Encoding using Constant Rate Frame (CRF)

[fatammag]

Hello,

I have recorded some Standard Definition TV episodes via a TV tuner and I’m using VideoReDo TV Suite 5 to remove the commercials.
The output format is a .ts file.

Now, I need to convert the TS file into x264.mkv. I’m using RipBot264 for that. I prefer encoding using Constant Rate Factor (CRF) compared to 2-pass encoding. The CRF 22 value is giving me a video bitrate about 2500 kbps and the total size is about 950 MB.

Is there a way to achieve a proper encoding using CRF and the file size kept bellow 450-500MB for about 45 minutes?

Thank you!

[theewizard]

Yes use a smaller resolution
Cut resolution down the same CRF setting will produce a smaller size

It is not an option I like, I do not do this
I encode for quality, and the size is what ever it is

[Dougster]

I don't use RipBot264 myself but CRF means Constant Rate Factor not Constant Rate Frame. If you want a specific file size the you should use a
2 pass average bit rate to achieve that. CRF is also referred as Constant Quality by many applications.

[fatammag]

Yes use a smaller resolution
Cut resolution down the same card setting will produce a smaller size

It is not an option I like, I do not do this
I encode for quality, and the size is what ever it is

The source’s Aspect Ratio is 1.333 – The width and height I chose is 720x540. I need a video bitrate not greater than 1700-1800 kbps. Source’s video bitrate is 4761 kbps.
CRF 22 is giving me 2500 kbps video bitrate as I mention in the first post. I think if I use CRF 24 or 25 the video bitrate will be fine. But, is that correct?
I forget to mention: Most frames in the video has a lot of light.

I don't use RipBot264 myself but CRF means Constant Rate Factor not Constant Rate Frame. If you want a specific file size the you should use a
2 pass average bit rate to achieve that. CRF is also referred as Constant Quality by many applications.

That's correct. I meant Contant Rate Factor.
If I can't figure it out with CRF I make a rip using 2-pass.

2-pass takes a long time for encoding. CRF is about the half time.

[jagabo]

CRF gives you a known quality, but you don't know what the bitrate will be. 2-pass VBR gives you a known bitrate (and hence, file size), but you don't know what the quality will be.

You may be getting 2500 kbps at CRF=22 with this particular video. But the next may give you 1200 kbps, or 5000 kbps, depending on the content. Bright, high contrast video requires more bitrate. Noisy video requires more bitrate. Quick action, flickering lights and flames, billowing smoke, wavy water -- all those things require more bitrate. On the other hand, a slide show with no motion, no noise, and no fancy transitions will compress down to nearly nothing without losing quality.

[fatammag]

CRF gives you a known quality, but you don't know what the bitrate will be. 2-pass VBR gives you a known bitrate (and hence, file size), but you don't know what the quality will be.

You may be getting 2500 kbps at CRF=22 with this particular video. But the next may give you 1200 kbps, or 5000 kbps, depending on the content. Bright, high contrast video requires more bitrate. Noisy video requires more bitrate. Quick action, flickering lights and flames, billowing smoke, wavy water -- all those things require more bitrate. On the other hand, a slide show with no motion, no noise, and no fancy transitions will compress down to nearly nothing without losing quality.

I know that,
But I can't find a way to calculate the video bitrate before. So, I need to try different CRF values each time.

I have 53 TV episodes. I'll use the 2-pass encoding.

Thank you all for your help!