Selecteven () with DGSource ?

[zappa]

very good night a greeting to all

for example if I have a video at 59.940 (60000/1001) FPS and I want to convert it to 29.970 FPS, all I know you have to apply the Selecteven ()

But how could this be done with DGSource() ????

Thank you very much for your help

-------------------------------------------------------------

DGSource("C:\Users\c\Desktop\rio-part1.dgi")

-------------------------------------------------------------

the question is very specific please, I do not want alternatives I want to know if you can and how would be the parameter to do it with the DGSource

in other words equivalent to Selecteven() in DGSource() ???

thanks so much

[MeteorRain]

It couldn't be done inside dgsource.

[zappa]

It couldn't be done inside dgsource.

hello what about

DGTelecide() or DGDecimate() ???

thanks for your help

[poisondeathray]

hello what about

DGTelecide() or DGDecimate() ???

thanks for your help

They are standalone filters . They don't work "inside" DGSource

Why are you asking this ? What do you want to do exactly ? What is wrong with SelectEven() ?

[zappa]

They are standalone filters . They don't work "inside" DGSource

Why are you asking this ? What do you want to do exactly ? What is wrong with SelectEven() ?

There is nothing wrong with Selecteven() , that is not the question , I just want to know if you can do it with the DGIndexNV is a personal curiosity , do you know how you could do it please ?

maybe with DGTelecide() or DGDecimate() , etc

Thank you very much

[wonkey_monkey]

DGSource("...")
SelectEven

That's how it's done...

It can be helpful to explain your intentions sometimes, otherwise we won't be able to understand why just using selecteven() isn't enough.

[Groucho2004]

very good night a greeting to all

for example if I have a video at 59.940 (60000/1001) FPS and I want to convert it to 29.970 FPS, all I know you have to apply the Selecteven ()

But how could this be done with DGSource() ????

Thank you very much for your help

-------------------------------------------------------------

DGSource("C:\Users\c\Desktop\rio-part1.dgi")

-------------------------------------------------------------

the question is very specific please, I do not want alternatives I want to know if you can and how would be the parameter to do it with the DGSource

in other words equivalent to Selecteven() in DGSource() ???

thanks so much
I suppose the closest to Selecteven() in dgsource is "deinterlace = 1" for single rate de-interlacing. It's in the manual for DGSource.

[johnmeyer]

He's asking the same thing over in the videohelp.com forum and no one there can figure out what he's asking.

[videoh]

Hey zappa. Author of DGSource() here. As others have said you cannot implement a SelectEven() in DGSource(). There is no reason to do so anyway, because SelectEven() has effectively zero processing demand in your script.

Thank you for your interest in DG tools. :thanks:

[StainlessS]

"I just want to know if you can do it with the DGIndexNV is a personal curiosity"

No particular reason to know at all, kinda like, "just curious, can I get to the moon in a tractor ?, if so, how ?".

[zappa]

"I just want to know if you can do it with the DGIndexNV is a personal curiosity"

No particular reason to know at all, kinda like, "just curious, can I get to the moon in a tractor ?, if so, how ?".

and.. whats the problem ? :D

It's a question,... if you don't know the answer just don't answer..

"Thank you for nothing".

If you like drama, go look for it somewhere else..

Have a nice day:D

[zappa]

Hey zappa. Author of DGSource() here. As others have said you cannot implement a SelectEven() in DGSource(). There is no reason to do so anyway, because SelectEven() has effectively zero processing demand in your script.

Thank you for your interest in DG tools. :thanks:

Thank you very much

have a good day :D

[zappa]

He's asking the same thing over in the videohelp.com forum and no one there can figure out what he's asking.


The question is very clear and direct I don't understand what your confusion is, if you don't know the answer then you can't help me, it's that simple. Thank you for trying.

Have a good day

[zappa]

I suppose the closest to Selecteven() in dgsource is "deinterlace = 1" for single rate de-interlacing. It's in the manual for DGSource.


Thank you very much for your help, apparently you are one of the few who understood my question. Thanks you Sr. !

...Unfortunately , there are people who understand the problem and play dumb.... (not You)

Also thanks the user "jagabo" @videohelp.com, who always helps in an effective and good way.

