I must have found my workflow for my SD videos! it's a gift for those who are interested :)

Hello! I post my original vs upscale comparison of my SD source in HD, I have better results than before which was already not bad and in the end I feel like I gain a lot of detail, without really having any artifacts and noise, I still apply neat video at the end with vegas pro but I have no loss of detail because on my first pass, I increase the detail a lot.
(thanks to those who advised me to produce a pass in hd with artemis then use iris afterwards in hd)

Here is how I proceed:

.
πŸ‘† ***** HYBRID *****
.

  • Resize (square pixel 768 x 576 BICUBIC)

  • Deinterlacing with gtgmc, important: EZkeepgrain: 8 (it allows to bring up the details even if the noise is amplified, the fact of using artemis hight and iris thereafter, gives better results)

  • Colors (thanks to β€œHDFactory” for putting me on the right track to recover some details in the dark areas)

  • Remove grain: 24

  • Dering: hq dering

  • Source SD 720 x 576 (1.0926 pixel) exported in ffv1 lossless
    (768 x 576, square pixel)

.
πŸ‘† ***** VAI *****
.

  • Export FFV1 artemis hight (recover original details: 100%) HD

  • Export ffv1 iris HD

.
πŸ‘† ***** SHUTTER ENCODER *****
.

I use it because my videos in ffv1 do not open with vegas pro so
Since it doesn’t work converting again to ffv1, I choose the prores, I don’t do it directly with VAI because I apply the rec 709 color setting and I convert the levels to 0-255 again. I thought it corrected the colors better because apparently going to convert back to 16-235 if I understood correctly. Unless it is the fact of exporting in prores with VAI which does not give good colors, I seem to have read that people complained about that. anyway, there are always ways to correct the colors even if it requires an export in addition :slight_smile:

.
πŸ‘† ***** VEGAS PRO *****
.

The filters I use in order:

** Gaussian Blur **
.
.


.
.

  • Export (not yet seen for export but probably in uncompressed avi and convert with another software because vegas pro has a reputation for bad encoding.

==================
⇩⇩⇩⇩⇩⇩⇩⇩⇩⇩⇩⇩⇩⇩⇩⇩

here is the comparison of two different scenes in three images.
The first, the original, the second, the improved version without final grain and the third, improved with final grain





Well, if you have any questions, I’ll try to answer haha, I’m really happy with these results because I have a really minimal loss of detail, in most images all the original noise is removed, the artifacts too, and there are more fine details. I will even be able to produce my videos without adding grain so much detail is preserved, but I prefer the final grain which gives the impression of having more detail, which can hide minimal artifacts and which gives a cinematic look, for our old family videos, it looks more professional I think :slight_smile:

And finally, here is the link to the video:

And thank you very much to the topaz team who are constantly improving their software and the users who have helped me a lot sometimes! :slight_smile: force to you for your projects :slight_smile:

:grin::v:

