Editing video and audio - Degradation

[shans]

Though the following topics were discussed in this forum, I have doubts about these now and then. So, I want to get it cleared once for all.

1) Video of any format will lose its quality when it is re-encoded even if the video bit rate is increased. To put it in detail, a video is decoded to lossless format, increased the video bit rate and re-encoded to the same format or different format, re-sized to higher resolution. By doing so, the quality can't be enhanced.

2) Similarly, audio file is converted to wave format and re-converted to same or different format increasing the bit rate. Even then, the quality of the audio can't be enhanced.

I think I have made my doubts clear. Appreciate comments.

[poisondeathray]

Though the following topics were discussed in this forum, I have doubts about these now and then. So, I want to get it cleared once for all.

1) Video of any format will lose its quality when it is re-encoded even if the video bit rate is increased. To put it in detail, a video is decoded to lossless format, increased the video bit rate and re-encoded to the same format or different format, re-sized to higher resolution. By doing so, the quality can't be enhanced.

2) Similarly, audio file is converted to wave format and re-converted to same or different format increasing the bit rate. Even then, the quality of the audio can't be enhanced.

I think I have made my doubts clear. Appreciate comments.

Yes - compression wise - you cannot objectively increase the "quality" compared to the original for audio or video . You cannot get better than what you started with, no matter the bitrate, no matter lossless or lossy


But you can subjectively increase the quality by performing other operations or filtering. E.g. if video is blurry you can sharpen. If video is too dark , you can brighten etc... If video is too noisy, you can denoise it. If speech frequencies are too low, you can enhance them etc... You might be able to subjectively do a better job of upscaling than your software or hardware player.

"Quality" is a difficult thing to describe or measure, but objectively, by metrics, you cannot get better than you started with

[Cornucopia]

To add to what pdr said:
You can increase certain aspects of the signal, in order to get a result that is subjectively improved for the target audience, but this does come at a cost, and other things might/will decrease in the result, quality-wise. Brightness might increase, but overall dynamic range may decrease as a result. Or, perceived sharpness might increase, but noise & halos might increase as well (making it harder to compress).

After "creation" or capture, it's always a level or downhill progression in overall quality. Which is why you want to start with the very best possible, and do as little damage as possible.

Note: there are lossless and lossy (compression) FORMATS and there are lossless and lossy PROCESSES. You want to manage both sets to arrive at minimal loss overall.


Scott

[jagabo]

There is an exception to the "will lose its quality when it is re-encoded" rule: when using lossless codecs. The losses video will be exactly the same quality as the source. The problems is that lossless codecs (huffyuv, lagarith, UT video codec, x264 lossless, etc.) are very inefficient compared to the lossy codecs (MPEG 2, h.264, h.265, etc.). If you start with a lossy compressed video and recompress it with a lossless codec your final file will likely be 10 or more times larger. So it's a waste of time unless you have some special need (like an editor that can't handle your lossy compressed source).

[shans]

Yes, I understood that it is very difficult to define the quality precisely under various scenarios. Thanks @poisondeathray @Scott @ jagabo