(Outdated) Britney Spears - ...Baby One More Time (Perfect AI Remaster, 1080p x264 AAC)

1. QTGMC does reduce video quality with default settings because of its de-noising options, but I didn't use it as a de-interlacer, I used it a post-processor, to do additional clean up. I used TFM to IVTC, which caused no detail loss, and QTGMC as cleanup. I compared the screenshots, there was not loss, in fact, all it did was increase detail on moving parts of the image. It did cause detail loss on scenes that had bad aliasing, but the AI kind of brings back that sharpness, and you can't remove the aliasing without some detail loss anyway.

2. Most video players (like VLC or mobile video players) don't use a high quality resizing algorithm, such as Lanzcos, or in this case, NNEDI3. If I resize to proper AR beforehand, the player dosen't have to do any (possibly subpar) resizing.

3. I would be very surprised if you can tell the difference between iTunes encoded AAC's files vs. WAVs. I consider myself an "audio quality enthusiast", and i cant tell the difference in blind tests. I find that AAC works perfect to keep file size down.

 

Thank you! @embis2003 could you list the programs you used and in which order? Where would one learn to do all this stuff? I would love to learn

 

Thank you!

 

thank you very much

 

gracias por el video

 

thanks

 

thank you

 

let's see

 

thanks

 

thanksssssss

 

thanks

 

thanks

 

Thanks!

 

thanks

 

https://cloud.mail.ru/public/Asyn/6BveYAF75

 

Thanks

 

Thanks

 

thanks



***Hidden content cannot be quoted.***

There is 4k upscales, yes, but they have a lot of artifacts. This was VERY involved, however, in the end I think it turned out nearly perfect. Started out with the telicine VOB file (the one without the color ghosting), and removed combing using AviSynth's TFM. Used parameter "pp=5, slow=2" for some additional cleanup, removing some of the harsh aliasing and other field issues. Elected to not decimate due to video being mastered at 29.97, and some transitions would be messed up, also decimation, to my eye, just causes this video to be more jumpy because it doesn't remove all the correct frames. Used QTGMC(inputtype=1) for even more clean up, which REALLY helped out the overall detail levels and corrected some more aliasing and shaking artifacts. It seems inputtype preserves NEARLY ALL DETAIL and cleans way more artifacts then simply using it as opposed to TFM(). No harsh denoisers were used and almost all fidelity is preserved. I then exported out in SD, and loaded into the AI software for noise reduction and deblocking ONLY because actually scaling up using it seems to cause artifacts. What came out was stunning, it looked like true prores. I scaled up using NNEDI3 by a power of 2, getting rid of nearly all aliasing, and corrected AR using Lanczos . I then fixed up the audio, adding CD quality, and encoded to x264 2-Pass 8000kbps 10bit Slower and QAAC 256kbps. Pretty sure it doesn't get much better quality than this.[/QUOTE]

 

Gracias

 

Thanks

 

tnks bro.... ...