Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > Video Encoding > MPEG-4 Encoder GUIs

Reply
 
Thread Tools Search this Thread Display Modes
Old 5th December 2022, 10:23   #19801  |  Link
slalom
Registered User
 
slalom's Avatar
 
Join Date: Jan 2010
Posts: 435
Quote:
Originally Posted by Ryushin View Post
Thank Atak for the update and the new features. I appreciate it.

If you have time would be nice to have SMDegrain replace MDegrain. The results are just staggering and I resisted doing it for so long because I thought MDegrain did such a good job it could not really improve that much with SMDegrain. SMDegrain's output is so much improved it's staggering. It's like the difference MDegrain made over HQDN3D and that I would never go back to HQDN3D after seeing the results for MDegrain. Now SMDegrain has made MDegrain obsolete.

Now I'm at the point where the output is so clean with SMDegrain I think something must be wrong as the file sizes are so much smaller and I must have lost detail. Every detailed comparison I've made shows that I was wrong and the output from SMDegrain has not only removed the grain, but since the grain is removed, details actually stand out and everything appears crisper. I've actually started re-encoding older films and watching them again such as The Terminator, The Edge, etc. just because the difference in detail is so dramatic it is like experiencing the film all over again and it is a joy to watch.

SMDegrain also does not produce the detail blur I saw with MDegrain3 in a characters face when they turned their head.

The only tweak I have started doing is blksize=8 when dealing with video that has a lot of dark scenes in it.

Please replace MDegrain with the following SMDegrain strength settings:
Light: SMDegrain(video,tr=2,thSAD=200,thSADC=100,contrasharp=true,prefilter=0,refinemotion=true)
Medium: SMDegrain(video,tr=4,thSAD=400,thSADC=200,contrasharp=true,prefilter=1,refinemotion=true)
Strong: SMDegrain(video,tr=8,thSAD=800,thSADC=400,contrasharp=true,prefilter=2,refinemotion=true)

Thanks again Atak for everything.
What about the speed?
What's the difference in speed between SMDegrain & MDegrain at those 3 settings you suggest?
__________________
E5 2697 v2 @ 3.0GHz on P9X79 Deluxe 24GB
Xeon E5-2680 v2 @ 3.1GHz 16GB
Sony Vaio VPC-F13Z1E/B

Last edited by slalom; 5th December 2022 at 11:50.
slalom is offline   Reply With Quote
Old 5th December 2022, 12:35   #19802  |  Link
TDS
Formally known as .......
 
TDS's Avatar
 
Join Date: Sep 2021
Location: On a need to know basis.
Posts: 543
Quote:
Originally Posted by slalom View Post
What about the speed?
What's the difference in speed between SMDegrain & MDegrain at those 3 settings you suggest?
Hi slalom,

https://forum.doom9.org/showthread.p...20#post1979320
__________________
This can be SO "TeDiouS"..
Long term RipBot264 user.

#1 Ryzen 7950X
#2 Intel i9-13900KF
#3 Ryzen 5950X
#4 Ryzen 5900X
#5 Ryzen 3950X
TDS is online now   Reply With Quote
Old 5th December 2022, 13:09   #19803  |  Link
cypher007
Registered User
 
Join Date: Jul 2003
Posts: 98
Quote:
Originally Posted by TDS View Post
Come on cypher, I think it's time you gave up on this...

I have a 5.2kW solar array, and my power is approx $0.38 per kwh, but my "rule", if the solar isn't generating enough power so I can encode for "free", I don't do any, and with the weather in my area being really crappy for months & months, I don't get much done , and time's running out.

I would think that if there were other users that use this function, you would have had some replies, but in my way of thinking, as you haven't, no body uses it !!!!

If Atak would explain what it does to encodes in progress, then it's probably not a really economical process, as you would be losing encodes, everytime it shuts down !!!

As rlev11 suggested, maybe you should set up a Windows Task to shut down & wake up the laptop, but then that wouldn't restart the interrupted encode, would it
if i wasnt on such crap money i would just let it run, but im now looking at using works pc to do the encodes. the problem i have is a typical 2hr movie takes about 24hrs on my i5 10th gen.
cypher007 is offline   Reply With Quote
Old 5th December 2022, 15:39   #19804  |  Link
Ryushin
Registered User
 
Join Date: Mar 2011
Posts: 396
Quote:
Originally Posted by slalom View Post
What about the speed?
What's the difference in speed between SMDegrain & MDegrain at those 3 settings you suggest?
Well, thing is, SMDegrain and MDegrain are so different from one another the speed should not be comparable. SMDegrain calls a lot more functions and I would think it would be vastly slower, but it's not. SMDegrain is comparable in speed to MDegrain. All depends on functions used.

