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 > Capturing and Editing Video > Avisynth Development

Closed Thread
 
Thread Tools Search this Thread Display Modes
Old 9th February 2020, 23:14   #5141  |  Link
qyot27
...?
 
qyot27's Avatar
 
Join Date: Nov 2005
Location: Florida
Posts: 1,419
Quote:
Originally Posted by Stereodude View Post
Windows 7 is already out of mainline support and people are still using XP? I wouldn't advocate dropping it just to spite XP users, but at this point you can barely even run Windows 7 on the latest hardware (definitely not officially supported). Encoding has to be quite painfully slow on the most recent hardware that actually supports XP.

Are there really people trying to use their ~20 year old P4 with their nearly 20 year old OS to work with modern video formats and processing techniques? Given the hardware and OS they're using it seems like they shouldn't have any issues with running an old version of AVIsynth(+) too. They should be used to it since pretty much every application has already left them behind.
Let's also be fairly clear, too: anyone with one of those setups could just as easily install a 32-bit version of Linux (or 64-bit, if after the advent of Athlon64/Core 2) and run current versions of AviSynth+ using Wine.

For the ancient systems, like the Coppermine-128 that's gathering dustš under my desk, the Win7 or Win10 targetting isn't the problem (under Linux); it's the lack of SSE2. And as far as that goes, AviSynth+ has the MSVC_ARCH configuration option to set the C++ code optimization and generate builds compatible with that if you so choose (or you could go nuts and make builds for 64-bit that require AVX2, the option swings both ways).

šI barely turn the thing on anymore, since the Byte3 on top of my desk runs circles around it. For a while it served as an archival system due to having a large hard drive in it, but if I continue doing anything with it, it'll probably be as a ReactOS testbed.
qyot27 is offline  
Old 9th February 2020, 23:26   #5142  |  Link
Groucho2004
 
Join Date: Mar 2006
Location: Barcelona
Posts: 5,034
Quote:
Originally Posted by pinterf View Post
... as long as there is a relatively effortless option for building xp-friendly stuff, I'll keep it for a while.
Sounds reasonable.
__________________
Groucho's Avisynth Stuff
Groucho2004 is offline  
Old 10th February 2020, 00:11   #5143  |  Link
FranceBB
Broadcast Encoder
 
FranceBB's Avatar
 
Join Date: Nov 2013
Location: Royal Borough of Kensington & Chelsea, UK
Posts: 2,883
Quote:
Originally Posted by pinterf View Post
Yep, I've seen the usage statistics in the latest poll over there, which was showing the demand for XP - I don't know whether it comes from the typical age tree of the visitors of forum-like social media , anyway as long as there is a relatively effortless option for building xp-friendly stuff, I'll keep it for a while.
Thank you, Ferenc!

Quote:
Originally Posted by Stereodude View Post
Windows 7 is already out of mainline support and people are still using XP?
Yes we are, besides with the extended support we kept it updated 'till April 2019. By the way, companies paying Microsoft for "Premium Support" still receive updates every second Tuesday of the month; it's called "Patch Tuesday", but unfortunately there has been no one that shared those across the web (for obvious legally binding reasons ).

Quote:
Originally Posted by Stereodude View Post
at this point you can barely even run Windows 7 on the latest hardware (definitely not officially supported). Encoding has to be quite painfully slow on the most recent hardware that actually supports XP.
Not really. For instance I have Windows 10 at work, but I keep my Windows XP VM at home since I use Fedora + XP and I hate Windows 10 (truth to be told I have it but I fire it up only when I have to use Visual Studio, Matlab, Davinci Resolve, AVID and then I quickly shut it down).
There are countless of people on VMs with XP for many different reasons.

Quote:
Originally Posted by Stereodude View Post
Are there really people trying to use their ~20 year old P4 with their nearly 20 year old OS to work with modern video formats and processing techniques?
Look at Manolino! He's absolutely happy with his setup and nothing is stopping him from encoding files and that's the way it is!
(If I remember correctly even StainlessS was happily on XP up 'till few months ago).

