Is Handbrake 1.3.3 Slower?

[Shibblet]

Description of problem or question:
It seems as if Handbrake 1.3.3 is running slower than it's predecessor.

Steps to reproduce the problem (If Applicable):
I'd like to download an older version of Handbrake to test if this is the case.

HandBrake version (e.g., 1.0.0):
1.3.3

Operating system and version (e.g., Ubuntu 16.04 LTS, macOS 10.13 High Sierra, Windows 10 Creators Update):
Same problem on both Windows 10 2004 19041.508 / Kubuntu 20.04.1

I have been encoding 1080p content with x264 for quite some time using Handbrake. Usually my encodes (with my basic settings, no filters, audio passthrough, 24q on Very Slow, Frame Rate Same as Source) would get around 12-15fps on a Core i7-3770. It usually took "Double Time" to encode a movie. i.e. 4 hours to encode a 2 hour movie. I am fine with this speed, and accept it. I get the best quality and the smallest file.

Now, I am getting around 6-10 fps. The average is around 7 fps. It's taking more than "Triple Time" to encode a movie. And the only difference I can think of, is updating to 1.3.3.

I didn't think anything of this when I updated the program to 1.3.3. A lot of the times I would start a couple of encodes, and then go to bed. But lately, I have noticed that I am getting up in the morning, and the encoding is not done yet. It still has a couple of hours to go. This is what prompted me to look into this problem.

I'd like to run 1.2.1 again to see if this situation is improved, but not really sure where to find an older copy.

Has anyone else noticed any performance issues with 1.3.3?

[mduell]

Same settings performs the same, within the margin of error.

Probably a change in settings, including possibly a change in defaults (although none come to mind for x264).

Post your logs from 1.2.1 and 1.3.3 (they're all saved to disk, as documented) and we can point out the differences.

Several people have made the same claim, and it always came down to different settings they chose.

[Shibblet]

Same settings performs the same, within the margin of error.
Probably a change in settings, including possibly a change in defaults (although none come to mind for x264).

I haven't changed any of my settings. Were the implementation of them changed between versions?

Post your logs from 1.2.1 and 1.3.3 (they're all saved to disk, as documented) and we can point out the differences.

I don't have any 1.2.1 logs... I upgraded to 1.3.3 a long time ago, and have just been using it anyway. I wasn't sure this was an issue until recently.

Several people have made the same claim, and it always came down to different settings they chose.

Any idea which settings those would be so that I can check? I searched the forums for "slow" and "slower," but those return VERY different results for x264.

[Shibblet]

Oh... I forgot to mention: In the preferences menu, there is an option for using QSV for decoding.
I have turned this on and off, and it doesn't seem to do a thing.

[Deleted User 11865]

https://handbrake.fr/old.php

[mduell]

Same settings performs the same, within the margin of error.
Probably a change in settings, including possibly a change in defaults (although none come to mind for x264).

I haven't changed any of my settings.

Sure, that's what everyone says. Anyway, you have the link for old versions so you can post both logs instead of engaging wild speculation.

[Shibblet]

Sure, that's what everyone says. Anyway, you have the link for old versions so you can post both logs instead of engaging wild speculation.

Have I done something to offend you?

I came to this forum and asked a legitimate question. However, instead of addressing that it could be an issue, you called my claim "wild speculation."
Far be it from me to have asked a question that may result in the answer of "Yep, we're aware of that problem, and are working on a solution." Which seems to be the case with a lot different types of software.

Instead, you seem to have taken it as some kind of insult. I mean... is this how questions on this forum are handled? "Sure, that's what everyone says?" As in complete disbelief? The only validation I have, is that others have asked the same because apparently that's "what everyone says..."

Being dismissive and confrontational are the absolute opposite of being helpful.

Regardless. I plan on checking out 1.2.1, or earlier versions now that I know they are available for download from a reputable source. And I will come back at a later time with my findings.

[Woodstock]

Presets have been evolving over time. As machines get faster, handbrake has been moving from "very fast presets" to "better quality presets" by default. That makes it harder to benchmark different versions via the GUI.

To compare on equal terms, you need to use the same source and the command line version, so you can make sure ALMOST everything is the same.

[mduell]

I came to this forum and asked a legitimate question. However, instead of addressing that it could be an issue, you called my claim "wild speculation."

We've seen this post before, again and again. Spoiler: it's not HB getting slower with the same settings.

Far be it from me to have asked a question that may result in the answer of "Yep, we're aware of that problem, and are working on a solution." Which seems to be the case with a lot different types of software.

Presumably you followed best practices and searched the open issues and requests before posting? So that would be an unlikely response.

