changing deblock to 1:1 or 2:2 will also smooth image.
Try and see what you like.

[zmejce]

Ok, thanks, I'll try it these days.

[zmejce]

detmek, and which settings for dvd's that are already ripped to divx 2000kbps (they have excellent quality)?

[Groucho2004]

detmek, and which settings for dvd's that are already ripped to divx 2000kbps (they have excellent quality)?
You should really use your original DVDs.

[Bloax]

If you don't want a smooth image, you should probably try --tune Grain, which both adjusts AQ to prefer noise, and reduces the deblocking strength.