Quote:
Originally Posted by Stereodude View Post
Given the hardware and OS they're using it seems like they shouldn't have any issues with running an old version of AVIsynth(+) too. They should be used to it since pretty much every application has already left them behind.
Not true. As I said there are many people running XP in VMs on modern hardware and there's a hugely active community on MSFN of XP users and I'm proudly part of it. As a matter of fact there are custom build of x265, custom builds of ffmpeg, custom builds of modern browsers based on Firefox and some official programs that still support XP like PotPlayer, AnyDesk, Miranda and so on. Life on XP is still not impossible.
As you can see from the net market share, XP doesn't lag so far behind many popular OS and is still worth it:





The only problem I see for Windows XP encoding-wise is the lack of modern indexers. LSMash is several years old and the latest build that is XP compatible is limited to 16bit stack or interleaved (I don't remember which one) not even planar 16bit.
ffms2 kept support for way longer, however in the end even qyot27 decided to drop XP support for the C plugin, eventually.
In case of foreign weird new codecs, people running XP are left with the very last version of ffmpeg so that they can decode the file, save it to something lossless like huffyuv, ffv1, utvideo and so on and then index and work with it in Avisynth, which is not exactly the best thing ever but it works.
Anyway, I still forecast a not-so-bad future for XP die-hard, especially 'cause there are many people willing to help and there's a very good community which I'm happily part of, so I thank Ferenc for keeping XP-Support.

Last edited by FranceBB; 10th February 2020 at 00:16.
FranceBB is online now  
Old 10th February 2020, 02:01   #5144  |  Link
Stereodude
Registered User
 
Join Date: Dec 2002
Location: Region 0
Posts: 1,436
Quote:
Originally Posted by StainlessS View Post
Who cares, set it going and tell it to auto shutdown when done, easy peasy. [I prefer not to encode on my No1 Machine].
So in a month or two your encoding job will be done? A 3.06gHz P4 with HT is likely 40-50x slower (extrapolating from Passmark scores) than my E5-2687W v2, which I think is too slow to be practical. I don't consider x265 encoding at 0.00625fps to be viable. At some point the cost of just the electricity to run the job on an old system vs. modern HW would pay for the new system.

Quote:
Originally Posted by Groucho2004 View Post
Not true. I just checked Asus support page, most AM4 boards support W7 and these Ryzen CPUs are very fast when it comes to encoding video.
Which is great, but Ryzen 3000 and the Intel 9xxx stuff doesn't work with Windows 7 (unless you somehow count not having USB as working) regardless of whether there are Windows 7 drivers for a motherboard that will take the CPU. I stand by my comment that the latest HW barely works under Windows 7.

Last edited by Stereodude; 10th February 2020 at 02:04.
Stereodude is offline  
Old 11th February 2020, 10:42   #5145  |  Link
tormento
Acid fr0g
 
tormento's Avatar
 
Join Date: May 2002
Location: Italy
Posts: 2,542
Quote:
Originally Posted by pinterf View Post
With XP support in our mind, releasing an LLVM clang build is impossible as well, LLVM dropped XP support three (?) years ago. (Though not considering it at the moment, I did not test for speed difference yet) Of course I can make installers for all different flavours, but I'd prefer only one version, I don't have infinite time and to tell the truth it's quite inconveniant.
I clearly understand.