Greetings and thank you for your answer.

[zappa]

DGSource("...")
SelectEven

That's how it's done...

It can be helpful to explain your intentions sometimes, otherwise we won't be able to understand why just using selecteven() isn't enough.

Thanks Sr.

selecteven() equivalent in DGindexNV ??


for example if I have a video at 59.940 (60000/1001) FPS and I want to convert it to 29.970 FPS, all I know you have to apply the Selecteven () in Avisynth.

DGSource("C:\Users\c\Desktop\0001.dgi") = DGindexNV
SelectEven() = Avisynth

SelectEven() Equivalent in DGindexNV ????



Many of you have already understood my question and I thank you for the answer.

Thank you all for trying to help me.:thanks:

[zappa]

Hey zappa. Author of DGSource() here. As others have said you cannot implement a SelectEven() in DGSource(). There is no reason to do so anyway, because SelectEven() has effectively zero processing demand in your script.

Thank you for your interest in DG tools. :thanks:

it would be possible to implement in DG tools some new tools like Denoisers that is widely used by everyone as the "SMDegrain" or maybe the "DFFTEST" or maybe the QTGMC for the interlaced videos ?.

the QTGMC is I think the best for interlaced video. But it has functions that involve other functions and plugin in Avisynth, I don't think it's easy. If for some reason you would implement it it would be very used.

What you have currently implemented in DG tools is very good but does not have much quality (in my opinion, very personal), for example in interlaced video the QTGMC is much better, another example to apply noise reduction is much better the SMDegrain or the DFFTEST or the FFT3DFilter, All Avisynth users use it a lot.

is just pure curiosity would be very complex to implement ? and if you can not, DG tool is very useful to me equally. especially when I use it with the NVEncC64. It saves a lot of time when encoding. Too much,

The quality is almost the same as long as you know how to configure the options.

Greetings and thank you very much for your help.

[videoh]

it would be possible to implement in DG tools some new tools like Denoisers that is widely used by everyone as the "SMDegrain" or maybe the "DFFTEST" or maybe the QTGMC for the interlaced videos ?. There are already some CUDA denoising/decombing filters in DGDecodeNV.dll. The only reason to duplicate the ones you mention would be to CUDA accelerate them. I am always investigating opportunities for new CUDA filters. Practically speaking, however, porting such a thing as QTGMC to CUDA is a massive undertaking. I have limited time in my life and prefer to use it for new things and for expressing my creative urges, not just for porting existing tools. If it was easy I'd have done it long ago.

the QTGMC is I think the best for interlaced video. But it has functions that involve other functions and plugin in Avisynth, I don't think it's easy. If for some reason you would implement it it would be very used. In fact it is easy to use these things in your script. If you are having difficulties probably some more study of the available resources/examples would serve you well. It's not my job to make setting up/using QTGMC easy for you. Did you ask the LSmash or FFMS guys the same questions? ;)

What you have currently implemented in DG tools is very good but does not have much quality (in my opinion, very personal), for example in interlaced video the QTGMC is much better, another example to apply noise reduction is much better the SMDegrain or the DFFTEST or the FFT3DFilter, All Avisynth users use it a lot. See above. Go ahead and use whatever you think is best for you and your use cases. Word to the wise: Telling a developer his stuff has "not much quality" is a good way to ensure you won't get much love from him/her. I've been patient here with you so far but that is not unlimited as you've given us good reason to believe you're either a troll or a clueless newbie. Maybe both.

[real.finder]

@zappa

all you asked for has cuda already, KTGMC and KSMDegrain https://github.com/nekopanda/AviSynthCUDAFilters/releases

[zappa]

There are already some CUDA denoising/decombing filters in DGDecodeNV.dll. The only reason to duplicate the ones you mention would be to CUDA accelerate them. I am always investigating opportunities for new CUDA filters. Practically speaking, however, porting such a thing as QTGMC to CUDA is a massive undertaking. I have limited time in my life and prefer to use it for new things and for expressing my creative urges, not just for porting existing tools. If it was easy I'd have done it long ago.

