DV-AVI best format for viewing on a TV

[Sharc]

I can only see 3 vertical lines on my windows media player and VLC player, What do I need to do to get my player to comply with full range? Thanks.

Your second picture in post#36 indicates that your video player (or browser) ignores the full flag

And if you adjust the player or the graphics card's settings to compensate for that -- all your limited range videos (pretty much everything else) will display incorrectly (washed out blacks, dull whites). And what if you use the media player built into your TV? Or some other media player device (Roku, AppleTV, GoogleTV, etc.)? Are all of them going to follow the range flag? This is why I recommend people stay away from full range encoding.

Fully agree. I didn't promote using full. Not at all. I just provided the 2 examples for letting the OP try how his players react.

[jagabo]

In the script I've replace jagabo's color ColorYUV command with ColorYUV(levels="PC->TV") I'm guessing they both achieve the same final result?

No they do not. I recommended ColorYUV(gain_y=-15, off_y=-10). ColorYUV(levels="PC->TV") is approximately the equivalent of ColorYUV(gain_y=-36, off_y=16).

[lollo]

Your processing did not change the levels, it produced super-white artifacts. You might want to make those artifacts visible by lowering the white level but then other parts of the picture that were perfect white (Y=235) before processing will now be a light grey instead (Y=~215), making the picture a bit dull. And the rest of the picture proportionally darker. And if you do the same at the dark end your perfect blacks (Y=16) will become dark greys (Y=~32), making the picture look washed out.

I agree, but the "meaning" of sharpening is to produce such artifacts and keep them.

And concerning the darker and the washed out part of the pictures, it is the same as when you capture YUV in the 0-255 range (only few cards can do it) and then reduce the range to 16-235 before an AviSynth processing. I am acting on the levels at the end of the processing because this last "expanded" the range from limited to full, but it is equivalent.

The issue here is your definition of "levels". You are using the term just to describe the range of Y values. I'm using the term to mean the theoretical full black and full white values.

Full black and full white are just a part of the whole picture. I just want to avoid clipping after processing, that's all I tried to explain.

[Sharc]

In the script I've replace jagabo's color ColorYUV command with ColorYUV(levels="PC->TV") I'm guessing they both achieve the same final result?

No they do not. I recommended ColorYUV(gain_y=-15, off_y=-10). ColorYUV(levels="PC->TV") is approximately the equivalent of ColorYUV(gain_y=-36, off_y=16).

^^^This.
And @SkyBlues2021: remember I suggested ColorYUV(levels="PC->TV") for rgr's 'full' version of post#26 only which I mistakenly took as your encode . Sorry again for the confusion this has created.

[SkyBlues2021]

Thanks jagabo, Sharc & lollo. Something came up yesterday and I wasn't able to practice the ColorYUV scripts. However I will research and practice on them on Saturday. Thank you for all the advice.

[SkyBlues2021]

I've tried ColorYUV and adapted jagabos settings for a different clip. It seems to me to be within the legal range. I was wondering when you would use the U&V in ColorYUV? My previous clips script used dehalo_alpha but now I've changed over everything to 64 bit do I need this plugin to be 64 bit or do some plugins work on both? It does seem to work with my Avisynth+ 64 bit. I couldn't find a 64 bit on the Avisynth site. Thanks.

[jagabo]

That black level looks too high to me. The main point of adjusting levels is to get parts of the frame that you know are supposed to be full black at Y=16, and parts that you know should be full white at Y=235. Having some small excursions outside that range isn't a problem. Not all shots will have full blacks and full whites -- so you may have to look at a few different shots to verify the levels. Of you may want to adjust the black/white levels differently for different shots. These are judgement/artistic calls. Of course, sometimes you may want to use such adjustments for other things -- like when you don't care about black levels and just want to make dark details more or less visible.

off_u and off_v can be used to adjust white balance (though RGB adjustments are usually better for this). Sometimes you'll have a video where the chroma channels are consistently shifted up/down and/or left/right. That type of problem can be fixed by shifting them back to the center (U=V=128).

cont_u and cont_v are for adjusting saturation. Occasionally you'll need to use different values for U and V.