β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β–‘β–ˆβ–ˆβ–ˆβ–‘
β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β–‘β–ˆβ–‘β–‘β–‘β–ˆβ–‘
β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β–‘β–‘β–‘β”€β”€β”€β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β”€β”€β”€β”€β”€β”€β”€β”€β”€β–‘β–ˆβ–ˆβ–ˆβ–‘β”€β”€β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β”€β”€β”€β”€β”€β”€β”€β–‘β–ˆβ–ˆβ–‘β–‘β–‘β–ˆβ–ˆβ–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β”€β”€β”€β”€β”€β”€β–‘β–ˆβ–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–ˆβ–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β”€β”€β”€β”€β–‘β–ˆβ–ˆβ–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–ˆβ–‘
β”€β”€β”€β–‘β–ˆβ–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–ˆβ–‘β–‘β–‘β–ˆβ–‘
β”€β”€β–‘β–ˆβ–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–ˆβ–‘
β”€β”€β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–ˆβ–‘
β”€β”€β–‘β–ˆβ–‘β–‘β–ˆβ–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–ˆβ–‘
β”€β–‘β–ˆβ–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–ˆβ–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–ˆβ–‘
β”€β–‘β–ˆβ–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–ˆβ–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–ˆβ–‘
β”€β–‘β–ˆβ–‘β–ˆβ–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–ˆβ–ˆβ–‘β–‘β–‘β–‘β–ˆβ–‘
β–‘β–ˆβ–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–ˆβ–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β–‘β–ˆβ–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β–‘β–ˆβ–‘β–ˆβ–‘β–‘β–‘β–‘β–ˆβ–ˆβ–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β”€β–‘β–ˆβ–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–ˆβ–ˆβ–ˆβ–‘β–‘β–‘β–‘β–ˆβ–ˆβ–‘
β”€β–‘β–ˆβ–‘β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–ˆβ–ˆβ–‘β–ˆβ–‘
β”€β”€β–‘β–ˆβ–‘β–‘β–ˆβ–ˆβ–‘β–‘β–‘β–ˆβ–ˆβ–‘β–‘β–ˆβ–‘β–‘β–‘β–ˆβ–‘
β”€β”€β”€β–‘β–ˆβ–ˆβ–‘β–‘β–ˆβ–ˆβ–ˆβ–‘β–‘β–ˆβ–ˆβ–‘β–ˆβ–‘β–‘β–ˆβ–‘
β”€β”€β”€β”€β–‘β–ˆβ–ˆβ–‘β–‘β–‘β–ˆβ–ˆβ–ˆβ–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–ˆβ–‘
β”€β”€β”€β”€β”€β”€β–‘β–ˆβ–ˆβ–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–ˆβ–‘
β”€β”€β”€β”€β”€β”€β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–‘β–‘β–‘β–ˆβ–‘
β”€β”€β”€β”€β”€β”€β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β”€β”€β”€β”€β”€β”€β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β”€β”€β”€β”€β”€β”€β–‘β–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β–ˆβ–ˆβ–ˆβ–ˆβ”€β”€β–‘β–ˆβ–‘β–ˆβ–ˆβ–ˆβ–ˆβ–‘β–‘β–‘β–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β–ˆβ”€β”€β–ˆβ”€β”€β–ˆβ–ˆβ–ˆβ–ˆβ”€β”€β–ˆβ–ˆβ–ˆβ–ˆβ–‘β–‘β–‘β–‘β–‘β–ˆβ–‘
β–ˆβ”€β”€β–ˆβ”€β”€β–ˆβ”€β”€β–ˆβ”€β”€β–ˆβ”€β”€β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ
β–ˆβ”€β”€β–ˆβ”€β”€β–ˆβ–ˆβ–ˆβ–ˆβ”€β”€β–ˆβ”€β”€β–ˆβ”€β”€β”€β”€β”€β”€β–ˆ
β–ˆβ”€β”€β–ˆβ”€β”€β–ˆβ”€β”€β–ˆβ”€β”€β”€β”€β–ˆβ–ˆβ”€β”€β–ˆβ–ˆβ”€β”€β–ˆ
β–ˆβ”€β”€β–ˆβ–ˆβ–ˆβ–ˆβ”€β”€β–ˆβ”€β”€β–ˆβ”€β”€β–ˆβ”€β”€β”€β”€β”€β”€β–ˆ
β–ˆβ”€β”€β”€β”€β”€β–ˆβ”€β”€β–ˆβ”€β”€β–ˆβ”€β”€β–ˆβ”€β”€β–ˆβ–ˆβ–ˆβ–ˆβ–ˆ
β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ”€β”€β–ˆβ–ˆβ–ˆβ–ˆβ”€β”€β–ˆβ”€β”€β”€β”€β”€β”€β–ˆ
β”€β”€β”€β”€β”€β”€β–ˆβ–ˆβ–ˆβ–ˆβ”€β”€β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ

Edit: we can remove the shutter encoder phase and export with vai in png 8 bit, it does not remove details and we can import a sequence of images with Vegas Pro, after maybe DaVinci Resolve or others can import AVI FFV1 files but you need the same filters that I use with Vegas on Resolve or other

7 Likes

Hallo, ssbroly, thanks for your detailed workflow for Hybrid. I followed exactly all steps, but Hybrid converts only 1-2 frames. With default settings however, it works. Could you help here please?

