qcomp - what is it?

[TheLastOfThem]

Can anyone please explain what exactly this does? I heard that at 1 - qcomp attributes more bitrate to faster scenes and less bitrate to motion scenes at lower qcomp values.

Is this true? Anything else about this that's relevant?

[TheLastOfThem]

Also another question - any ideas why when I do SelectEvery the FPS of my video changes to decimals - .105 fps? .120fps?

[TheLastOfThem]

One last question - im working with a NTSC DVD and not noticing any interlacing or 3:2....The DVD is being reported as 29.97 however....confused as to what this is then in terms of source?

[manono]

Your questions are so imprecise that no one can answer. I'll assume you're using AviSynth scripts. And, if so, please give the exact script and provide 10 seconds of the DVD, untouched and with about 10 seconds of steady movement. No one likes to guess what you have. And what program are you using to encode, and to what - what format?

any ideas why when I do SelectEvery the FPS of my video changes to decimals - .105 fps? .120fps?

No idea what that question is about. Nothing uses those framerates.

The DVD is being reported as 29.97 however...

All NTSC DVDs output interlaced 29.97fps (59.94 fields per second).

[TheLastOfThem]

Nevermind on 2nd question..I did SelectRangeEvery..which acts appropriately..but does anyone know why SelectEvery DOESN'T and changes fps? weird!

[TheLastOfThem]

But I see absolutely no 3:2 pulldown method or interlaced lines?

[manono]

Nevermind on 2nd question..I did SelectRangeEvery..which acts appropriately..but does anyone know why SelectEvery DOESN'T and changes fps? weird!

Not weird at all. Maybe if you read the doc before spouting nonsense you'd understand what it does. It changes the framerate by the percentage of remaining frames, keeping the length the same. For example, if the source framerate is 29.97fps, and you SelectEvery 4 out of every 5 frames:

SelectEvery(5,0,1,2,3)

you are selecting 80% of the original frames and get a 23.976 framerate (80% of the source framerate).

But I see absolutely no 3:2 pulldown method or interlaced lines?

Okay, so? Can you confirm that with a 100% FILM D2V file? Maybe you have a deinterlacer turned on in your player. One doesn't determine hard or soft pulldown just by looking at bits of the video. Still waiting on the sample.

[TheLastOfThem]

Nevermind on 2nd question..I did SelectRangeEvery..which acts appropriately..but does anyone know why SelectEvery DOESN'T and changes fps? weird!

Not weird at all. Maybe if you read the doc before spouting nonsense you'd understand what it does. It changes the framerate by the percentage of remaining frames, keeping the length the same. For example, if the source framerate is 29.97fps, and you SelectEvery 4 out of every 5 frames:

SelectEvery(5,0,1,2,3)

you are selecting 80% of the original frames and get a 23.976 framerate (80% of the source framerate).

But I see absolutely no 3:2 pulldown method or interlaced lines?

Okay, so? Can you confirm that with a 100% FILM D2V file? Maybe you have a deinterlacer turned on in your player. One doesn't determine hard or soft pulldown just by looking at bits of the video. Still waiting on the sample.

D2V reports frame type as interlaced. I'm not looking at bits of the video I'm looking at ALL of it. There's absolutely no pulldown on this video. And if one can't determine it from bits of the video, why do you spew nonsense of wanting to see a sample?

[TheLastOfThem]

removed my diatribe

[TheLastOfThem]

removed my diatribe

[TheLastOfThem]

removed my diatribe

[TheLastOfThem]

removed my diatribe

[manono]

D2V reports frame type as interlaced. I'm not looking at bits of the video I'm looking at ALL of it. There's absolutely no pulldown on this video. And if one can't determine it from bits of the video, why do you spew nonsense of wanting to see a sample?

I'll ignore your multi-post diatribe. It's nonsense wanting to see a sample? You're the one asking for help and to provide it we have to know what you're dealing with. It's standard practice to provide information when asked, such as the script used, the encoder used and the video source. A video sample is worth much more than all the words you used in describing what you're dealing with, and some of the information you've provided was misleading at best.

If it's 100% video then it could be any one of a number of things. It could be all progressive 29.97fps but encoded as interlaced. It could be hard telecined film. It could be true interlaced video. And other things. If you don't see any interlacing then it could be that whatever you're using to play it is deinterlacing it. Samples are very important, whether you realize it or not.

What about if you open the AviSynth script in Virtual Dub using MPEG2Source and no other filtering? No interlacing when advancing a frame at a time through a scene with movement? No duplicate frames? If not then you have a progressive 29.97fps video.