I find, anyway, unreasonable to keep on dragging a useless weight for 1.34% of the users when the other ~98% would benefit from CLANG compiles of Avisynth+, MVTools2, MaskTools and RGTools.
__________________
@turment on Telegram
tormento is offline  
Old 11th February 2020, 10:53   #5146  |  Link
pinterf
Registered User
 
Join Date: Jan 2014
Posts: 2,309
Clang is not magic.
There are cases when clang is better, but sometimes msvc build is quicker. The race result even depends on whether we are seeing SSE4 or AVX2 code paths or just a different function.
In RgTools for example one mode is quicker with clang build, others are better with microsoft compiler.
pinterf is offline  
Old 11th February 2020, 18:13   #5147  |  Link
StainlessS
HeartlessS Usurer
 
StainlessS's Avatar
 
Join Date: Dec 2009
Location: Over the rainbow
Posts: 10,980
Quote:
I find, anyway, unreasonable to keep on dragging a useless weight for 1.34% of the users when the other ~98% would benefit from CLANG compiles of Avisynth+, MVTools2, MaskTools and RGTools.
Usage Poll, 10 XP users versus 29 W10 users of Avisynth, is a bit more than 1.34%.

EDIT: Current total OS usage sample 58.
I use 2 * XP, 1 * W7, 1*W10, and voted for all 3, [was only supposed to vote for 1]
__________________
I sometimes post sober.
StainlessS@MediaFire ::: AND/OR ::: StainlessS@SendSpace

"Some infinities are bigger than other infinities", but how many of them are infinitely bigger ???

Last edited by StainlessS; 11th February 2020 at 18:20.
StainlessS is offline  
Old 15th February 2020, 12:12   #5148  |  Link
real.finder
Registered User
 
Join Date: Jan 2012
Location: Mesopotamia
Posts: 2,587
expr don't have >> like lut? I got error with expr("x 100 scalef <= x 1 >> x 1.4 ^ ?",scale_inputs="floatf")
__________________
See My Avisynth Stuff
real.finder is offline  
Old 15th February 2020, 13:10   #5149  |  Link
pinterf
Registered User
 
Join Date: Jan 2014
Posts: 2,309
No binary shifts. You have to use /2 or *0.5 instead. Plus an integer truncation if being integer is important.
pinterf is offline  
Old 15th February 2020, 14:41   #5150  |  Link
TheFluff
Excessively jovial fellow
 
Join Date: Jun 2004
Location: rude
Posts: 1,100
Quote:
Originally Posted by real.finder View Post
expr don't have >> like lut? I got error with expr("x 100 scalef <= x 1 >> x 1.4 ^ ?",scale_inputs="floatf")
expr converts all inputs to single precision float and works with that internally, so binary left/right shift isn't really meaningful in the way it is for integers.
TheFluff is offline  
Old 15th February 2020, 15:24   #5151  |  Link
real.finder
Registered User
 
Join Date: Jan 2012
Location: Mesopotamia
Posts: 2,587
thank you both pinterf and TheFluff

Currently I work on port ContraSharpen_mod for Nnedi3_resize16 both by mawen1250 to avs+ and it has this expr, I will leave it in mt_lut for now until I fully complete the script
__________________
See My Avisynth Stuff
real.finder is offline  
Old 16th February 2020, 12:15   #5152  |  Link
real.finder
Registered User
 
Join Date: Jan 2012
Location: Mesopotamia
Posts: 2,587
if I go from out-of-range float clip (not 0..1.0 luma nor -0.5..0.5 chroma) to 8 bit int I got kinda wrong output unless I use mt_lut("x -0.5 0.5 clip",yexpr="x 0 1 clip",u=3,v=3) or Limiter(0,1,-0.5,0.5) before I convert to 8 bit, is this bug or this is normal?
__________________
See My Avisynth Stuff
real.finder is offline  
Old 17th February 2020, 10:32   #5153  |  Link
pinterf
Registered User
 
Join Date: Jan 2014
Posts: 2,309
Quote:
Originally Posted by real.finder View Post
if I go from out-of-range float clip (not 0..1.0 luma nor -0.5..0.5 chroma) to 8 bit int I got kinda wrong output unless I use mt_lut("x -0.5 0.5 clip",yexpr="x 0 1 clip",u=3,v=3) or Limiter(0,1,-0.5,0.5) before I convert to 8 bit, is this bug or this is normal?
If it does not even fit into a 32 bit integer, there may be problems, but I'll check it when you specify how much it is out of range.
pinterf is offline  
Old 17th February 2020, 12:12   #5154  |  Link
real.finder
Registered User
 
