In this "stay at home" Covid-19 era, I am working in VPX while my primary collaborators have Adobe Premiere Pro.
We have been exchanging files back and forth, exporting as MP4, full HD (1080i, 1920x1080, 29.97), via h.264.
Is this optimal?
Is this approach likely degrading our video through multiple back-and-forths? (i.e. Are 3 round-trips worse than 1, or the same?)
Is there a better, recommended approach?
Like, can Premiere accept AVCHD Transport Stream?
Thank you so much!
Hi
. . . . We have been exchanging files back and forth, exporting as MP4, full HD (1080i, 1920x1080, 29.97), via h.264. Is this optimal? . . . .
That depends on what the edit/render cycle is - these files are a compressed format and every time you re-render them you lose a bit of quality.
Ideally you should be exchanging in an uncompressed format, eg AVI, however this is often not practical due to the massive files sizes ~ 150 - 250 MB per second of video.
AVCHD files are compressed more than h.264. Premiere can read them depending on the version.
As with many things , this is a balancing act and the acceptable final product is your and your collabs final decision on what is acceptable for the final product.
HTH
John EB
That depends on what the edit/render cycle is - these files are a compressed format and every time you re-render them you lose a bit of quality.
This gets to the heart of what I wanted confirmation on.
I'm an audio guy. I know what I'm dealing with when I compress a .wav file to .mp3; or re-compress an .mp3 to .aac or vice versa (which can sound truly terrible).
What I've found encouraging on the audio side of things is that exporting an existing .mp3 audio file to .mp3 does NOT recompress it. It remains the same. I believe this to also be true with .mp4/.aac audio files. I was hoping that MP4 video exported as MP4 video would similarly not take an additional quality hit, if the bitrate was the same.
Wondering if the difference is:
- In Audio software, I can work with either the actual .mp3/.aac file OR bring it in as a .wav file, and when it exports back to .mp3/.aac it does the exact same compression as the original process (unless I've added further audio processing, obviously).
- In Video software, I assume now that VPX, like FCPX and Premiere Pro, they are actually first converting the MP4 info to a proprietary program format, and then re-exporting via an additional conversion process (so the MP4 is changed twice each time, rather than either remaining the same or being re-encoded identically). Thus, an additional quality hit each round trip.
Does this sound correct?
I wonder if anyone has done a controlled test on this to determine a "rule of thumb" (i.e. 1 RT - undetectible, 2RT - slightly noticeable in contrast or shadowing, 3RT - color fading, action begins to blur, 4RT - unrecommended; or something of this sort. I'd be VERY interested in such data or known/established info., if it exists.
Hi
IIRC mp3 compression is a constant compression technique the only thing that changes the compression ratio is the bitrate - which is not as complex like video which requires more complex techniques then audio to maintain quality.
. . . . VPX . . . . they are actually first converting the MP4 info to a proprietary program format . . .
Only if you are using proxy files, however these are only to improve timeline playback.
The proxy files are not used for rendering, the original clips are used, they are in effect imported / decoded to create a cmplete image per frame of video, manipulated if you have any effects and then encoded again into the new file.
There is a quality hit on each round trip.
. . . . I wonder if anyone has done a controlled test on this to determine a "rule of thumb" . . . .
A general rule of thumb would be as few times as possible - however with round robin trips tthe 'rule' depends on the final destination, video resolution and acceptable limits of quality degredation you can accept there are too many variables to generate -a 'general rule of thumb'.
In your situation the ideal process would be to change the workflow so as not involve any round trips, eg one person does all the video editing using the original video clips from the collabs.
HTH
John EB
No I don't think you are correct on a second encoding of an MP3 file not producing a deteriorating second copy. If you keep recompressing that format there will be a general eventual deterioration of the sound quality not only effecting frequencies with phase complications but an increase in floor noise levels. Each pass is a rounding off of varying bit depths. It's the same with any format that is not lossless. Wav. is a direct copy of the waveform if bit depths and sampling frequencies are matched.
The same is true of digital still images or video compressed formats.
For digital still images there are two ways to capture a non-compressed image within the camera if the camera supports the file formats but there are still compromises with both file types.
One is the tiff format and only normally found on higher end digital cameras. They can produce, depending on just how high end the camera is, 14bit or 16bit images as taken directly from the sensor with each photosite's recorded photon capture level recorded and placed into it's corresponding place within the image. This results in what looks like a noisy image that is also very soft looking with no white balance, noise reduction, sharpening, or other alterations that the cameras' second processing unit dedicated to those tasks would normally clean up. The advantage to those who are good enough at editing is the wider tonal range you can get using greater bit depths. But as most sensors and their corresponding DACs work in 16bit, those cameras that can only give a 14 bit output will lose out quite a lot in that respect.
The second file type is DNG or Digital NeGative (More commonly referred to as raw files) is not an image file at all but an instruction set that within an editor capable of reading it can reproduce an image from the instructions within the file. If you have the corresponding codec on your system you can view thumbnails within windows, but they are not the image. They are smaller embedded jpgs within the file to show up within folders for easier identification.
The advantage of a DNG file is it also records all the camera settings at the time the shot was taken, including white balance, sharpening, noise reduction and so forth, so you can at least have a starting point. Any converting or altering of such a file is not permanent and can be reworked as and when needed. The converted output to a useable file format will be slightly compressed though compared to what you will see in the editors window if your monitor has a high enough bit depth of reproduction that few if any have.
When you use a video editor, as far as I'm aware, the file is loaded 'as is' and only loses any quality on export as it also works on instruction sets as far as I'm aware. This is why there is a list of useable codecs on every NLEs specification sheet.
As john said, there are lossless file types you can choose for export, but the resultant file sizes as much larger than the original files unless the original files came directly from a camera sensor and encoded into a digital raw format that MEP at least can't handle.
I agree with John's summary.