In fact it is easy to use these things in your script. If you are having difficulties probably some more study of the available resources/examples would serve you well. It's not my job to make setting up/using QTGMC easy for you. Did you ask the LSmash or FFMS guys the same questions? ;)

See above. Go ahead and use whatever you think is best for you and your use cases. Word to the wise: Telling a developer his stuff has "not much quality" is a good way to ensure you won't get much love from him/her. I've been patient here with you so far but that is not unlimited as you've given us good reason to believe you're either a troll or a clueless newbie. Maybe both.


I apologize, but it was not my intention to disrespect you, I have only questions regarding DG tools.

I don't have problems with the QTGMC I use it perfectly, but it takes time because it does a good job, the idea would be to implement it in DG tools to speed up that process And that is the question nothing else. But as it is so complex and involves other functions and plugins of Avisynth is very difficult to implement (I understand that, but I wanted to ask you personally) is not that I do not know how to use ..

Don't take my questions the wrong way please it's just Curiosity, I also like DG tools a lot, it helps a lot to reduce the time of the encodes.

Thank you very much for your time and your work.

:thanks:

[zappa]

@zappa

all you asked for has cuda already, KTGMC and KSMDegrain https://github.com/nekopanda/AviSynthCUDAFilters/releases

I didn't know there was a project

WOW thanks so much !!!

:thanks::thanks::thanks::thanks:

[real.finder]

I didn't know there was a project

WOW thanks so much !!!

:thanks::thanks::thanks::thanks:

you are welcome

in case you don't know, you also need this https://github.com/nekopanda/AviSynthPlus/releases (called AVSNeo or AvisynthNeo)

and here how to use both https://forum.doom9.org/showthread.php?p=1820381&highlight=cuda#post1820381

[zappa]

you are welcome

in case you don't know, you also need this https://github.com/nekopanda/AviSynthPlus/releases (called AVSNeo or AvisynthNeo)

and here how to use both https://forum.doom9.org/showthread.php?p=1820381&highlight=cuda#post1820381

I'm sorry to bother you but I have a little problem with this plugin.

I'm trying to use it but I get this error, I already checked all the information and find nothing

-Install Avisynth Neo the latest version x64
-I read all the information and nothing.
-I have a Nvidia 1060 6 GB Card
-Updated Windows 10 x64
-KTGMC x64
-everything is in its place

but even so when I try to use it so the avisynth script I get the following error

KTGMC ()

https://i.imgur.com/6vwKzL7.jpg

https://i.imgur.com/zavAMsC.jpg

I'll be very grateful for your help.

[MeteorRain]

You are mixing things up and that's what people were keeping telling you.

DGSource is a video decoder that decodes video using Nvidia hardware decoder. That's it, nothing more and nothing less. And there are other DG tools that do their own job.

If other tools have done some other job, then there's no point wasting precious developing time to reinvent the wheel. And even someone does reinvent, people would still use the original one anyway.

SMDegrain and other tools have nothing to do with DGNV or videoh. It just sounds as stupid as asking Microsoft to implement those tools because they ran on Windows.

Software developing is very expensive. You can hire and pay for what you want (hint, it's gonna be thousands of dollars to build a faster version of QTGMC and SMDegrain). You can also politely ask if someone could make it. You can also bear with it, you know, maybe get a faster CPU.

As a side note, claiming that you are polite has nothing to do with behaving impolitely.

[MeteorRain]

使い方
OnCPUとOnCUDAで囲って使ってください。

OnCPUやOnCUDA、SetMemoryMaxの詳しい説明はAviSynthNeoを見てください。
https://github.com/nekopanda/AviSynthCUDAFilters/wiki/KTGMC

CUDA用にNeoで追加された関数

OnCPU(clip, int "num_prefetch")
指定したクリップをCPUで処理します。

OnCUDA(clip, int "num_prefetch", int "device_index")
指定したクリップをGPUで処理します。
https://github.com/nekopanda/AviSynthPlus/wiki/Avisynth-Neo

[zappa]

You are mixing things up and that's what people were keeping telling you.

DGNV is a video decoder that decodes video using Nvidia hardware decoder. That's it, nothing more and nothing less.