To cut a 10 second sample, open a VOB in DGIndex. Use the [ and ] buttons to isolate a small piece. Then go File->Save Project and Demux Video. You can upload the resulting M2V file here here.

[jagabo]

D2V reports frame type as interlaced. I'm not looking at bits of the video I'm looking at ALL of it. There's absolutely no pulldown on this video.

It means the source wasn't film but rather some 29.97 fps progressive video. Or, if it was film, it was produced by duplicating frames rather than using 3:2 pulldown.

[TheLastOfThem]

....Thank you manono for ignoring that diatribe........

Can we maybe just by default assume I'm stupid / nonsensical / moronic when you post next time ? It'd save a lot of your comments on my personality.

Getting back to the sample.....I didn't want to share it because it's a commercial DVD (not sure if allowed)

but here is a sample of it...(raw mkv rip)

http://www.mediafire.com/download/3mjv4l7a1c6eidi/themesongdvd.mkv

I just noticed....

Looking at 5 frames...

Frame 1 & 2 = the same


Frame 3 & 4 = same

frame 5 = different.

Not sure what's going on here. Frames do seem duplicated....

I thought DVD spec did not allow for progressive video to be saved on to it without interlaced fields?

[manono]

I didn't want to share it because it's a commercial DVD (not sure if allowed)

Sharing short samples of copyrighted works for discussion and study is permitted under the Fair Use and Fair Dealing laws of virtually all countries.

but here is a sample of it...(raw mkv rip)

I already described in detail how to take a video sample from the DVD using DGIndex. But that's okay, at least you didn't reencode it.

Looking at 5 frames...

Frame 1 & 2 = the same


Frame 3 & 4 = same

frame 5 = different.

That's just one part. Others are different.

Well, it's progressive with every fifth frame being a duplicate. To return it to its original 23.976fps use the animation mode of TDecimate:

TDecimate(Mode=1)

TDecimate is part of TIVTC.

I thought DVD spec did not allow for progressive video to be saved on to it without interlaced fields?

You thought wrong. It's all progressive content, but encoded as interlaced. Here the two fields making up each frame happen to be from the same moment in time. What's important is how it's output for display. For example, 23.976fps can be be encoded progressively, but with 3:2 pulldown to output interlaced 29.97fps.

Oh, I can't answer the original qcomp question, although I probably know it as a different term. If you give us the name of the encoder you're using (MeGUI?), perhaps someone that uses it can answer the question.

[TheLastOfThem]

I'm using x264 encoder. No I didn't re-encode the sample fortunately .

When I put the MKV in AviSynth using FFMpegSource2 (just this)

I'm seeing that one part with duplicate frames....in the way I mentioned.

Can you tell me how exactly you noticed every fifth frame is a duplicate?

Another question I had about TDecimate - is it used for mostly DVD's ? I'm just wondering because I had another source that was a PAL SAT Rip (25 fps) but not 29.97 and it was interlaced during SOME parts. No 3:2 pulldown and no duped frames from what I can tell (though I may suck at this)

Bit rate : 2 460 Kbps
Maximum bit rate : 15.0 Mbps
Width : 720 pixels
Height : 576 pixels
Display aspect ratio : 4:3
Frame rate : 25.000 fps
Standard : PAL
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Interlaced
Scan order : Top Field First
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.237
Time code of first frame : 00:00:00:00
Time code source : Group of pictures header
Stream size : 388 MiB (91%)

I just did a decomb in handbrake with these files (all of them) and kept the same FPS. I believe the original source was drawn (it's a cartoon) in 23.976 however..and all I did was change the FPS in the .mkv container to 23.976...I believe unless I chose 23.976...handbrake does not remove duped frames...

I always wonder what duplicate frames will do to something that was meant for 23.976? Would it make it jerky?

[TheLastOfThem]

Thanks for still helping.....manono

[jagabo]

First of all, cartoons are not animated at 24 fps. Typically, character animation is at 12 fps, even 8 fps, only in panning shots does the background move at 24 fps. So you expect to see lots of duplicate frames as cells without panning are duplicated 2 or 3 times when put on film (12 fps to 24 fps, or 8 fps to 24 fps). Secondly, the theme song animation is often different than the body of the cartoon. Often shots from the body of the cartoon are sped up or slowed down to match the timing of the theme song. So you expect to see duplicate or missing frames during the theme song.

The MPEG 2 video is encoded interlaced but the content of the frames is progressive -- so there are no combing artifacts. Looking at the shots with panning (eg, the shot where Courage is carrying Muriel and Eustace within the TV frame) you can see there is a duplicate frame every 5th frame. So the film was converted from 23.976 fps to 29.97 fps by duplicating every 4th frame.

