[Chris Wright]

-How to edit video on a slow computer in Premiere. It makes super small file sizes that can play back on slow cpu and slow hard drives. Think 400x smaller(400MB clip becomes 1MB clip!)

Easy to follow youtube video to make proxies in premiere since you can customize the video size to abnormally low quality like 320×240 in resolution(needs to be even multiple fraction size) and 1 mbps bitrate.

The real trick here is in export video, advanced video-keyframes. set it to “1”. This will make h.264 turn into an all-INTRA codec(i.e. no delta frames so your cpu doesn’t have to work at all)

If you’re working in a small window, you don’t need 4k proxies.(You can toggle off for grading.)
Additionally, other good proxies are cineform(smoother than prores if using offline grading) if you go that route and dnx proxies)
You can also set playback quality to 1/2 or 1/4.

making custom ingest proxy presets-
https://www.youtube.com/watch?v=nLn4gzcM3no

Some contents or functionalities here are not available due to your cookie preferences!

This happens because the functionality/content marked as “Google Youtube” uses cookies that you choosed to keep disabled. In order to view this content or use this functionality, please enable cookies: click here to open your cookie preferences.

[Chris Wright]

LUFS isn’t like normalization. it takes the average of the whole sequence( like a perceptual RMS averaged across several minutes worth)
It almost always changes the levels.
it’s always a trade between dynamic range and gain. there’s some plugins called loudness maximizers
that attempt to automate this. additionally, compressors average against EQ, so it’s always a good idea to get your EQ smoothed out when reducing dynamic range as most dialogue is recommended a ratio of 2:1. LUFS should always be at the end of the pipeline as it’s a polishing standard tool, not a repair tool.

[Chris Wright]

you used to use a precomp and/or solid if you were slowing down footage or it clipped off the last third of the footage. i don’t know if they fixed that but I used it extensively.

[Chris Wright]

I did find a really amazing use for white noise, to find the harmonic distortion frequencies of mics. Record white noise with the mic, then use that sample to remove learned noise. it got rid of an extra 30% noise without hurting the dialogue. i’m sure that doing some of this before you even hit record, ie. an in-lined mixer will help the sound sample’s lattitude especially if you’re not shooting 96khz 24 bit.

[Chris Wright]

is it lagging with the proxies enabled? if your cpu supports it, see if h.264 acceleration is turned on in preferences. any intraframe proxies should work well. the smaller the resolution, the less cpu overhead. also, you can disable high quality playback, right click on the program monitor. also, lower the playback resolution. cineform, dnx, prores proxies, should work well.

if you are low on hard drive space, you can create a custom h.264 with very low resolution and in advanced preset, enable “1” every keyframe(disables Delta interframes) at 1 mbps. This will make your proxies 400x smaller than normal.
I believe cineform has the lowest cpu overhead, then dnx, then prores. mxf is also really fast.

[Chris Wright]

pardon me, but did you try toggling the proxy master button in the timeline controls?

[Chris Wright]

you can verify this by right clicking in project window and enabling metadata view for “attached proxy status” Then you’ll know for sure.

[Chris Wright]

https://www.synchroarts.com/products/vocalign-pro/overview

[Chris Wright]

Wait, you’re converting VHS directly to DVD mpeg-2 low quality, then compressing again to mp4?
Then you edit the mp4 in Premiere and then export out again to H.264? That’s like 3 generations of quality loss!

These codecs don’t re-compress when editing in Premiere.
https://helpx.adobe.com/ca/premiere-pro/using/smart-rendering.html

I’d capture VHS in a non-interframe codec so there’s no motion artifacts, deinterlace with QTGMC, then create a graded master using dnxhd, prores, cineform, and use that for rendered dvd and h.264 compression. if exporting to handbrake, use RF 18 or less for quality. bitrate controls the file size, but the resolution per bitrate controls the overall quality. 2-pass may be a bit smaller.

[Chris Wright]

i was testing some auto EQ plugins to see if I could get better audio and realized that you could simply compare your audio against pink noise and EQ out the interference.
1. copy pink noise into mono left channel LUFS -23
2. copy post audio into mono right channel -23 LUFS (they both have to be the same loudness)
3. mixdown to mono to create destructive interference
4. view Frequency analysis at 256 FFT size to view interference
5. while playing back, use parametric EQ to make a flat response.

This standardized my mics.
Now, I just need to automate this.
I tested Gullfoss and Hornet Thirty-One(got closer)
but these are made for mastering, not for fixing bad audio. Manual was still better.
I also tried using RX against generated internal pink noise but RX seemed to like white noise better for its algorithm.