Saturation and gain often go hand in hand. If you adjust gain_y (or cont_y) then you may want to adjust cont_u and cont_v by similar amounts.

[SkyBlues2021]

Thanks jagabo. Sorry if I'm not grasping it properly. I'm guessing that my staying within the bars on the waveform is correct? but my levels are wrong and my blacks need to be darker and less washed out? I've redone the levels but only visually? How should I analyze it better? I tried to make the dark area on my shirt darker however I'm not sure if this is the correct thing to do.

[Sharc]

On the .avi source of post#66 you could for example try

colorYUV(gain_y=-17,off_y=-5)

But apart from "technical correctness" it's also a matter of personal preference.

[SkyBlues2021]

Thanks Sharc. I will try your code tomorrow. What I struggle with is how you work out how to get to colorYUV(gain_y=-17,off_y=-5)? I've been reading various past forum and Avisynth pages today regarding ColorYUV but I'm not sure I'm understanding it any better. I even printed jagabos post#39 and read it at work this week to try to understand the concept better. My thinking was probably wrongly that you stay within the 16-235 range and then toggle between colorYUV(gain y=,off_y=) to get the correct levels. I know my explanation probably sounds crude I'm just hoping I've got even a small part correct so I can build on that. Thanks.

[jagabo]

Here's a waveform monitor of the original video, frame 124:

TurnRight().Histogram().TurnLeft()


[Attachment 71628 - Click to enlarge]

Given how much of the waveform is bunched up just above the green line I suspect that is the correct black level. Little bits of the waveform below that level are probably just the result of oversharpening halos. Those aren't "real" portions of the picture but rather artifacts of the camera's and DVD recorder's processing. At the high end the crushed peaks at Y=255 indicate the white level should have been higher than 255. We don't really know how much higher. So we'll just bring those existing peaks down to 235.

Before seeing Sharc's latest reply I came up with ColorYUV(gain_y=-12, off_y=-8), a little lower at the black end, higher at the white end. Either would be adequate.

[Sharc]

What I struggle with is how you work out how to get to colorYUV(gain_y=-17,off_y=-5)? ....... My thinking was probably wrongly that you stay within the 16-235 range and then toggle between colorYUV(gain y=,off_y=) to get the correct levels.

Your thinking is correct. It's all about shifting the waves of the waveform monitor close to 16 for the darks (16 means black) and close to 235 for the brights (235 means white). You can usually ignore minor short occasional excursions Y<16 or Y>235 (typically caused by artifacts).
AvsPmod is an excellent tool for such adjustments as you see the result of the changes immediately. Once you try out and understand the basic impact of gain_y= and off_y= you can usually find the "correct" values in a few iterative steps (trial and error).

Edit:
In case you feel that dark areas and shadows look too dark you may want to play with gamma_y rather than just shifting the dark (black) level upwards.

[SkyBlues2021]

Thanks jagabo & Sharc. What you have both explained has been really helpful to me thank you. I think one of the mistakes I'm making is that I'm trying to get everything within the bars. I will play with both your codes and post some new clips that hopefully I'm able to get nearer the desired standard. Thank you for your patience.

[Sharc]

.....and, at the end and referring to the title of this thread, play an encoded testclip on your (hopefully properly adjusted) TV ......

[SkyBlues2021]

I've uploaded 4 new clips. Hopefully I'm getting closer to where I need to be? The last 3 come from one tape (tape 38) and I'm hoping it works on the indoor/outdoor scenes. Thanks.

[SkyBlues2021]

Thanks Sharc. Yes good point, I need to do that more often.