Generally, and I don't know why, SMDegrain Light and Medium is faster than equivalent MDegrain. While SMDegrain Hard is about 20-30% slower than MDegrain. SMDegrain gives vastly improved visuals over MDegrain. Just as MDegrain gave huge improved visuals over HQDN3D.

Edit: blksize=8 drops the speed down by half, which is okay as it's processing four times as many blocks. I use this to prevent "blotchy" in content that the background is almost dark, or having two similar colors, such as in Untouchables, where Sean Connery is wearing a navy blue uniform at night on the bridge. Though I have to zoom up on the image to notice it.

Last edited by Ryushin; 5th December 2022 at 15:46. Reason: More Information
Ryushin is offline   Reply With Quote
Old 5th December 2022, 17:10   #19805  |  Link
Atak_Snajpera
RipBot264 author
 
Atak_Snajpera's Avatar
 
Join Date: May 2006
Location: Poland
Posts: 7,716
Quote:
Originally Posted by cypher007 View Post
if i wasnt on such crap money i would just let it run, but im now looking at using works pc to do the encodes. the problem i have is a typical 2hr movie takes about 24hrs on my i5 10th gen.
Can you check if WAKE function even works on your PC?

Try this
1) open Task Scheduler
2) create new task

3) Type WorkHours as name

4) Click triggers tap -> new -> specify time

5) Actions -> start program -> rundll32.exe

6) Conditions -> Wake the computer ...


Now suspend your PC.
Atak_Snajpera is offline   Reply With Quote
Old 8th December 2022, 02:31   #19806  |  Link
Pauly Dunne
Grumpy Old Man.
 
Pauly Dunne's Avatar
 
Join Date: Jul 2019
Location: Out There....
Posts: 675
update package for RipBot264

This is targeted at the standard version(s)...

This will update the main components of the program, and nothing else.

https://www.mediafire.com/file/x110q...Bot264.7z/file

please read the two included .txt files.

Stay tuned for updated & revised PD builds, coming soon..
__________________
Not poorly done, just doin' it my way !!!
Live every day like it is your last,
because one day, it will be !! (M$B)
Pauly Dunne is offline   Reply With Quote
Old 8th December 2022, 12:50   #19807  |  Link
Ryushin
Registered User
 
Join Date: Mar 2011
Posts: 396
Ryzen 7950x Threads and Tuning

So I got the Ryzen 7950x installed in the computer on Monday night I've been testing RB against it using many different threading options which I'll post at the end of the message.

Instead of reading through all the options, my threading options (for me) are as follows:
Two Encoding Servers
Set x264/x265 threads to 16: /x265-threads 16 /x264-threads 16 (RB default is 16 for the 7950x)
Set Avisynth Threads to 12: (In settings, use multiple processing threads) (RB default is 16 for the 7950x)

Last night I down clocked the Ryzen 7950x to all cores at 4700Mhz based on this information:
https://www.youtube.com/watch?v=7JiYAwKIHRY
My silicon is probably not as good as the one in the video or my motherboard and memory (running at 6000Mhz) won't me reach the same as the video. Though I might just have to do some more tuning. Right now I'm giving up 7% of the processors power to save more than double the power savings. AMD (and Intel for that matter) are really chasing every last bit of performance no matter the power costs.

Stock Power Consumption and Temperature:
Idle: 38 watts, 46 Celsius
Full Load:175 watts, 94.5 Celsius

Underclocked Power Consumption and Temperature:
Clock Speed: 4700 Mhz, Voltage: 1.005 volts
Idle: 8 watts, 40 Celsius,
Full Load: 88 watts, 58 Celsius

So essentially, in my case, I loose 7% of the maximum performance of the CPU, to save double the power savings at full load and five times the power savings at idle. I also noticed, at default power settings, full load had half of the cores at 5.1Ghz, while other half was at 4.8Ghz. I have not done any RB thread testing under the Underclocked settings yet as I still want to do some more tweaking and I'm hoping to hit that 4.8Ghz at 0.985 volts, the same as the video. I'm only encouraged going down this route.