If other tools have done some other job, then there's no point wasting precious developing time to reinvent the wheel. And even someone does reinvent, people would still use the original one anyway.

SMDegrain and other tools have nothing to do with DGNV or videoh. It just sounds as stupid as asking Microsoft to implement those tools because they ran on Windows.

Software developing is very expensive. You can hire and pay for what you want (hint, it's gonna be thousands of dollars to build a faster version of QTGMC and SMDegrain). You can also politely ask if someone could make it. You can also bear with it, you know, maybe get a faster CPU.

As a side note, claiming that you are polite has nothing to do with behaving impolitely.


You're totally wrong. You didn't understand the question. I'm not looking for anything for free, and I'm not asking meaningless questions either. YOU ARE TOTALLY WRONG.

I am free to ask in a good way what I want and when I want, as long as it is respecting the rules and regulations always and without disrespecting anyone.

This forum is to help, not to criticize or politicize an issue or try to disrespect a person who asks respectfully. etc. .

The idea is to save time by sending some processes to the GPU without losing quality in the encode, "DG tools" has other tools like Denoisers , Deinterlaces , etc , etc .

So you can ask if they could include more

What's wrong with asking that?

If you do not like the question or do not understand the meaning of the question is your problem, I do not like the drama, if you look for protagonism or make clear your point is fine, I respect you, but do not say that I do "stupid questions" ... That's disrespectful to a person. And I never disrespect you.

Thank you for your comment and have a good day.

[zappa]

https://github.com/nekopanda/AviSynthCUDAFilters/wiki/KTGMC


https://github.com/nekopanda/AviSynthPlus/wiki/Avisynth-Neo

Thank you very much just using the translator I found that answer, thank you very much for the answer.

Now is working 100 %

this way it works very well


DGSource("C:\Users\c\Desktop\VideoFile.dgi").OnCPU(0)

KTGMC().OnCUDA(0)

[Groucho2004]

You are mixing things up and that's what people were keeping telling you.

DGNV is a video decoder that decodes video using Nvidia hardware decoder. That's it, nothing more and nothing less.
Not quite. DGDecodeNV can utilise Nvidia's hardware de-interlacer, remove soft pulldown and recent versions have denoisers and whatnot. Who's mixing things up?

[zappa]

https://github.com/nekopanda/AviSynthCUDAFilters/wiki/KTGMC


https://github.com/nekopanda/AviSynthPlus/wiki/Avisynth-Neo

Thanks so much for your help

[MeteorRain]

Not quite. DGDecodeNV can utilise Nvidia's hardware de-interlacer, remove soft pulldown and recent versions have denoisers and whatnot. Who's mixing things up?

My bad. Should have listed all video decoder related technologies in that reply to make people understand that SelectEven is not the job of DGNV. I was in a hurry and didn't spend the time to write a multi page essay. I'll be careful next time.

[Groucho2004]

I was in a hurry and didn't spend the time to write a multi page essay. I'll be careful next time.Smooooth. :)

[MeteorRain]

The idea is to save time by sending some processes to the GPU without losing quality in the encode, "DG tools" has other tools like Denoisers , Deinterlaces , etc , etc .

Zappa. This is my opinion. Could be right, could be wrong.

You either asked the wrong question or you asked in the wrong way.

I noticed that you are not a native English speaker, so there could be misunderstanding and miscommunication.

If your idea is to save time by sending processes to the GPU, you should have asked so. With that, your intention is very clear, and it's easier for people to help. By simply asking about SelectEven, people barely understood your intention, and answers were very hard to come.

Besides, functions like SelectEven() don't need "processing" and thus are not affected by whether they are using CPU or GPU. SelectEven simply requests its upstream every other frame. Frames that are not requested are simply skipped and do not consume time. And thus, writing DGSource().SelectEven() would be exactly the same as implementing and using SelectEven inside DGSource.

As I mentioned in other threads before, it's kinda a X-Y problem, where you want X but you ask Y, and people are confused.

EDIT: I hope this post doesn't look too offensive. My intention is to help you on understanding things and how to efficiently ask questions.

[zappa]

Zappa. This is my opinion. Could be right, could be wrong.