Join Date: Jan 2012
Location: Mesopotamia
Posts: 2,587
Quote:
Originally Posted by pinterf View Post
If it does not even fit into a 32 bit integer, there may be problems, but I'll check it when you specify how much it is out of range.
test with this fat lut/expr from ContraSharpen_mod

Code:
ColorBars(width=640, height=480).converttoyv12
ConvertBits(32)
mt_lut("x 128 == x x 128 - abs 11 / 22 ^ 11 * "+string(33/100.0)+" * x 128 > 1 -1 ? * x 128 - 2 ^ 11 2 ^ 44 + * x 128 - 2 ^ 44 + 11 2 ^ * / * 1 55 0 == 0 11 55 / 4 ^ ? + 1 55 0 == 0 x 128 - abs 55 / 4 ^ ? + / * 128 + ?",scale_inputs="allf",use_expr=2, U=3, V=3)
ConvertBits(8) #ok with isvideofloat? Limiter(0,1,-0.5,0.5).ConvertBits(8) : ConvertBits(8)
__________________
See My Avisynth Stuff
real.finder is offline  
Old 17th February 2020, 14:14   #5155  |  Link
pinterf
Registered User
 
Join Date: Jan 2014
Posts: 2,309
Thanks for the sample. When the pixel value is greater 128 (float format) - not that the usual max. is 1.0 - then the 8 bit conversion will not clamp to 255. Rare case, nevertheless I have fixed it on my side.
pinterf is offline  
Old 17th February 2020, 15:34   #5156  |  Link
amichaelt
Guest
 
Posts: n/a
Quote:
Originally Posted by StainlessS View Post
Usage Poll, 10 XP users versus 29 W10 users of Avisynth, is a bit more than 1.34%.

EDIT: Current total OS usage sample 58.
I use 2 * XP, 1 * W7, 1*W10, and voted for all 3, [was only supposed to vote for 1]
https://en.wikipedia.org/wiki/Selection_bias
 
Old 17th February 2020, 16:29   #5157  |  Link
StainlessS
HeartlessS Usurer
 
StainlessS's Avatar
 
Join Date: Dec 2009
Location: Over the rainbow
Posts: 10,980
Yep I see what you mean, I should have only voted for XP, as that is the one I use on most machines.
__________________
I sometimes post sober.
StainlessS@MediaFire ::: AND/OR ::: StainlessS@SendSpace

"Some infinities are bigger than other infinities", but how many of them are infinitely bigger ???
StainlessS is offline  
Old 17th February 2020, 21:27   #5158  |  Link
tormento
Acid fr0g
 
tormento's Avatar
 
Join Date: May 2002
Location: Italy
Posts: 2,542
Quote:
Originally Posted by StainlessS View Post
Usage Poll, 10 XP users versus 29 W10 users of Avisynth, is a bit more than 1.34%.
This forum is everything but a representative sample.
__________________
@turment on Telegram
tormento is offline  
Old 17th February 2020, 21:54   #5159  |  Link
StainlessS
HeartlessS Usurer
 
StainlessS's Avatar
 
Join Date: Dec 2009
Location: Over the rainbow
Posts: 10,980
Yep you are correct, there will be MANY more XP users than show up on Internet Stats, many are totally offline (One of mine is occasionally given Internet access [several months since] the other one NEVER).
__________________
I sometimes post sober.
StainlessS@MediaFire ::: AND/OR ::: StainlessS@SendSpace

"Some infinities are bigger than other infinities", but how many of them are infinitely bigger ???
StainlessS is offline  
Old 18th February 2020, 03:14   #5160  |  Link
StainlessS
HeartlessS Usurer
 
StainlessS's Avatar
 
Join Date: Dec 2009
Location: Over the rainbow
Posts: 10,980
Quote:
Originally Posted by StainlessS View Post
Dont know if below is bugged, or misleading error message, or both.