BTW, the 7950x requires a 360mm or larger AIO (All In One) Cooler to stay within it's thermal limits. I chose the Fractal Design Celsius S36 360mm for my system, based on cost. But a 420mm cooler would be best:
https://www.youtube.com/watch?v=8kckURYqTKI

I had to buy a Phanteks Enthoo Primo PH-ES813P_SWT case to hold my four optical drives and provide enough room to go up to 420mm cooler. I still chose the 360mm based on cost. I thought the 360mm would be plenty. But there is still headroom that the 420mm gives you if you are going to run it at full tilt. With me underclocking the CPU, the 360mm is more than adequate. But still the 420 is the best option.

Here are all my RipBot thread tests and notes I wrote during testing:
RipBot264 default threads for Ryzen 7950x:
x264/x265 threads: 16
Avisynth Threads: 16
Encoding used for all tests: CQ18

4K Tests:
Source 3 Minutes: Top Gun Maverick 4K with changing aspect ratios (7:00 to 10:00)
No Degraining
One Encoding Server
x265 16 Threads: 5m:32s (13.01 fps)
x265 24 Threads: 4m:20s (16.62 fps)
x265 32 Threads: 4m:10s (17.28 fps)

Source 6 Minutes: Top Gun Maverick 4K with changing aspect ratios (7:00 to 13:00)
No Degraining
Two Encoding Servers
x265 16 Threads: 8m:7s (17.74 fps)


Source 3 Minutes: Top Gun Maverick 4K with changing aspect ratios (7:00 to 10:00)
SMDegrain Medium Setting: SMDegrain(video,tr=4,thSAD=400,thSADC=200,contrasharp=true,prefilter=1,refinemotion=true)
x264/x265 Threads: 16

Single Encoding Server and single chunk.
06 Avisynth Threads: 8m:04s (8.92 fps)
07 Avisynth Threads: 7m:20s (9.82 fps)
08 Avisynth Threads: 6m:45s (10.67 fps)
10 Avisynth Threads: 6m:46s (10.64 fps)
12 Avisynth Threads: 6m:36s (10.90 fps)
14 Avisynth Threads: 6m:33s (10.99 fps)
16 Avisynth Threads: 12m:30s (5.73 fps)
Noticed the CPU usage had been climbing from 50% for 8 threads to 100% for 16 threads. 16 threads also showed fluctuations from 30-100%. Theory is some threads are waiting on other threads to finish processing before they can continue.


Source 6 Minutes: Top Gun Maverick 4K with changing aspect ratios (7:00 to 13:00)
SMDegrain Medium Setting: SMDegrain(video,tr=4,thSAD=400,thSADC=200,contrasharp=true,prefilter=1,refinemotion=true)
x264/x265 Threads: 16

Two Encoding Servers and two chunks (3 minutes each).
06 Avisynth Threads: 12m:09s (11.85 fps)
07 Avisynth Threads: 12m:09s (11.85 fps)
08 Avisynth Threads: 12m:13s (11.79 fps)
10 Avisynth Threads: 12m:55s (11.15 fps)
12 Avisynth Threads: 12m:51s (11.21 fps)
14 Avisynth Threads: 13m:15s (10.87 fps)
16 Avisynth Threads: 14m:37s (9.85 fps)
Not much of an improvement compared to single threads, even starting from 6 threads.


Source 3 Minutes: Top Gun Maverick 4K with changing aspect ratios (7:00 to 10:00)
SMDegrain Medium Setting: SMDegrain(video,tr=4,thSAD=400,thSADC=200,contrasharp=true,prefilter=1,refinemotion=true)
Set x264/x265 threads to 32: /x265-threads 32 /x264-threads 32

Single Encoding Server and single chunk.
08 Avisynth Threads: 7m:45s (9.29 fps)
10 Avisynth Threads: 7m:04s (10.19 fps)
12 Avisynth Threads: 6m:50s (10.54 fps)
13 Avisynth Threads: 7m:00s (10.29 fps)
14 Avisynth Threads: 6m:53s (10.56 fps)
16 Avisynth Threads: 10m:45s (6.70 fps)


Source 3 Minutes: Top Gun Maverick 4K with changing aspect ratios (7:00 to 10:00)
SMDegrain Medium Setting: SMDegrain(video,tr=4,thSAD=400,thSADC=200,contrasharp=true,prefilter=1,refinemotion=true)
Set x264/x265 threads to 24: /x265-threads 24 /x264-threads 24

Single Encoding Server and single chunk.
08 Avisynth Threads: 7m:35s (9.49 fps)
10 Avisynth Threads: 6m:57s (10.40 fps)
12 Avisynth Threads: 6m:40s (10.80 fps)
14 Avisynth Threads: 6m:25s (11.22 fps)
16 Avisynth Threads: 10m:56s (6.59 fps)


Summary Best 4K Results:
Two Encoding Servers with 16 x264/x265 threads and 6-8 Avisynth threads. Note using the RipBot default of 16 avisynth threads is the slowest with all tests.



