Wolverine-Hawkeye Telecine

[Stan Jelavic]

Nice work David. Ah, yes remember the script copy and paste issues. What a pain.
HTML tags should have helped.
I ran into issue when using stabilize ahead of the dirt filter with the hi rez clip.

[David Brown]

There is a lot to take in, wish I had started post production earlier. Thanks Stan for helping with the Dirt filter. I have added your changes to my Option4.avs. What helped me was copying the lines from the forum and pasting them directly into the avs. There must have been 10-15 errors.

When copying from the forum, extra formatting is hiding the pasted script! Seems like I remember this from some time ago.

I fixed them all! The error message provides the line number and column position and eventually I got the script to run!

There is another Dirt filter available for Vdub. Spent a lot of time with both and the AVISYNTH version is best. Vdub filters are easier to use and often have descriptions and live feedback of the sliders.

The artifacts I had from a couple of days ago came from the Dirt filter. Too much "strength" together with fast motion in the scene will turn to a mess. The cartoon I've been using has been out in the open at various positions on the rewinds. It had lots of dust that could have been wiped clean with a cloth, but it was a great mess for testing the Dirt filter.

I've not been able to see any differences when changing the Denoising strength. That is part of the Dirt filter magic. Default at 300, I've had it at 900 and 1.

Remove dirt completely cleans up the scene at dirt_strengtht=80 and remove the artifacts. I had first tried it at 100. Except for a scene change. The last frame of the scene and the first frame of the next are not cleaned. There is a forwardclense and backwardclense function I need to find for last/first frame of scene changes.


Kamel, you might be able to sell your 12mm and 72BUOC to someone building the Hawkeye upgrade. Helps defray the cost.

So many many filters to try..................

[Stan Jelavic]

Kamel,
here is a copy of invoice from TIS in May this year:


The USB3.2 cable is $29.

[David Brown]

That's crazy Kamel. In June I paid 324 euros and that included shipping. There must be a market shortage for the 37ux226.

[Stan Jelavic]

Kamel, I do not see much of advantage except for higher resolution and the color tones are somewhat better. With USB3 it is a faster camera and can give you better scan rates for the same resolution.
So, if you just use a laptop to view your videos I would say it is not worth spending that much more money. But if you have a good quality large monitor you will notice a difference.
I will dig the quote out for you.

[Kamel Ikhlef]

Thanks for your kind suggestion Stan.

for the moment I have not yet decided because I have a lot of delay in my projects.
And I don't know if the difference in quality is worth it

But, Yes it can be interesting. Tell me the price with the shipping costs.

I will also order the new V13 to be on top!

About V13, one other more small request.
I know that space is lacking but it would also be good to add the possibility of connecting a second fan to cool the camera.
Maybe, that may interest some people...

[Stan Jelavic]

David, the script is avisynth. You get it from VideoFred download:
https://forum.doom9.org/showthread.php?t=144271

Download --- unzip and then go to the Film Restoring/scripts dir. (you probably know this but anyways)
The original script that I used was Film_Restoring_Frame_Interpolation.avs

I then modified it and created the option4.avs script.
https://drive.google.com/file/d/1vC0...ew?usp=sharing

You can then follow my instruction above to modify option4.avs to just do dirt removal and perhaps play with some more filters.

When done open the script from VirtualDub2 as any other video file. Edit in VirtualDub2 or just save it as a raw avi file or compressed or whatever.

Fernando, RPI is very slow for image saving which will give you very slow FPS. I can try it.

Kamel, how about if I buy it for you here and ship it. It may be cheaper.

[Kamel Ikhlef]

Bruce, at the moment I have the BUC02.
I wanted a better camera like Stan or David but in France it is overpriced (600 € + taxes and shipping costs).
That's why I looked at the e-con products. Stan had done some tests and finally I haven't decided yet.
I would like 4K...

Thanks Stan for advancing on V13.

[David Brown]

WOW, thanks Stan!

Is this script used for Avisynth, some language? Like C+ or Python? Just wondering. I'll test that tomorrow.

Congrats Fernando.

Sorry I can't answer your question except to say spend as much as you can afford!

I did buy a (I need to find it) Windows 10 pc that is the size of a large thumb drive. If I find it, I'll give it a test. If you just want something to do the capture and save the files, it should work as would anything that small. Don't forget a tv or monitor.

Processing the images/videos might be a problem, unless you intend to do that on another more powerful machine.

[Stan Jelavic]

Hi David,
I still have to move the camera mount.
I looked at the script tonight and got the RemoveDirtMC working. That is the only thing done to the video for now. No cropping - nothing just RemoveDirtMC.

Here is how you do it.

The execution of the script is from the bottom:
Eval(result)#.converttoRGB24()

converttoRGB24 not required.
Next find the result definition.

result="result4" # specify the wanted output here

Next result4

#RESULT4: MANUAL LEVELS, MANUAL COLOR CORRECTIONS
#................................................. .................................................. .................................................. ................
result4= cleaned.coloryuv(off_U=blue,off_V=red).levels(blac k_level,1.0,white_level,0,255)\
#.addborders(bord_left, bord_top, bord_right, bord_bot)

See my changes.. I changed the input video for result4 to cleaned. Color tweak is OK I believe --- hmm maybe not for the UX226 -- will check

Now find where cleaned is defined
cleaned= RemoveDirtMC(AviSource(film).converttoYV12(),dirt_ strenght).unsharpmask(USM_sharp_ness1,USM_radi_us1 ,0)\
.unsharpmask(USM_sharp_ness2,USM_radi_us2,0).Lancz os4Resize(W,H)

Note my change to go straight to the source video. Had to convert it to YV12 for the dirt filter to work.

