Some chose a lower RF for HD video than SD. Why?

[mas5acre]

I've seen a lot of people over many different forums choose a lower RF (like say 18) for HD video and a higher RF for SD video like 20.

This always seemed backwards to me. I always saw it as the crappier the source, the more the quality you want to preserve. The better the source the more you can afford to lose.

I actually do think a majority of people do over bitrate their encodes. Though, sometimes I can understand using a high RF of 18 on the appletv 1 and 2 preset because you are throwing away some detail since appletv 1 and 2 can only do 720p and you are trying to retain some extra detail, and might as well since file size is smaller for 720p vs. 1080p.

Thoughts?

[JohnAStebbins]

We tend to agree withe you here. If you do a little search of the forum, you will find we always recommend a higher RF for HD. I use 22 for HD and 19 for SD myself.

I think most people who recommend lower RF are just misguided. A rare few might know what they are doing and use the lower value for *archival* purposes. I.E. they can't see the extra detail, but they know it is there and it is useful if you plan to make copies of the copy.

[BradleyS]

I'll add that viewing context is paramount. The larger your display, the more apparent artifacts will be at normal viewing distances. The following assumes local playback on wired connections where storage and bandwidth are plentiful.

Way long ago before RF was a thing, I encoded SD at what would now be about RF 21. I mostly viewed on a 20" or 27" CRT. Artifacts were mostly too small to be visible, and the analog nature of CRT further dampened any digital quality issues.

At some point, I got a 40" LCD. With some testing I found that RF 19-20 worked well for SD, but I decided to go with pristine 18 to avoid worrying about it. I'm a stickler for quality and storage is cheap. About the same time, I began encoding HD at RF ~22, give or take +-1. At normal viewing distances, RF 21 was pretty excellent for HD.

Then I upgraded to a 55" plasma (RIP Panasonic tech). My SD was mostly fine since I was very conservative in my last revision. However, my HD was not quite as solid. Most people wouldn't notice the difference, but I have a trained eye and had just bought a really nice display--I wanted the best. So I upped my HD encodes to RF 19 for tune:film and 18 for tune:grain. RF 16 for tune:animation and select SD titles that need it. Really. Again, storage is cheap, and my hardware plays it fine. This is where I'm at now.

With 4K sets emerging, I'm pretty confident that my recent encodes will be fine, judging by past experience and the fact that my encodes look nearly identical to their sources zoomed in to 200%.

A rare few might know what they are doing and use the lower value for *archival* purposes. I.E. they can't see the extra detail, but they know it is there and it is useful if you plan to make copies of the copy.

For archival purposes (assuming you want to transcode and not remux), my advice is to investigate what RF you believe is a comfortable level of quality for your current equipment, then consider lowering the RF by -1 (adding about +12.5% bitrate) or even -2, depending on your starting point, if you're concerned about future investments necessitating reencoding. The only caveat to this approach (again, assuming local transport) is extended encoding time.

[Deleted User 13735]

The vernacular usage notwithstanding, "archive" means "not lossy" in my corner of the business.

[BradleyS]

Indeed. I'd remux if hardware like AppleTV better supported sliced encoding employed by modern content. Closer, but still not lossless. I certainly keep a lossless copy of all my personal projects.

[Deleted User 13735]

That's two copies for me.
One archive on a hard drive that I spin up once a year, and another playback version on my movie drive.