1080P Tests:
Source 6 Minutes: Cast Away HD (23:07 to 29:17)
No Degraining
Two Encoding Servers
x265 16 Threads: 3m:29s (42.49 fps)


Source 6 Minutes: Cast Away HD (23:07 to 29:17)
SMDegrain Hard Setting: SMDegrain(video,tr=8,thSAD=800,thSADC=400,contrasharp=true,prefilter=2,refinemotion=true)
Set x264/x265 threads to 16: /x265-threads 16 /x264-threads 16

Two Encoding Servers and two chunks (3m:0s first, 3m:10s second).
06 Avisynth Threads: 11m:51s (12.49 fps)
08 Avisynth Threads: 10m:05s (14.68 fps)
10 Avisynth Threads: 9m:58s (14.85 fps)
12 Avisynth Threads: 8m:50s (16.75 fps)
14 Avisynth Threads: 8m:48s (16.82 fps)
16 Avisynth Threads: 8m:56s (16.57 fps)


Atak's x265 Benchmark Tool:
Ryzen 1800x: encoded 2500 frames in 108.78s (22.98 fps), 7025.74 kbps, Avg QP:37.21
Ryzen 7950x: encoded 2500 frames in 21.74s (115.00 fps), 7025.74 kbps, Avg QP:37.21

Last edited by Ryushin; 8th December 2022 at 13:19. Reason: Fixed CPU to 7950x
Ryushin is offline   Reply With Quote
Old 8th December 2022, 13:10   #19808  |  Link
TDS
Formally known as .......
 
TDS's Avatar
 
Join Date: Sep 2021
Location: On a need to know basis.
Posts: 543
Quote:
Originally Posted by Ryushin View Post
So I got the Ryzen 9750x installed in the computer on Monday night
Hi Ryushin,

So there is SO much info here, and tbh, I don't understand much of it.

But I have to confess, I have not used my 7950X yet

But these Avisynth Thread settings & x264/x265 Threads settings, are they the new option in the Encoder command line ??

I have been using the Affinity settings on the 3950X & 5950X, that rlev11 documented a while back.

So, I'm going to have to look into this, soon, I guess.
__________________
This can be SO "TeDiouS"..
Long term RipBot264 user.

#1 Ryzen 7950X
#2 Intel i9-13900KF
#3 Ryzen 5950X
#4 Ryzen 5900X
#5 Ryzen 3950X

Last edited by TDS; 8th December 2022 at 13:21.
TDS is online now   Reply With Quote
Old 8th December 2022, 13:25   #19809  |  Link
Ryushin
Registered User
 
Join Date: Mar 2011
Posts: 396
Quote:
Originally Posted by TDS View Post
Hi Ryushin,

First of all mate, you need to stop calling it a 9750X.

So there is SO much info here, and tbh, I don't understand much of it.

But I have to confess, I have not used my 7950X yet

But these Avisynth Thread settings & x264/x265 Threads settings, are they the new option in the Encoder command line ??

I have been using the Affinity settings on the 3950X & 5950X, that rlev11 documented a while back.

So, I'm going to have to look into this, soon, I guess.
Fixed the 9750x to 7950x in my post. Not sure why I kept saying that. LOL

I'm not using any of the affinity settings, just the thread settings. So the x264/x265 threads settings is in the Encoding server command line options.

Essentially, using the default RB Avisynth threads of 16 for the 7950x has the FPS fall off a cliff with 4K content. Dialing it back to 12 gives the best overall performance with 4K and 1080p content.

The underclocking might be of serious interest for those running solar or off grid.
Ryushin is offline   Reply With Quote
Old 8th December 2022, 14:04   #19810  |  Link
TDS
Formally known as .......
 
TDS's Avatar
 
Join Date: Sep 2021
Location: On a need to know basis.
Posts: 543
Quote:
Originally Posted by Ryushin View Post
Fixed the 9750x to 7950x in my post. Not sure why I kept saying that. LOL

I'm not using any of the affinity settings, just the thread settings. So the x264/x265 threads settings is in the Encoding server command line options.