[TheLastOfThem]

Here the two fields making up each frame happen to be from the same moment in time

How did you figure this out? That sounds tricky. So if a source is not showign interlaced fields - it means the fields are from the same moment in time? I assumed the point of fields was Field 1 showed one half of a frame, and Field 2 showed another half in the next moment...or something like this.

[TheLastOfThem]

The MPEG 2 video is encoded interlaced but the content of the frames is progressive -- so there are no combing artifacts. Looking at the shots with panning (eg, the shot where Courage is carrying Muriel and Eustace within the TV frame) you can see there is a duplicate frame every 5th frame. So the film was converted from 23.976 fps to 29.97 fps by duplicating every 4th frame.

Thank you. I do see it where you mention the duplicate 4th frame. But ironically when the TV is being switched off 7 frames in a row are different...I don't see a duplicate 4th..not sure what's happening here?

I'm confused in how a video can be encoded interlaced and the content of frames is progressive. I thought putting anything on DVD ended up adding combing artifacts....

So progressive source always ends up progrssive on DVD? When does it improperly end up having combing artifacts?

[jagabo]

I'm confused in how a video can be encoded interlaced and the content of frames is progressive. I thought putting anything on DVD ended up adding combing artifacts....

No. When progressive frames are encoded interlaced there is no combing. Think of the usual hard 3:2 pulldown process. 3 of every 5 frames look progressive (because both fields come from the same film frame) even though they are encoded interlaced. Only two of the frames show combing, those which contain fields from two different film frames. Normally 3:2 pulldown is used to get smoother motion. But it looks like the producers of this DVD decided to duplicate frames rather than apply 3:2 pulldown.

[TheLastOfThem]

I'm confused in how a video can be encoded interlaced and the content of frames is progressive. I thought putting anything on DVD ended up adding combing artifacts....

No. When progressive frames are encoded interlaced there is no combing. Think of the usual hard 3:2 pulldown process. 3 of every 5 frames look progressive (because both fields come from the same film frame) even though they are encoded interlaced. Only two of the frames show combing, those which contain fields from two different film frames. Normally 3:2 pulldown is used to get smoother motion. But it looks like the producers of this DVD decided to duplicate frames rather than apply 3:2 pulldown.

Thank you so much!! this makes complete sense. Though I still wasnt' able to see exactly how each frame was duplicated. It seems to vary. I noticed the 4th frame when he's running with Muriel and Eustace...but when Eustace clicks off the TV every single frame is different

[manono]

About qcomp:

https://en.wikibooks.org/wiki/MeGUI/x264_Settings

says this:

qcomp

Default: 0.60

Quantizer curve compression factor. 0.0 => Constant Bitrate, 1.0 => Constant Quantizer. Note that adaptive quantization partially replaces the effect of qcomp and x264 will internally raise qcomp to compensate based on the adaptive quantization strength.

It looks to me, as for most of its settings, that you can leave it alone without worry. I think it's used to give more or fewer bits depending on the complexity of the scene. That is, the closer to one you choose, the closer to true VBR (constant quantizer) you get. Maybe someone else that cares about such things will chime in.

[jagabo]

but when Eustace clicks off the TV every single frame is different

As I mentioned earlier, all kinds of odd things can happen during the theme music as the video is modified to match the music. Even in the body of cartoons you can get similar speed ups and slow downs.

[TheLastOfThem]

Thank you both for your help. I used TDecimate(2) as it seems to be exactly what this duped frame source needed.

[TheLastOfThem]

Had one final question -

In terms of animations - does duplicating frame necessarily create smoother motion?

[jagabo]

Had one final question -

In terms of animations - does duplicating frame necessarily create smoother motion?

No, it's the opposite. It makes for less work in animating the characters. Instead of having to draw 24 different character positions for every second of film they only draw 12 or 8 or whatever. But the result is jerky motion.

[TheLastOfThem]

Had one final question -

In terms of animations - does duplicating frame necessarily create smoother motion?

No, it's the opposite. It makes for less work in animating the characters. Instead of having to draw 24 different character positions for every second of film they only draw 12 or 8 or whatever. But the result is jerky motion.

A lot of people mention deinterlacing using 'Bob' and outputting what was a 23.976 source to 50fps for "smoother motion" But this concept I don't understand. If you use Bob deinterlace to extract frames from repeated fields then how does this "Create smoother" motion?

[manono]

For your source it doesn't. It's usually meant for sources with a lot of interlacing when you want to make the source progressive for one reason or another. It turns each field into its own frame. It won't do you any good at all. You'll double the framerate but also double the number of duplicate frames.