[PornstarsLikeitBig.com / Brazzers.com] Lichelle Marie; Andrew Andretti - A Wet Surprise [2007.11, Blonde, Big Tits, Posing, Straight, OldBrazzers, 960p, Upscale]

Стоит попробовать Dione, а то слишком пластмассово получилось

попробуем,спасибо

попробовал все модели из имеющихся в топазе в версии 3.3.0 и все шляпа
лучше протеуса(это этот ролик) нифига не вышло, но согласен - есть пластилин
может кто знает что еще попробовать?может есть еще программы какие стоящие?

Is there any chance you have the picture set for this scene? I've been looking for it for a long time. Thanks in advance regardless!

Unfortunately no dude, i searched myself and found only a few photos and they are in low resolution and as I understand it, these are official photos

Мда, с лицами лютый пиздец вышел.

Согласен. Есть тема, можно нейросетью лица дорисовывать там, где вообще невменяемый вид, но надо запариться конкретно - сделать декомпозицию на кадры и частично апскейлить. Но результат тоже не прям огонь.

Sad Mars
Не, я так понял - ты большой фанат этого конкретного ролика

Да нет, с чего бы это?
Просто ответил на коммент. Другие ролики не коментят.
Вдруг чего подскажешь по поводу исправления качества - ролики браззерсов еще есть и пока с ними хз как решать. Слишком плохая запись

Ah damn! It is unfortunate that a lot of old scenes and photosets are missing now. I will keep searching for it and I'll let you know if I find it. Thank you very much for responding!
As to the issue with the AI upscale model, I tried to do this myself before. It gave me a very large file (4GB) that I'm downloading again now to check if there is any bad artifacting or blurring. I don't know what model I used but the extension that Topaz automatically did was alq-13.
Since you mentioned that you tried a lot of models, I'm not sure this helps. In any case, this is my experience and I hope it helps! Once the download is done and I can check the quality, I'll respond.

The essence of the problem is that no matter how you clean the view with topaz, perfect detailing will not work due to a lack of initial data - for example, low quality faces in some scenes can only be redrawn (I tried to do this with a neural network, but the quality of new faces is too strange) , but it’s almost impossible to circle the contour lines (which makes topaz).
Yes, I tried the alq-13 algorithm and it looks good, removes noise. Maybe in this case, as in cases with other old videos from brazzers, this algorithm will be the best option.
I also agree with you on the lack of photosets for old videos. It's depressing, but what can we do?
It is necessary to work in brazzers, suddenly they have the source of the video and photo))

Помимо прочего, как минимум аватар.

Так совпало. Просто в этот момент я начал изучать как улучшить визуальную составляющую раздач и гиф-аватар входил в вопрос. А видео с нарезкой было под рукой.
Да и нормально вроде бы вышло, не так ли?

It makes sense that the neural network wasn't able to figure things correctly. Those old video files are probably too lossy to be able to be cleaned up like that. I wonder if using the alq-13 algorithm and then trying the detailing might help, since it probably has "more data" even if said data isn't perfect. My suggestion does sound silly though, even to me, since I think the result will still be improper faces and details.
There are some scenes and entire sites that were lost because of the credit card company issue and I don't think they'll ever come back. I know it's not the best place to ask, but since I had a Brazzers account, I asked the guys who were in Support for old stuff like Asses in Public, or King Dong (old Muffia/Reality Kings stuff) and they said that that data is mostly just stuck somewhere archived and most likely will never come back online. It's just shitty.
All that I've done is find the random images from those promotional sites and I think I found about 40 to 50, which is a fraction of the possible 1000 images. I've asked another person on Reddit and I hope he has some connections to data hoarders to see we find anything. It's probably a bust but it's worth asking, you know?
Thank you very much for your explanation and your reply! If I get any update, I'll be sure to send you a message.