Essentially, using the default RB Avisynth threads of 16 for the 7950x has the FPS fall off a cliff with 4K content. Dialing it back to 12 gives the best overall performance with 4K and 1080p content.

The underclocking might be of serious interest for those running solar or off grid.
You must have quoted my post while I edited mine...

Have you read my 7950X post ??

https://forum.doom9.org/showthread.p...40#post1979340

So the Affinity settings do the same thing, setting it back to 12 cores, but I guess if you want to do it that way, that's your choice

So you must be using this setting :-

Code:
Server1CommandLine=/port 1000 /priority normal /restart-if-no-progress /avisynth-prefetch-threads 12
And with the settings on the Main page :-

Use multiple processing threads..."0" ??

Limit to following filters only (checked or unchecked) ??
__________________
This can be SO "TeDiouS"..
Long term RipBot264 user.

#1 Ryzen 7950X
#2 Intel i9-13900KF
#3 Ryzen 5950X
#4 Ryzen 5900X
#5 Ryzen 3950X
TDS is online now   Reply With Quote
Old 8th December 2022, 16:04   #19811  |  Link
Ryushin
Registered User
 
Join Date: Mar 2011
Posts: 396
Quote:
Originally Posted by TDS View Post
You must have quoted my post while I edited mine...

Have you read my 7950X post ??

https://forum.doom9.org/showthread.p...40#post1979340

So the Affinity settings do the same thing, setting it back to 12 cores, but I guess if you want to do it that way, that's your choice

So you must be using this setting :-

Code:
Server1CommandLine=/port 1000 /priority normal /restart-if-no-progress /avisynth-prefetch-threads 12
And with the settings on the Main page :-

Use multiple processing threads..."0" ??

Limit to following filters only (checked or unchecked) ??
Atak fixed that bug, so setting Use multiple processing threads to 12 sets the avisynth-prefetch-threads 12 in the Encoding Servers Wrench.

Setting the encoder threads this way I have 16 x264/x265 threads and 12 avisynth threads. I think this gave me the best results. Though I did not test limiting to 12 threads for x264/x265 as that that seems to like threading.
Ryushin is offline   Reply With Quote
Old 9th December 2022, 04:29   #19812  |  Link
TDS
Formally known as .......
 
TDS's Avatar
 
Join Date: Sep 2021
Location: On a need to know basis.
Posts: 543
Quote:
Originally Posted by Ryushin View Post
Atak fixed that bug, so setting Use multiple processing threads to 12 sets the avisynth-prefetch-threads 12 in the Encoding Servers Wrench.

Setting the encoder threads this way I have 16 x264/x265 threads and 12 avisynth threads. I think this gave me the best results. Though I did not test limiting to 12 threads for x264/x265 as that that seems to like threading.
I sent you a PM regarding this, but I think I have now figured it out...

So something like this, then :-

Code:
Server1CommandLine=/port 1000 /priority normal /restart-if-no-progress /avisynth-prefetch-threads 12 /x264-threads 16 /x265-threads 16
Server2CommandLine=/port 2000 /priority normal /restart-if-no-progress /avisynth-prefetch-threads 12 /x264-threads 16 /x265-threads 16

OR 

Server1CommandLine=/port 1000 /priority normal /restart-if-no-progress /avisynth-prefetch-threads 12 /x264-threads 12 /x265-threads 12
Server2CommandLine=/port 2000 /priority normal /restart-if-no-progress /avisynth-prefetch-threads 12 /x264-threads 12 /x265-threads 12
__________________
This can be SO "TeDiouS"..
Long term RipBot264 user.

#1 Ryzen 7950X
#2 Intel i9-13900KF
#3 Ryzen 5950X
#4 Ryzen 5900X
#5 Ryzen 3950X
TDS is online now   Reply With Quote
Old 9th December 2022, 05:40   #19813  |  Link
Pauly Dunne
Grumpy Old Man.
 
Pauly Dunne's Avatar
 
Join Date: Jul 2019
Location: Out There....
Posts: 675
ALL new Lite & full builds, ready to use..

Well, after about a week of re-hashing, re-configuring & updating the builds, they are ready to go...again

Lot's of important installation instruction's & hint's.

So PLEASE read them, it can prevent any problems.