Hello, 1 to 2 images? In fact I recently discovered that you can export to PNG with Hybrid, if you’re talking about export. It’s the same quality as the ffv1 almost pixel-perfect after comparison. But the image sequences open with everything, including vegas pro. :wink:

Thank you for a quick answer. I wanted to say that something is wrong with the settings or I do something wrong. I insert a clip that is 1 min long, but the output file lasts 0 second (has only 1-2 frames in it). Playing around with Hybrid I changed β€œvapourtsynth” to β€œavisynth” and it suddenly worked! But there is no option EZkeepgrain: 8. I hope I explained my problem clearly. Could you help please with this problem? Thank you very much in advance

in the second photo, ezkeepgrain is in the qtgmc settings

Personally, I like the one without the grain the most. I don’t understand why people like grain.

the one without the grain? what are you talking about ?

What am I talking about? Your examples/results. They are good, but I personally like the one without the added grain the most. That is all I am saying.

I had the same opinion as you when I started, but the final grain is more cinematic and more professional as a result. and then the grain adds detail, visually reduces imperfections and removes the β€œplastic” effect. quite simply :slight_smile:

I have found ways to work the media without grain so that it doesn’t have that plastic/cartoony look. Been working great after doing 600+ files so far. :grin:

if you don’t have any grain, you definitely have artifacts, and if you really don’t have any artifacts, you have a plastic look. impossible to keep all the details of an old SD video without any grain or artifacts

I may get some artifacts here and there, but I can fix them with processing after topaz. You can have clear crisp restored old videos without grain and without the plastic look. I have done it hundreds of times now. :slight_smile:

1 Like

can I see an original/restored comparison? do you have a youtube channel?

+1 ! nice work , i arranged your workflow for my needs, and the result is marvelous !

thanks !

1 Like

great! :slight_smile: I’m experimenting with another workflow similar to this one for even better quality :slight_smile:

on my side i could not go with the Vegas part :

I followed your reco about Hybrid (cried a bit as i was not familiar with it),
then i went to TVAI, but directly with FHD upscale, and Iris Manual settings as below

the output is set as H265 Main , Constant bitrate , @4Mb/s, as mp4.

it is the best result i got so far since months and tries, also compatible with animes and movies, orginally as 720p interlaced/telecined!

so here my 50cts by sharing this alternative. of course, I think Vegas postprod should improve dark and color quality, but i was at least happy with that result on 576p and 720p .

I have just one question : i repeated your action of setting width 768 in hybrid, but i did not get your reason. Fortunatly, i did not see any bad behavior by following the config.

if you export in png or ffv1, it will be lossless and you will have even more details, unless you don’t want to make any more improvements behind then I understand for h265. with hybrid I export in 768x576 because that is the size by defining the square pixel, or 720x540 otherwise. because I read in different places on the forum that it is advisable to send a video in square pixel but personally I don’t really see any differences between 720x576 and 768x575… on the other hand to go from sd to fhd, Your video really needs to have no artifacts, you don’t have any unreal effects in the final video?

I prefer using this settings as it is a perfect resolution of 1440x1080 when UPSCL at x2…

1 Like

yep , i have some, but really really less than using proteus :
in fact the great improvement is the hybrid part with the settings you recommended to deinterlace, lossless and color .
also brightness is a bit increased during this step, and it seems tvai is able to detect more details , and then the result is better .

yes, it is perfectible , but far better than before.

up to keep improving !

regarding the h265 , yep it was for final result.
i can switch to ffv1 for more post treatment once i found what i can do in next steps

yes I was talking about ezkeepgrain on 0.20, it reduces the accentuated details but ultimately we don’t lose any. setting ezkeepgrain low, there is less noise, so ultimately better for Vai models.

shit instead of answering you I modified my message from before lol