Hello specialists in movie editing !
I’m looking for advice on how to preserve the quality of the sound in my work.
A while ago I did an interview with the developer Eskil Steenberg in Stockholm. Now I’m finishing this interview, which is 20 minutes long. It will be published under creative commons, very soon (I Will probably use the one that the B.F. uses for the open movies or even public domain).
I have put only my own resources to make this interview, just out of curiosity.
I worked with uncompressed .WAV sounds and already have all the footage rendered as .png images to keep the quality.
I worked in sections, now it’s time to put everything together.
So I’m looking for advice on how to keep the sound quality while rendering the master. My plan is to render those sections again, with low resolution for the movies, and good quality for the sounds. Later I will rip out the low res movies and put the high quality .png images in place, in order to obtain the final “final” rendering which will be Mpeg 4, Mpeg 2 and Mp3 (Target: youtube) .
The idea is to make sound syncing easier because right now I have many cut sound stripes per segment and I want them to became few (one per section).
But, as you may have notice until here, I’m not a movie editing specialist. I don’t know which format is better for sound preservation. I will use .MP3 for the final render to put it on youtube. So, I’m afraid about the intermediary pass: if I use .MP3 and render it again, what will Blender do? Just copy the sound to one strip or… Re render it in .MP3 lowering the quality again?
Will it blend?
Thank you for your patience on reading the whole thing.
EDIT: question added to StackExchange. If anyone else has more information, be welcome!
REPLY, By Troy_S:
"It boils down to if you are increasing the compression or not, and has nothing to do with Blender.
The reason for this is that any given level of compression will “shape” the encoded values into the compressed shape required. Subsequent compression at the same degree using the same compression technique seeks the exact same shape. Given that the algorithm is seeking the shape the data is already in, very little further degradation will occur typically.
You can test this by trying to encode a JPEG image twice at the same compression level. What you should observe is very little encoded data value shift, and even possibly a growth in file size.
As a general rule, this is a poor option for a pipeline, and it would be much more prudent to keep your interchange data in a lossless and uncompressed format, and encode once and only once. Even expanding an MP3 to a WAV can help you here in theory."