[Shibblet]

We've seen this post before, again and again. Spoiler: it's not HB getting slower with the same settings.

Ultimately, you are right. Handbrake did not get any slower. However, none of the settings were changed... So, what's the dealio-yo?

A friend of mine explained what the problem was when we were talking about it on the phone last night. What it came down to was the content of what I have been encoding as of late. I have been ripping and encoding my Blu-Ray complete series of "Battlestar Galactica (2006). This show has a lot of film grain in it. Apparently the more variation in the picture, the longer it takes per frame to encode. The movie "300" takes longer to encode at the same settings as say "The Incredibles." This has nothing to do with Handbrake itself, it's more to do with the variation of pixel-to-pixel. Film grain varies that quite a bit.

Presumably you followed best practices and searched the open issues and requests before posting? So that would be an unlikely response.

Nope. Can I not ask questions in this forum? That's all that I did. These responses are acting as if I made false claims and accusations. I did not. I simply said "It seems like..." It's very hard to ask the right questions where you are not even sure of the problem.

[Woodstock]

Ah, grain and static... you can invoke NLMeans to clean it up substantially, making for smaller files, but... That will make the perceived slowness even worse.

Are you using batch processing so you don't have to wait on things while handbrake chews its way through a set of encodes?

[Shibblet]

Ah, grain and static... you can invoke NLMeans to clean it up substantially, making for smaller files, but... That will make the perceived slowness even worse.

Are you using batch processing so you don't have to wait on things while handbrake chews its way through a set of encodes?

Oh heck yeah. I set up about 4 episodes in the queue, then go to bed. The previous Blu-Ray collection I had encoded was Big-Bang Theory. I would get around 12-15 fps on each episode... but those were all filmed digital and have a super clean picture. Battlestar Galactica has a LOT of film grain

I don't use any filters, because I want the video to look exactly like it does from the Blu-Ray.

I may play around with the filters and see what it actually does, one of these days.

[Woodstock]

I just completed a 150+ file batch over the weekend (Saturday through Tuesday morning), and tried several versions of handbrake in it (advantage of using the CLI, just change the path), and there wasn't a significant difference in reported FPS between 0.10.3, 1.12.x, and 1.3.3, the main difference being in whether I was processing 1080p->1080p or 1080p->720p.

(I also have a "medium capability" machine that I can dedicate to the process)

[Shibblet]

I just completed a 150+ file batch over the weekend (Saturday through Tuesday morning), and tried several versions of handbrake in it (advantage of using the CLI, just change the path), and there wasn't a significant difference in reported FPS between 0.10.3, 1.12.x, and 1.3.3, the main difference being in whether I was processing 1080p->1080p or 1080p->720p.

(I also have a "medium capability" machine that I can dedicate to the process)

Good to hear. I have noticed my encodes are a lot faster if I down-sample to 720p. But the whole idea is to keep 1080p.

I'm still not yet convinced that 4K is worth jumping up to, and by the same concept, I think 8K beyond overkill.

A 4K TV makes a big difference. The pixel density gives a lot of clarity to the image. But on that 4K TV, I cannot see any difference between 4K content and 1080p content when I watch it on a 4K TV. Even on a larger TV, sitting at "reasonable" distance in the living room. 10' - 15' away from a 65"
But that's just me. Some people claim they can see a major difference in the source resolution, but not me. I'm good with 1080p.

I used to work at a Home Audio/Video store. I set up our "Theater Room." The boss gave me the "all out" when it came to setting it up too, so used the high end Klipsch Reference Series speakers, and a High End Marantz Receiver. The receiver was for the source, and was used to power the 2 of the 4 Surrounds and the 2 Atmos (above) speakers. A secondary 5 channel Marantz Amp was used to power the Front Left, Center, and Front Right, and the other 2 surrounds. Anyway... amazing theater setup.

After I had it all set-up, I started playing with some Blu-Ray's to get it tuned in. I switched between the Dolby True HD, and Dolby Digital tracks, and could not hear much of a difference. A bit, but not much. Then I started switching between the DTS-HD MA, and DTS tracks on a different Blu-Ray, and this time, could not hear any discernible difference AT ALL. Tried it at lower volumes, higher volumes, switching the amps to different speakers, tuning the receiver, etc. etc. etc. Absolutely NO DIFFERENCE AT ALL.

So, with my encodes, I only use the basic Dolby Digital or DTS audio tracks. And I used x264 for 1080p video. I encode at Very Slow and a CQ of 24. Most movies I encode are anywhere between 2 to 4 gigs with the Dolby or DTS audio tracks, and look and sound just like the original Blu-Ray on my system.

[mduell]

Glad you figured it out.