More Filter's in the "full" PD build than before, including :-
Anime filters
Debanding filters
Dehaloing filters
De-interlacing filters
Derainbowing filters
Deringing filters
Resizing filters
Sharpening filters
and some special combo filter scripts, mainly for Anime type video's..
https://www.mediafire.com/file/bdkj1...12-22).7z/file (500Mb)

More info in the changelog.txt


Please Note:-

The PD_Lite build, has many component updates, but limited SMDegrain filter scripts available, ready to use, and that's all you get.
https://www.mediafire.com/file/bz3tm...2%2529.7z/file (410Mb)

More info in the changelog.txt

Good luck..
__________________
Not poorly done, just doin' it my way !!!
Live every day like it is your last,
because one day, it will be !! (M$B)
Pauly Dunne is offline   Reply With Quote
Old 9th December 2022, 10:04   #19814  |  Link
Ryushin
Registered User
 
Join Date: Mar 2011
Posts: 396
Quote:
Originally Posted by TDS View Post
I sent you a PM regarding this, but I think I have now figured it out...

So something like this, then :-

Code:
Server1CommandLine=/port 1000 /priority normal /restart-if-no-progress /avisynth-prefetch-threads 12 /x264-threads 16 /x265-threads 16
Server2CommandLine=/port 2000 /priority normal /restart-if-no-progress /avisynth-prefetch-threads 12 /x264-threads 16 /x265-threads 16

OR 

Server1CommandLine=/port 1000 /priority normal /restart-if-no-progress /avisynth-prefetch-threads 12 /x264-threads 12 /x265-threads 12
Server2CommandLine=/port 2000 /priority normal /restart-if-no-progress /avisynth-prefetch-threads 12 /x264-threads 12 /x265-threads 12
The first one:
Code:
Server1CommandLine=/port 1000 /priority normal /restart-if-no-progress /avisynth-prefetch-threads 12 /x264-threads 16 /x265-threads 16
Server2CommandLine=/port 2000 /priority normal /restart-if-no-progress /avisynth-prefetch-threads 12 /x264-threads 16 /x265-threads 16
I was surprised at first with the testing as I thought I could use just one encoding server instead of two. For 4K sources, one encoding server is nearly adequate. But with HD sources, two encoding servers needed to be used.
Ryushin is offline   Reply With Quote
Old 9th December 2022, 10:51   #19815  |  Link
TDS
Formally known as .......
 
TDS's Avatar
 
Join Date: Sep 2021
Location: On a need to know basis.
Posts: 543
Quote:
Originally Posted by Ryushin View Post
The first one:
Code:
Server1CommandLine=/port 1000 /priority normal /restart-if-no-progress /avisynth-prefetch-threads 12 /x264-threads 16 /x265-threads 16
Server2CommandLine=/port 2000 /priority normal /restart-if-no-progress /avisynth-prefetch-threads 12 /x264-threads 16 /x265-threads 16
I was surprised at first with the testing as I thought I could use just one encoding server instead of two. For 4K sources, one encoding server is nearly adequate. But with HD sources, two encoding servers needed to be used.
Thanks,

I will give that a try, when I get around to firing up the 7950X.
__________________
This can be SO "TeDiouS"..
Long term RipBot264 user.

#1 Ryzen 7950X
#2 Intel i9-13900KF
#3 Ryzen 5950X
#4 Ryzen 5900X
#5 Ryzen 3950X
TDS is online now   Reply With Quote
Old 9th December 2022, 16:53   #19816  |  Link
rlev11
Registered User
 
Join Date: Aug 2020
Location: Pennsylvania
Posts: 39
So a quick test comparing fps on 4k stuff using either the affinity mask or the avisynth-prefetch-threads setting them to 12 on my 16 core ryzens (1-3950x,3-5950x,1-7950x). About halfway thru a 4k movie once everything settles out, the total fps showing on the distributed encoding window are very similar between them. Within a couple fps in the 70-72 fps range. If these numbers hold up for 1080p stuff, this will help me out as I won't have to have 2 separate command batch files for the encoding servers and switch between them when doing 4k and anything else. I can just start everything with the prefetch numbers and be done with it. (720 and 1080 stuff runs much faster without the affinity mask set to 12 cores)

I encode everything to 10bit HEVC using default encoder settings with a CRF of 18

