60i to 24P dvd. Where is the 20% savings?

[Mike Kujbida]

I just tried on an 80 min. video using the DVD Architect 24p NTSC Widescreen video stream template (23.976 + 2-3 pulldown).
When I rendered this out using a custom DVD Architect NTSC video stream template (VBR of 8,000,000 7,200,000, 3,600,000), the file size was 4.03 GB (4,333,539,328 bytes).
Using the 24p template and the same VBR settings, it came out to 3.37 GB (3,623,215,104 bytes).
Assuming my math is correct, this is a 16% file size reduction.

BTW, where do you get a 24P progressive setting setting in DVDA?
I fed it that file and it said OK and rendered it.

[George Wing]

[Mike Kujbida] “When I rendered this out using a custom DVD Architect NTSC video stream template (VBR of 8,000,000 7,200,000, 3,600,000), the file size was 4.03 GB (4,333,539,328 bytes).
Using the 24p template and the same VBR settings, it came out to 3.37 GB (3,623,215,104 bytes).
Assuming my math is correct, this is a 16% file size reduction.”

Those results are very interesting — as the filesizes (using the same bitrates) should be about the same. It would appear that the 24p render DID NOT maintain the stated AVERAGE bitrate of 7,200,000bps. Perhaps using the 2-pass option would give closer results (as the first pass does the analysis, and the 2nd pass encodes accordingly). Or if you use the same CBR Bitrate (DVDA and DVDA 24p Templates), you should get files that are about the same size.

The bottom line is that in order to reduce size, you have to lower your encoding bitrate (either CBR, or AVERAGE VBR).

The 24P allows the encoder to allocate the bits-per-second to less frames (24 vs. 30), and so that’s why you might get better looking results (using the same bitrates).

It’s similar to folks sometimes getting confused by thinking that lowering resolution to 352×480 or 352×240 will produce smaller files — but that is NOT true either. In other words, if you use the same bitrates to encode a 60-minute video at 720×480, and a 60-minute video at 352×480 or 352×240, you will end up with about equal filesizes.

Regards,
George

[Mike Kujbida]

“It would appear that the 24p render DID NOT maintain the stated AVERAGE bitrate of 7,200,000bps.”

I have no reason to think that it didn’t do that.

Perhaps using the 2-pass option…

I’ll do a few more renders tonight trying the various options you suggested and post the results here.

[Terje A. bergesen]

As George is pointing out, the size of the rendered file is independent of framerate and most other things. The only thing that matters is the length of your movie and the bitrate. Since it is bps(econd) and not bpf(rame), the file size should be more or less identical for 60i and 24p material.

The 24p material should be of higher quality at the same bitrate however, since you have more bits per frame available.

In theory, if you encode 24fps at 20% lower bitrate than 60i 24fps, you should get similar quality. Haven’t tried this though.

[Kert]

Thank you all for responding.

The same VOB file size for 60i and 24P now makes sense.
In fact when I originally compared 2 1/2 hour footage I rendered at 60i and then at 24P, I noticed that the 24P footage seemed higher quality, but I was not sure. Now I know.

Another way of looking at it is that if you have say 2 1/2 hours DV footage to render and you set 3.8 MBS compression, one gets a useable quality DVD. But with 24P with the same 3.8 MBS compression one can put on DVD approx. 3 hours of video with the same quality as the 2 1/2 hour at 60i.

Thanks again, John Kert

[Kert]

Another way of looking at it is that if you have say 2 1/2 hours DV footage to render and you set 3.8 MBS compression, one gets a useable quality DVD. But with 24P with the same 3.8 MBS compression one can put on DVD approx. 3 hours of video with the same quality as the 2 1/2 hour at 60i.

I meant to say: Another way of looking at it is that if you have say 2 1/2 hours DV footage to render and you set 3.8 MBS compression, one gets a useable quality DVD. But with 24P with 3.2 MBS compression one can put on DVD approx. 3 hours of video with the same quality as the 2 1/2 hour at 60i.

JK

[George Wing]

[George W.] “It would appear that the 24p render DID NOT maintain the stated AVERAGE bitrate of 7,200,000bps.”

[Mike Kujbida] “I have no reason to think that it didn’t do that.”

The resulting file size should be a reason to suspect the 24p Render DID NOT maintain the stated AVERAGE bitrate…

[Mike Kujbida] “I’ll do a few more renders tonight trying the various options you suggested and post the results here.”

How did those additional renders turn out?

Regards,
George

[George Wing]

[John Kert] “I meant to say: Another way of looking at it is that if you have say 2 1/2 hours DV footage to render and you set 3.8 MBS compression, one gets a useable quality DVD. But with 24P with 3.2 MBS compression one can put on DVD approx. 3 hours of video with the same quality as the 2 1/2 hour at 60i.”

Exactly 🙂

The cadence changes using the 24p templates — so on occasion, usually during fast scenes, the motion won’t be as “fluid”

Regards,
George

[Mike Kujbida]

[George W.] “How did those additional renders turn out?”

Original (60i) file: 4.03 GB (4,333,539,328 bytes)
24p CBR: 3.95 GB (4,244,844,544 bytes)
24p VBR: 3.37 GB (3,623,215,104 bytes)
24p 2-pass VBR: 4.04 GB (4,342,931,456 bytes)

File length was 1 hr. 20 min.
CBR setting: 7,000,000
VBR setting: 8,000,000 / 7,200,000 / 3,600,000

[George Wing]

Those numbers look to be in line with what I said earlier. Thanks for the confirmation…

Regards,
George