You either asked the wrong question or you asked in the wrong way.

I noticed that you are not a native English speaker, so there could be misunderstanding and miscommunication.

If your idea is to save time by sending processes to the GPU, you should have asked so. With that, your intention is very clear, and it's easier for people to help. By simply asking about SelectEven, people barely understood your intention, and answers were very hard to come.

Besides, functions like SelectEven() don't need "processing" and thus are not affected by whether they are using CPU or GPU. SelectEven simply requests its upstream every other frame. Frames that are not requested are simply skipped and do not consume time. And thus, writing DGSource().SelectEven() would be exactly the same as implementing and using SelectEven inside DGSource.

As I mentioned in other threads before, it's kinda a X-Y problem, where you want X but you ask Y, and people are confused.

EDIT: I hope this post doesn't look too offensive. My intention is to help you on understanding things and how to efficiently ask questions.


I still think I was very clear when I asked, but anyway I will take your advice as valid because you are asking respectfully and I understand that it is with good intention.

No problem, everything is fine, thank you very much for trying to help me, next time I will try to be clearer and more precise in my questions.

Greetings and have a nice day.

[TCmullet]

Hey zappa. Author of DGSource() here. As others have said you cannot implement a SelectEven() in DGSource(). There is no reason to do so anyway, because SelectEven() has effectively zero processing demand in your script.

Thank you for your interest in DG tools. :thanks:
I ran across this odd thread and felt I wanted to say something. DG, I greatly appreciate your tools DGIndexNV! I think I have 4 licenses, but use only 2 lately. Equally valuable but for a different reason, I greatly appreciate DGBob. Without that, all my i30 footage would not be convertible to p60, which Youtube (as of 2014) finally honors. I also use Mpeg2Source for the times I have to deal with Mpeg2 vs h264.

So a big public THANK YOU for these things!

P.S. I've just gotten a bit into H.265, and discovered (with some expectation) that DGIndexNV does not accept those files. That's okay... I'll be use H.265 ONLY as a delivery codec, keeping all files that ever may need work as H.264.

[videoh]

I've just gotten a bit into H.265, and discovered (with some expectation) that DGIndexNV does not accept those files. That's okay... I'll be use H.265 ONLY as a delivery codec, keeping all files that ever may need work as H.264. Nope, DGDecNV handles H.265 just fine. If you have a problematic stream, please give a link to a sample.

[DJATOM]

Nope, DGDecNV handles H.265 just fine. If you have a problematic stream, please give a link to a sample.

My guess is too old and cheap GPU.

[MeteorRain]

Very possible the GPU doesn't support HEVC decoding.

We have just tested a sample that is HEVC 10bit 60fps 7680x4320 and the latest DG is running smooth with that, zero problem, of course providing that you have the correct video card to handle it.

Do you have a GTX 950 or better?

[videoh]

Ah, yes, guys, thank you. Spoiled by my 2080 Ti, I didn't think about that.

[TCmullet]

Very possible the GPU doesn't support HEVC decoding.

We have just tested a sample that is HEVC 10bit 60fps 7680x4320 and the latest DG is running smooth with that, zero problem, of course providing that you have the correct video card to handle it.

Do you have a GTX 950 or better?
Where in the Nvidia Control Panel do I find that info?

FWIW, I do not consider my card to have been cheap. Has 192 CUDA cores.

[videoh]

Where in the Nvidia Control Panel do I find that info? Help/System Information. You can also get that from DGIndexNV Video/List GPU Devices.

FWIW, I do not consider my card to have been cheap. Has 192 CUDA cores. It's all relative. My card has 4362 CUDA cores. Sounds like you must have a low-end card. Something like a GTX 260?

https://www.studio1productions.com/Articles/NVidia-GPU-Chart.htm

[zappa]

Help/System Information. You can also get that from DGIndexNV Video/List GPU Devices.

It's all relative. My card has 4362 CUDA cores. Sounds like you must have a low-end card. Something like a GTX 260?

https://www.studio1productions.com/Articles/NVidia-GPU-Chart.htm

https://github.com/nekopanda/AviSynthCUDAFilters

hello.

Would it be possible to implement these filters in DGtools ?

