Can anyone see a difference from these 3Mbps 7.5Mbps 9Mbps captures?

[Christopher2222]

I've uploaded 3 commercials captured at 3Mbps, 7.5Mbps and 9Mbps. The source is 720p (1280x720 at 60fps - that's what the source detection was reading from the capture) and the capture is set to the same as the source aside from my bitrate settings.

On my monitor 20" wide screen Asus, I cannot see any difference between them except for file sizes. Do any experts notice an appreciable difference between the files?

[Christopher2222]

It seems hard to notice the 3mbps until you pause it and then you can see. But during motion it's hard to notice on a 20" monitor.

[badyu17]

Yes the most noticeable for me was armored vehicles on snow. A lot of ugly blocking. Also your sample is not that good, it consists of many 1 second cuts with huge black borders that are easy to compress.

[jagabo]

It's also much easier to see defects in longer shots. With a bunch of <1 second cuts by the time start looking for something the shot changes. Watch for posterization artifacts in grainy areas with shallow gradients, especially dark grainy areas. Also watch for loss of small, low contrast detail. You can definitely see this in your samples looking at still frames.

[Christopher2222]

Ok, yes I realized the still frames will reveal more. What do you mean 1 second cuts .. oh ok in the commercial itself they've stitched together 1 second shots ok I see.
Now when you say the sample is not that good, I should have used a commercial without black borders? The 7.5 is very decent, but I can't tell if the 9 is really that much better, is it worth the extra space to go with 9Mbps?

[jagabo]

The ~17 percent difference between 7.5 and 9 Mb/s will be subtle, even examining still frames. A lot of the time you will see a difference but you won't be able to say which is better without comparing to a high quality original (ie, which more closely resembles the original).

Regarding the black borders: Unchanging lack borders take almost no bitrate to encode. So the bitrate goes only to ~75 percent of the picture. In other words, 25 percent less bitrate is needed when 25 percent of the picture is unchanging black borders. Of course, if you're always going to be recording 2.35:1 movies that is a valid test.

[Christopher2222]

Ok. I see (I'm only just experimenting now) using handbrake autocrop actually crops the screen to 1280x546 (using mediainfo to check it) and then setting Handbrake to constant quality RF 20 it brought the 9Mbps file size down to ~6.5Mb! Dropped the size by 80%.

No I won't be doing movies at 2.35:1 so ok agreed that commercial was a bad one as an example. I tried another test news clip which was recorded at 5Mbps - better video size and had Handbrake bring the size down by about 40%! (again using RF quality set to 20)

My aim is to get the best compromise of quality and file size. It seems the best way to go about this is not to capture at the best bitrate setting but to capture at a higher setting and then use handbrake to compress it a little more.

Although if I was just going to capture the programs without editing (which was my original intention - don't want to spend too much time editing except to edit out commercials) then deciding at capturing at either 3, 7.5 or 9 might tempt me to stay away from 3 and just go with 7.5 and maybe 9 (in another post of mine I mentioned I settled on 5Mbps - however I think I'm slowly thinking I should go slightly more). With a few tests I couldn't help but think the 5Mbps video seems like it was slightly lower quality (but maybe it's all just in my head.. haha) , the 9 seemed a bit excessive but 7.5 seems ok for the testing I've done so far. And like you said, the difference between 7.5 and 9 will be subtle.

I appreciate everyone's input.

[KarMa]

I did a split screen comparison of the 3 videos you provided, outputting in lossless YV12 Lagarith. Midway through the "9 Mbit" video goes one frame out of sync due to what I believe to be a missing frame or maybe the duplicate frame pattern changes compared to the other two. Did not care to trouble shoot it much more, and it gets the point across.

[Christopher2222]

Out of sync with the audio?
Is there a simple program that will identify where out of syncs occur?

[KarMa]

Out of sync with the audio?
Is there a simple program that will identify where out of syncs occur?

I did not include audio. I was talking about the video frames being 1 frame out of sync in my video. With the right side being the one to drop a frame and be one frame ahead the center and left. Happens around the 13 second mark in my video but then goes back in sync around second 20. Not sure if it's my fault or the video itself is missing frames.