Code:
# OK
Blankclip(Length=1,Width=1,height=1,Pixel_type="YV24").PointResize(256,256)

# Resize: source image too small for this resize method. Width=1, Support=1.
Blankclip(Length=1,Width=1,height=1,Pixel_type="YV24").BilinearResize(256,256)

# Resize: source image too small for this resize method. Width=2, Support=2.
Blankclip(Length=1,Width=4,height=4,Pixel_type="YV12").BicubicResize(256,256)   # NOTE, YV12, Presumably error for chroma

# Resize: source image too small for this resize method. Width=2, Support=2.
Blankclip(Length=1,Width=2,height=2,Pixel_type="YV24").BicubicResize(256,256)   # NOTE, YV24

# Resize: source image too small for this resize method. Width=4, Support=4.
Blankclip(Length=1,Width=4,height=4,Pixel_type="YV24").Lanczos4Resize(256,256)

# Resize: source image too small for this resize method. Width=4, Support=4.
Blankclip(Length=1,Width=4,height=4,Pixel_type="YV24").GaussResize(256,256)

# Resize: source image too small for this resize method. Width=4, Support=4.
Blankclip(Length=1,Width=4,height=4,Pixel_type="YV24").BlackmanResize(256,256)

# Resize: source image too small for this resize method. Width=3, Support=3.
Blankclip(Length=1,Width=3,height=3,Pixel_type="YV24").LanczosResize(256,256)

# Resize: source image too small for this resize method. Width=4, Support=4.
Blankclip(Length=1,Width=4,height=4,Pixel_type="YV24").sincResize(256,256)

# Resize: source image too small for this resize method. Width=2, Support=2.
Blankclip(Length=1,Width=2,height=2,Pixel_type="YV24").Spline16Resize(256,256)

# Resize: source image too small for this resize method. Width=3, Support=3.
Blankclip(Length=1,Width=3,height=3,Pixel_type="YV24").Spline36Resize(256,256)

# Resize: source image too small for this resize method. Width=3, Support=3.
Blankclip(Length=1,Width=3,height=3,Pixel_type="YV24").Spline36Resize(256,256)

Return Last
For eg YV24, width=2, height=2, Says width = 2, Support=2 (when it clearly produces an error message, ie seems not to support 2).

Think its been like this 4E4 (believe it exists as above in avs standard).

EDIT: Presume similar on height, untried.
Above still happens.

Also,
Code:
W=1
H=1
CS="RGB32"
BlankClip(Width=256,Height=256,pixel_type=CS)
#BiLinearResize(W,H)   # Resize: Source image too small for this resize method. Width=256, Support=256.
BicubicResize(W,H)    # Resize: Source image too small for this resize method. Width=256, Support=512.
Return Last
Presume same for Height [and other Resizers/colorspaces], untested.
EDIT: "AviSynth+ 3.5 (r2947, master, i386)"

EDIT: Even chasing the error message is fun
Code:
W=1
H=1
CS="RGB32"
#BlankClip(Width=256,Height=256,pixel_type=CS)
#BiLinearResize(W,H)   # Resize: Source image too small for this resize method. Width=256, Support=256.
#BicubicResize(W,H)     # Resize: Source image too small for this resize method. Width=256, Support=512.

# Even chasing the error message is fun
#BlankClip(Width=512,Height=512,pixel_type=CS)
#BicubicResize(W,H)    # Resize: Source image too small for this resize method. Width=512, Support=1024.

BlankClip(Width=1024,Height=1024,pixel_type=CS)
BicubicResize(W,H)    # Resize: Source image too small for this resize method. Width=1024, Support=2048.

Return Last
__________________
I sometimes post sober.
StainlessS@MediaFire ::: AND/OR ::: StainlessS@SendSpace

"Some infinities are bigger than other infinities", but how many of them are infinitely bigger ???

Last edited by StainlessS; 18th February 2020 at 07:10.
StainlessS is offline  
Closed Thread

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 09:38.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.