Hey Fredo,
I was reading, with great interest, the series of posts you placed on the Nuendo Users group on Facebook, regarding issues with compressed video files and the resulting sync issues. You said that if we can’t get uncompressed files from the Video Editors and, instead, receive MP4 files we should convert them ourselves to Prores format, in order to guarantee the correct sync.
My question is, if you convert the file are you actually restoring it to it’s full value or just opening the compressed file to a larger file with it’s losses still intact? The reason I’m asking is because:
I don’t know squat about video editing and feel, as a media composer, I should be better informed about this kind of issue and,
I converted a short MP4 file to a Prores format and the picture looked a lot grainier to me. I was expecting a much sharper image as a result of the conversion (the file went from 199MB to 1.23GB!).
Consequently, I’ve managed to confuse myself quite a bit!
I totally get the sync issue. But am I wrong to expect a sharper picture? Is the larger file now too large for mere PC Video Cards to translate? Do I need a BMD Decklink card and 4K Monitor to get the full effect?
I’ll try to keep it as simple as possible, but feel free to ask for more details if you feel it’s not enough.
Basically, H264 is far more advanced than any of the codecs we consider “good” for DAW playback. That is why it looks far crisper while keeping the file size considerably small, and its professional name is AVC (Advanced Video Coding).
Because H264 is far more advanced, it also demands far more real-time system resources. Each time you press play or, even worse, scroll/scrab through the timeline, it pushes the system into extreme mathematical calculations, that happens because the most popular H264 encoding is done with what is known as GOP (Group Of Pictures) using IPB, i.e., i-frame p-frame b-frame, the i-frame is a complete video frame (full images). The p-frame (also known as Anchor Frame) and b-frame only hold data that tells the decoder what has been changed from the last I-frame. So a group of pictures may look like so: IBBPBBPBBPBBI but may easily look more like this: IBBBBPBBBBPBBBBI, imagine how many calculations are required just to present the 15th frame in the above sequence (Group Of Pictures).
That’s where ProRes and DNxHD(HR) come into the picture, the industry understood quickly that AVC (Advanced Video Coding) is great as an end product but horrible for editing work as it draws a lot of real-time system resources at the expense of more important editorial needs which leads to audio/video sync problems and stuttering video playback, so both Apple and Avid created their own standards, ProRes by Apple and DNx by Avid. Both codecs are almost the same in their essence, they use what is known as Intra frame compression so they only have i-frames, which means the compression is by far simpler and easily decoded but it uses significantly more disk space to present good quality video.
That’s why they created several “flavors” for their codecs to fit all industry needs:
ProRes Proxy, ProRes LT, ProRes 422, ProRes HQ, ProRes 4444, and there’s more.
Avid DNx has an even larger flavors set but we’ll not get into that.
Professional video editors who rely solely on the picture work with ProRes Proxy or DNxHD 36 through the entire editing process and only render the final product in full quality (H264/ProRes HQ/DNxHR HQX).
So I think it’s safe to say that audio producers of all types can and should work with a little unsharp image to increase their DAW productivity even if it means much more disk space and less enjoyable video reference, that being said, if you have a high-performance SSD and willing to spend the disk space required, you can always work with ProRes 422 or DNxHD 185 for a clear sharp image experience while keeping your system resources free for audio plugins, virtual instruments and other important things we use to produce our product and desperately need the real-time processing free from unnecessary interruptions.
There are other important issues with H264 and audio to video sync which are part of the H264 compromise to produce a high-quality video with minimum file size, but that’s a whole different subject.
You should just remember to never trust lip-sync when the video ever was or is H264.
Unfortunately, no DAW has a decent algorithm to properly correct the damage created by any of the H264 encoders even in 2021.
It’s well worth playing around and finding render settings that work well on your system and also work well for the task. I recently did a few batches remixing replaced stock music for a TV series so I was working with stems for everything except new music. Since that was the case I didn’t worry about sync at all as far as lip-sync goes, and it was rare I moved audio and had to worry about picture cuts. For that job I picked DNxHD (Windows) 8-bit, 720p, and a pretty low bitrate. There were so many episodes I needed to get through that I didn’t want a huge amount of video data on my video drive for no good reason.
But for stuff where I need lip-synch clearly visible I do no less than HD of course, and with a fairly hefty bitrate.
I have to say that on my modest ‘recent’ system playback of DNxHD in Windows in Nuendo is smooth as butter. I love it. The system I use in a couple of studios where I work running Macs and PT just blows in comparison.
I made a ProRes 422 copy of that same MP4 file. I copied it over to my SSD drive and played it back through Vegas Pro 16. It looked great on the 4" screen but still blurry on the 27" HP screen. Since you said NO DAW would ever be able to display a really sharp image, I thought I’d have better luck with a Video Editor App. I guess Vegas isn’t really “Pro Grade?”
Maybe you should look into the BMG intensity pro 4K, I like it and use only that for decades (before it went into 4K), but keep in mind that the Nuendo video engine is still unable to playback any 4k video, only FHD…