KTGMC = QTGMC with CUDA
KNNEDI3 = NNEDI3 with CUDA
KSDegrainN = SMDegrain with cuda
And more...

All of these have already been passed on to CUDA. with .dll

https://github.com/nekopanda/AviSynthCUDAFilters

https://github.com/nekopanda/AviSynthCUDAFilters/releases


These are the best filters for each case. this would be very good and we could use your program integrally without having to look for other options to complement it.

Thank you for your help

[videoh]

You're asking me to steal his code? No, it is not possible.

[StainlessS]

You're asking me to steal his code?
Surely not if you if you comply with the GPL. [ I notice some recognisable code in there, eg Grunt(), I'm sure that Big G dont mind ].

[videoh]

DG tools are closed source. I cannot use GPL code without special permission.

[StainlessS]

Oh yes of course, my bad. :(

EDIT: That link you provided for graphics cards, and what to avoid is real handy, I snapped a copy for when I finally do decide to get a decent one.

EDIT: Your post:- https://forum.doom9.org/showthread.php?p=1887024#post1887024
Direct link( https://www.studio1productions.com/Articles/NVidia-GPU-Chart.htm ) :- https://www.studio1productions.com/Articles/NVidia-GPU-Chart.htm

[MeteorRain]

Where in the Nvidia Control Panel do I find that info?

FWIW, I do not consider my card to have been cheap. Has 192 CUDA cores.

You should check your Windows device manager. Under display adapters you shall find NVIDIA Geforce GTX ABC, where GTX ABC is your model.

I have a low end GTX 950 which only has 768 CUDA cores and 2GB memory, and can barely do things.

192 CUDA cores card sounds like a GT710 or 720, which is sold at $25. I'd consider the price point as "cheap". The cheapest card you can get to decode up to 4K HEVC is GTX 950 at about $80, and the cheapest card to decode 8K HEVC is GTX 1650 at about $150.

[TCmullet]

Help/System Information. You can also get that from DGIndexNV Video/List GPU Devices.

It's all relative. My card has 4362 CUDA cores. Sounds like you must have a low-end card. Something like a GTX 260?

https://www.studio1productions.com/Articles/NVidia-GPU-Chart.htm
192 is a lot more than the 16 in my other card. And the 192 runs hot and takes a LOT of power. I'd be a bit scared to have 1000 cudas unless they take a lot less juice per unit.

Uh, what am I looking for in system information. Nothing I can see about "cuda level". Make/model of my card in left column and a bunch of stats in the right.

[Groucho2004]

192 is a lot more than the 16 in my other card. And the 192 runs hot and takes a LOT of power. I'd be a bit scared to have 1000 cudas unless they take a lot less juice per unit.

Uh, what am I looking for in system information. Nothing I can see about "cuda level". Make/model of my card in left column and a bunch of stats in the right.So, what model is your card?

[TCmullet]

Quadro FX 3800. But one should not have to be familiar with makes/models in order for the Nvidia control panel to give a SPEC (of some sort) indicating what "level" of graphics capability it has. If you can infer the answer from your broad knowledge of the graphics card market, fine (and thank you). But I would, in principle, want to see the answer in the specs.

[Groucho2004]

Quadro FX 3800.No wonder it gets hot. It's still based on a 55nm process whereas the latest cards are now on 12nm (!!). Much less power consumption and therefore heat. A GTX1650/1660 would be a vast improvement with a very good price/performance ratio.

[videoh]

Uh, what am I looking for in system information. Nothing I can see about "cuda level". Make/model of my card in left column and a bunch of stats in the right. Well, "cuda level" is not a thing. What exactly are you trying to determine? What features the GPU supports? If so, see here:

https://en.wikipedia.org/wiki/Nvidia_PureVideo

Use the chart to see what feature level your GPU is at and then look at the feature level descriptions. If your card is not listed there, you'll have to search for other similar tables from nVidia and others to determine which GPU architecture your card has.

Anyway, we already know you have a low-end card that does not do 10-bit HEVC.

EDIT: I did a quick search on your card and it has a GT200B GPU which is feature set A, so it belongs in a museum not in your PC.