I have now switched over to using 3 SMDegrain scripts. In the past I was using just mdegrain3 with strengths of 200 for clean sources, 300 for light grain sources, and 400 for grainy sources. I did some testing and specifically comparing final file sizes, I now use the following SMDegrain scripts:
Clean sources - video=SMDegrain(video,tr=3,thSAD=300,thSADC=150,contrasharp=true,prefilter=1,refinemotion=true)
Light grain - video=SMDegrain(video,tr=4,thSAD=400,thSADC=200,contrasharp=true,prefilter=1,refinemotion=true)
Grainy - video=SMDegrain(video,tr=6,thSAD=600,thSADC=300,contrasharp=true,prefilter=2,refinemotion=true)

As I said, the end files sizes are very close to the corresponding mdegrain3 settings I had been using.

The biggest difference is in the speed increase using SMDegrain. I am seeing about a 40 - 60 % increase in fps using SMDegrain over mdegrain using comparable settings. 4k Forrest Gump with mdegrain was around 49-50 fps and with SMDegrain is 73-75 fps.

I did come across 1 movie that just would not run using SMDegrain, but did run using mdegrain, not sure what was up with that, so mdegrain is not completely dead

Last edited by rlev11; 9th December 2022 at 17:20.
rlev11 is offline   Reply With Quote
Old 10th December 2022, 00:50   #19817  |  Link
TDS
Formally known as .......
 
TDS's Avatar
 
Join Date: Sep 2021
Location: On a need to know basis.
Posts: 543
Quote:
Originally Posted by rlev11 View Post
So a quick test comparing fps on 4k stuff using either the affinity mask or the avisynth-prefetch-threads setting them to 12 on my 16 core ryzens (1-3950x,3-5950x,1-7950x). About halfway thru a 4k movie once everything settles out, the total fps showing on the distributed encoding window are very similar between them. Within a couple fps in the 70-72 fps range. If these numbers hold up for 1080p stuff, this will help me out as I won't have to have 2 separate command batch files for the encoding servers and switch between them when doing 4k and anything else. I can just start everything with the prefetch numbers and be done with it. (720 and 1080 stuff runs much faster without the affinity mask set to 12 cores)

I encode everything to 10bit HEVC using default encoder settings with a CRF of 18

I have now switched over to using 3 SMDegrain scripts. In the past I was using just mdegrain3 with strengths of 200 for clean sources, 300 for light grain sources, and 400 for grainy sources. I did some testing and specifically comparing final file sizes, I now use the following SMDegrain scripts:
Clean sources - video=SMDegrain(video,tr=3,thSAD=300,thSADC=150,contrasharp=true,prefilter=1,refinemotion=true)
Light grain - video=SMDegrain(video,tr=4,thSAD=400,thSADC=200,contrasharp=true,prefilter=1,refinemotion=true)
Grainy - video=SMDegrain(video,tr=6,thSAD=600,thSADC=300,contrasharp=true,prefilter=2,refinemotion=true)

As I said, the end files sizes are very close to the corresponding mdegrain3 settings I had been using.

The biggest difference is in the speed increase using SMDegrain. I am seeing about a 40 - 60 % increase in fps using SMDegrain over mdegrain using comparable settings. 4k Forrest Gump with mdegrain was around 49-50 fps and with SMDegrain is 73-75 fps.

I did come across 1 movie that just would not run using SMDegrain, but did run using mdegrain, not sure what was up with that, so mdegrain is not completely dead
Hi rlev11,

It's great to know that you seem to have sussed it out, and are very happy with the results, both in quality & encoding speed

I will be interested in hearing your thoughts, once you've tried the "non" affinity option, and how your 720p & 1080p encodes perform without changing your settings...

I also noticed that you've "bumped up" the strength of the script for "clean" sources....interesting !!

I would be very interested to know what the movie was that SMDegrain didn't like.

Sometimes (well, nearly always) I will run a movie thru MKVToolNix just to gut out all the stuff that I don't want, (less work & time for RB) and then run it thru RB, sometimes that will "fix" any problems.

But as I have said many times, when it comes to 4K stuff, I run it thru MKVToolNix, and just get the video track, and run that thru RB, and most of the time, it's also in 20 - 30 minute "chunks", so I don't have to run the encoding process ALL day.
__________________
This can be SO "TeDiouS"..
Long term RipBot264 user.

#1 Ryzen 7950X
#2 Intel i9-13900KF
#3 Ryzen 5950X
#4 Ryzen 5900X
#5 Ryzen 3950X
TDS is online now   Reply With Quote
Old 10th December 2022, 06:08   #19818  |  Link
Pauly Dunne
Grumpy Old Man.
 