This is just a start and other filter can be added. Still not sure if stabilize may be needed ahead of the dirt filter. I think something breaks when you do that.

[Fernando de Peña]

Hi everyone. Just bought my Hawkeye from Stan and I’m thinking on buying a PC to work with it. Anyone had experience of some kind using the Wolverine/Hawkeye with a Raspberri? Which should be the minimum specs of the PC to have a good performance of the Hawkeye?

Im thinking in this raspberry model:
https://www.raspberrypi.org/products...ellerType=home

un abrazo desde Puerto Rico

[David Brown]

Boy, I'll say the V13 board is maxed out! I decided not to over-drill the camera mounting holes with V12 for fear of hitting a trace. As it is, one camera supporting leg is insulated from shorting a transistor.

My lament about the scripts is being unable to easily evaluate any one filter when so many are being applied.

Running multiple filters is the most efficient method for processing, especially when there is sometimes a specific order suggested.

I have not chosen a pre processing video size, it's over scanned to reveal partial sprockets and the frame lines. The current sample I'm using is 8mm:1872 x 1452. Now I see that I must select a certain size based upon the processing blocks. I hope not. * RemoveDirt requirements below. #3 is not something we care about.

IMHO stabilizing the video before RemoveDirt should improve the dirt removal, but I don't know.









*RemoveDirt Optimal Usage

1. If possible, crop after RestoreMotionBlocks. Modern codecs divide the frames in the same way as RemoveDirt into a grid of 8x8 pixel blocks to perform the crucial discret cosine transform for such blocks. Now if the clip is cropped after RemoveDirt, then the grid of RemoveDirt and the codec are likely to be different resulting in inferior compression. There is one exception, though: cropping afterwards does not hurt, if all four sides are cropped by a multiple of 8. For instance, crop(8,64,0,-72) is ok. On the other hand, one should crop after RemoveGrain/Repair if possible, because this filters cannot process the boundary pixels. Thus the optimal solution is to crop afterwards and then only by multiples of 8, which unfortunately is not always possible.
2. Crop only with "align=true". RestoreMotionBlocks heavily uses SSE/SSE2 instructions. If you crop without "align= true" before RestoreMotionBlocks, then the data on the frames may not be properly aligned and RemoveDirt will execute substantially slower. This is particularily important for the SSE2 version. As a consequence you should always crop with Avisynth and not with DVD2AVI or DGIndex.

3. Telecined movies must be inverse telecined before RemoveDirt. If a film is telecined some fields are doubled in order to increase the frame rate from 24fps to 30fps. Hence on such doubled fields the basic property of dirt, described above, is no more valid and no temporal cleaner can ever spot dirt on such doubled fields. On the other hand, after an inverse telecine usually every fourth frame is composed of fields, which originate from two different frames. Visually these two fields fit together well but both are from a different compression context, which can mislead RemoveDirt to false motion detection. In extreme cases, one field may be from an I- or P-frame, while the other is from a B-frame. But even if the fields are from from frames of identical type, the different compression context has a substantial effect. Consequently RemoveDirt performes less well on inverse telecined movies than on natively progressive movies. By the same reason also compression of inverse telcined movies is worse than of natively progressive movies. We in Europe should thank god every day that we are not getting telecined. However, here in Germany we have digital tv broadcasters, which like to comb progressive films (about 5% of all progressive movies from ARD and especially ZDF are combed). Fortunately these idiots are not able to double fields, so RemoveDirt should work, but on combed films the dirt is always split over two frames which clearly hurts RemoveDirt. On the other hand, if these combed films are uncombed, then we have the compression context problem for any frameand not only for any fourth frame. Stepping through the video with the builtin filter Bob()one can decide with near absolute certainty, whether the video is truely progressive, interlaced, telecined, field blended or progressive with a field shift.

4. Put other filters after RemoveDirt. Except those filters mentioned before, like crop and inverse telecine, all other filters should be put after RemoveDirt in the Avisynth script, because most filters have a negative rather than a positive impact on dirt detection.

[Stan Jelavic]

Hi David, somehow I missed your post.
So let me try to understand your problem.
You just want to keep the video as is and just use Clean_Dirt and DeGrain functions. Right?

What size is your video? I assume you are using the UX226?
I had some issues with VideoFred and larger size video. Just started looking at that a few days ago.
I have a pretty good handle on how the script works I think and can try to reproduce your problem here and possibly help you with that.

Kamel, I will add the 12V bias and the LED pad. Adding both side and vertical type A is a challenge because these connectors are through hole. The current USB2 connectors that we used in V12 are SMT and 4 traces only. These are 9 traces and through holes. I can do a V13a version with the vertical mount pretty easily if needed.
I am still using the Eagle free version and the board is the max size and cannot be made larger.
Kamel, you can use the 12V leds. Something like this:
https://www.amazon.com/Indicator-Lig...dp/B075F6KPNW/
Here is the V13

[Bruce Davis]

Hi Kamel, have you made a decision yet on which camera you are going to use.

[Kamel Ikhlef]

Hello everyone,
Thanks Bruce,

Stan for the leds I only use 12v for the power indicator. You just have to keep one.

Maybe we can save a little space by putting the usb A connector vertically and enlarge the card a little at the top right.
But I have a large unit (movie maker pro).
I don't know how much space we have for the other models (reflecta,somikon...).

very good for the sonalert.