[jagabo]

I will sometimes temporarily blur the picture and crop away the edges to reduce noise while making adjustments like this. For example:

BinomialBlur(5)
Crop(16,16,-16,-16)
TurnRight().Histogram().TurnLeft()

Once the adjustments are made I'll remove the blur and crop. And maybe further refine the adjustments.

[Sharc]

I've uploaded 4 new clips. Hopefully I'm getting closer to where I need to be? The last 3 come from one tape (tape 38) and I'm hoping it works on the indoor/outdoor scenes. Thanks.

Looks good to me.
You could probalbly use as well ColorYUV(gain_y=-12,off_y=-8) for all 4 clips I think. The differences are subtle only.

[SkyBlues2021]

Thanks jagabo & Sharc. Great news thank you, I finally got there! I will try the BinomialBlur as well and see if I get slightly different results and post. Thank you again for all your help.

[SkyBlues2021]

Thanks jagabo & Sharc. Yes I see what you mean it makes the waveform very clean to work with. I hope I put ColorYUV in the correct place? after the crop but before TurnRight().Histogram().TurnLeft(). I used Sharc settings which seemed to fully fill the waveform without encroaching into the brown bars.

[jagabo]

Instead of BinomialBlur() you can downscale and upscale the frame. BilinerResize(width/4, height/2).BicubicResize(width, height). Or for HD video I often just downscale. Whatever works for the situation.

And I agree that the DV device is pretty consistent and you can probably use a compromise setting for most shots.

[SkyBlues2021]

Thanks jagabo. I will read about BilinerResize & BicubicResize I think I've seen it in other people scripts but never researched it. I will encode the videos tomorrow after work and post. Thanks.

[SkyBlues2021]

I've encoded the videos. I deinterlaced and used dehalo_alpha but I will be careful when I use it. Hopefully that was ok to use it? I will watch the clips on my TV to check they look good. Thanks.

[Sharc]

I've encoded the videos. I deinterlaced and used dehalo_alpha but I will be careful when I use it. Hopefully that was ok to use it? I will watch the clips on my TV to check they look good. Thanks.

Looks good to me.
Whether you want to generally apply dehalo is your choice. It's a compromise between reducing halos and loosing fine details and structures (tendency to producing "wax skins" for example).

Edit:
In the first (....indoor bright....) .mp4 I see some color flicker (rainbows) at the edges of the shelf in the center. I think this is baked into the original though, and not caused by your processing.

[SkyBlues2021]

Thanks Sharc. Thank you for the advice. I will be careful when I use dehalo_alpha. Thanks.

[jagabo]

FineDehalo() is another option. You can use the same arguments as dehalo_alpha(). FineDehalo() uses a mask of sharp edges and only dehalos the worst halos. That keeps it from loosing small details. You may have fine tune the threshold values to get the look you want.

http://avisynth.nl/index.php/FineDehalo

I sometimes build my own edge mask and use it with dehalo_alpha().

edgemask = mt_edge(...)
dehaloed = dehalo_alpha(...)
Overlay(last, dehaloed, mask=edgemask)

[SkyBlues2021]

Thanks jagabo. I will have a look at the FineDehalo() that could be useful for a few of my videos that need more restoration and could be challenging. I will be posting them in the next few weeks. The edgemask looks a bit more daunting, I've seen it in other people scripts. I've found ColorYUV extremely hard to learn, I'm trying to retrain an old brain! I'm so grateful for the support I get. Thank you.

[SkyBlues2021]

I going to encode my first full video on Sunday and I wanted to check as I'm not 100% sure. I captured my analogue videos using Virtualdub 1:9:11 32 bit, now I've set up everything for 64 bit Virtualdub2 & Avisynth to edit and encode and I wanted to double check that this is ok? Thanks.

[jagabo]

That should be fine. Assuming that everything is working properly -- and from what we've seen that seems to be the case.

[Alwyn]

Fiddle around with your AVISynth scripts in the VDub script editor: Tools>Script Editor. Works well.