Pauly Dunne's Avatar
 
Join Date: Jul 2019
Location: Out There....
Posts: 675
Here we go again, 24 hours later...

Hey guy's,

I have a couple of small, but important updates.

Dogway just never stops "fiddling"

So here's the update for the PD "full" build :-
https://www.mediafire.com/file/j50ec...l%2529.7z/file 148kb

and for the "Lite" build :-
https://www.mediafire.com/file/ma59v...e%2529.7z/file 90kb

Simply unpack, and copy into the PD_Tools folder...that's it
__________________
Not poorly done, just doin' it my way !!!
Live every day like it is your last,
because one day, it will be !! (M$B)
Pauly Dunne is offline   Reply With Quote
Old 10th December 2022, 16:10   #19819  |  Link
rlev11
Registered User
 
Join Date: Aug 2020
Location: Pennsylvania
Posts: 39
Quote:
Originally Posted by TDS View Post
Hi rlev11,

It's great to know that you seem to have sussed it out, and are very happy with the results, both in quality & encoding speed

I will be interested in hearing your thoughts, once you've tried the "non" affinity option, and how your 720p & 1080p encodes perform without changing your settings...

I also noticed that you've "bumped up" the strength of the script for "clean" sources....interesting !!

I would be very interested to know what the movie was that SMDegrain didn't like.

Sometimes (well, nearly always) I will run a movie thru MKVToolNix just to gut out all the stuff that I don't want, (less work & time for RB) and then run it thru RB, sometimes that will "fix" any problems.

But as I have said many times, when it comes to 4K stuff, I run it thru MKVToolNix, and just get the video track, and run that thru RB, and most of the time, it's also in 20 - 30 minute "chunks", so I don't have to run the encoding process ALL day.
I do the same thing, run everything through MKVToolNix and strip out everyhting but the videoand audio tracks and a subtitle if needed for foreign language. The movie was Iron Man 3

So it looks like I have to re-evaluate my earlier belief that with 720 and 1080 sources run better on the 16 core Ryzen's running full bore. I did a somewhat controlled test with a 720 and a 1080 source. I used the exact same settings and smdegrain script EXCEPT either running the 16 core Ryzen's in my farm full out, 12core using affinity mask, or just changing the avisynth prefetch settings to 12. My results were quite surprising to me.

I am just using the fps number at the bottom of the distributed encoding window when the last chunk gets to 25% so it is at the same point in each test. This is running my full farm of encoding servers with a single thread on everything except the 7950x which I have 2 servers active.

With the 16 core Ryzen's running at full bore the 720 ran at 530fps and the 1080 227fps

Running the Ryzens at 12 cpu with affinity mask setting the 720 ran at 560 and the 1080 at 218

Running the 16 core Ryzens just adding the avisynth prefetch threads set to 12 the 720 ran at 535 and the 1080 at 225

This is combined with earlier tests that doing 4k running the 16 core Ryzens full bore drops fps off the cliff, the end result as I can see is it doesn't matter too much weather you use affinity mask or prefetch threads set to 12, you should just use one of them and leave everything alone regardless of what resolution you are encoding. YMMV
rlev11 is offline   Reply With Quote
Old 12th December 2022, 12:59   #19820  |  Link
Ryushin
Registered User
 
Join Date: Mar 2011
Posts: 396
RipBot Not Starting Up on Boot

Hi Atak,

I've been having this annoying problem on one particular machine. RB won't start up automatically on boot. It's been doing this for a couple of years not, but I just have not been able to get it resolved.

I checked the box to start automatically, but it does not seem to take. Going right back into settings shows the box as unchecked.

RB used to start up fine on this machine, until I had to recreate my Windows User Profile after it got some weird corruption that was preventing a program from working and that only way to solve it was to create a fresh profile. I copied what I could from the old profile into the new profile so I didn't have to start from scratch.

I've tried going through the registry and deleted RB entries at one point but that did not solve the problem. Any insight into what registry keys I need to delete so RB can start back up on windows startup?
Ryushin is offline   Reply With Quote
Reply

Tags
264, 265, appletv, avchd, bluray, gui, iphone, ipod, ps3, psp, ripbot264, x264 2-pass, x264 gui, x264_64, x265, xbox360

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 07:04.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2023, vBulletin Solutions Inc.