Digitizing Hi8 for storage/editing and dvd

[Ranubis]

Hello all,

I have a PC using Windows 7. I'm using a Sony CCD-TRV66E (PAL) with something like 20 Hi8 tapes from various family events. My aim is to convert all this footage to an external drive for storage, editing and also some dvd converting.

The Sony is outputting S-Video to an AVerMedia DVD EZMaker 7 device that connects through USB to the computer. Unfortunately this camcorder lacks the Firewire interface and for now this is the system I can use. For capturing I was first using the Cyberlink PowerDirector 10 with the following settings:

AVI High Quality Profile

Video

25 fps

640x480

True Color (24 bit)

Audio

PCM 48 Khz 16 Bit Stereo

This results in really huge files like for 01:06 hrs of footage I was getting 58,2 gb!

I then tried the STOIK Capturer. Since I've read that DV-AVI is the best for compression I was using it and the results were very nice albeit for some reason there is no sound but just the sound of the camcorder itself! How can I configure that properly?

I would like to mention that I'm using a 500 GB drive and possibly it won't be enough but for now it will have to do the trick.

After this process I would do some basic editing on Premiere and then export to DV-AVI for digital storage or MPEG-2 for dvd conversion.

What settings for this btw?

Thank you very much!

[Ranubis]

Anyone?

It's a bit urgent. Thanks!

[aedipuss]

powerdirector allows you to change the format you are capturing to. try dvd spec mpeg-2. in pal land that is 720x576 at 25fps. you can change the bitrate and audio settings also.

[Ranubis]

But isn't capturing to MPEG-2 already degrading the quality? Wouldn't it be better to capture at DV AVI first for storage purposes?

I just need a software that does that...

[Ranubis]

I can't capture as DV - AVI from Power Director... The only thing I can get is a huge 640x480. I would like to be able to capture an 720x576 DV-AVI. Any suggestions?

[Richard_G]

I would like to be able to capture an 720x576 DV-AVI. Any suggestions?

Use a camcorder with Firewire output and WinDV. Then the only format you will be able to save as is 720x576 DV-AVI.

[smrpix]

If you have firewire on your computer you have several options.

Replace your Avermedia box with a D/A converter like ADVC 110.
Replace your AverMedia box with a DV camera that has passthrough.
Replace your Avermedia box with a Digital 8 Camera that will play your Hi8 tapes

Even if Powerdirector cannot capture the DV directly from those devices, you can use WinDV and import.

You can also consider converting your huge 640x480 raw file to DV for archiving AFTER you're finished editing.

edit: just re-read this thread more closely. Why are you using Powerdirector or Stoik at all if you have Premiere?

[Ranubis]

The problem is I don't have Firewire and the camera doesn't have a Firewire interface...

Atm I can't replace it for any of those options, maybe later...

I tried WinDv and it didn't work. I suppose it just works with Firewire interface..

I'm currently on a laptop and I can't run heavy software so even running Premiere Pro CS3 is a bit of a risk.

On my own computer I have a Firewire interface but I still have the problem of my camera. I can just output S-Video.

Thanks for the suggestions though

[sanlyn]

The problem is I don't have Firewire and the camera doesn't have a Firewire interface...

Good. Now the DV fanboys and artifact druggies can't talk you into ruining your tapes if you can't get them to DV.

I tried WinDv and it didn't work. I suppose it just works with Firewire interface..

Saved again.

I'm currently on a laptop and I can't run heavy software so even running Premiere Pro CS3 is a bit of a risk.

Other than cut your videos to size and do some color correction (you won't be doing color correction anyway), or getting a second-rate encode, you've again saved your tape collection ftom further quality loss.

On my own computer I have a Firewire interface but I still have the problem of my camera. I can just output S-Video.

Another fortunate turn. Now you won't have to degrade your images with dot crawl filters.

Why don't you get a DVD recorder? It's easier and faster, and high-birtrate DVD/MPEG2 at least won't kill your tapes any worse than DV would.

[Ranubis]

Wow, I didn't know it would degrade that much quality...

Well I can also convert to MPEG-2 with software

[sanlyn]

Wow, I didn't know it would degrade that much quality...

Well I can also convert to MPEG-2 with software

I might have overstated the case, which depends on the quality of the source. Retail analog sources are one thing, but most home recorded tapes and videos have problems that typically turn into garbage when they hit lossy encoders, especially if they're encoded twice. The old analog hardware was rather forgiving and not all that precise. Digital processing expects perfect sync and little or no noise. If your final goal is DVD or standard-def h264, the "typical" analog source requires some cleanup -- depending on the final quality that you expect. If your requirements aren't that high, you can lower the bar without losing sleep over it.

Much depends on your source. If you already have some captures that represent your source, you can submit a few seconds here for more precise answers. Otherwise, readers are really being very general because they have no idea of the condition of your source.

[Ranubis]

Wow, I didn't know it would degrade that much quality...

Well I can also convert to MPEG-2 with software

I might have overstated the case, which depends on the quality of the source. Retail analog sources are one thing, but most home recorded tapes and videos have problems that typically turn into garbage when they hit lossy encoders, especially if they're encoded twice. The old analog hardware was rather forgiving and not all that precise. Digital processing expects perfect sync and little or no noise. If your final goal is DVD or standard-def h264, the "typical" analog source requires some cleanup -- depending on the final quality that you expect. If your requirements aren't that high, you can lower the bar without losing sleep over it.

Much depends on your source. If you already have some captures that represent your source, you can submit a few seconds here for more precise answers. Otherwise, readers are really being very general because they have no idea of the condition of your source.

Thank you sanlyn for keeping the thread flowing with useful posts!

My final goal is: be able to capture in a format that I can edit and I mean really basic stuff, just cutting and cropping, no meddling with the colours. Also be able with that file to have the choice if someone wants to put in on a dvd.

If capturing the footage in MPEG-2 with the DVD HQ from PowerDirector is fine I'm comfortable with it since I really think all these huge avi files are a bit of a waste. I just need to know the best setting for the encoding.

I have the famous "head switching noise" and a stripe on the side that changes colur depending in the scene. I'm sure I can take those out...

See for yourself: http://netload.in/dateiO9XbWdHg77/2007v2.avi.htm

[sanlyn]

Border stains, rainbows, scratches, dropouts, etc., will likely require that you work in YUV and AVI. Head switching noise alone can be covered with a black border, which is what most people do, and many users mask edge stains the same way (if it's not too extensive); Premiere can do that in MPEG, and some other editors can do it. Altering the image along those lines requires re-rendering (re-encoding) -- unless other readers can suggest otherwise.

[Ranubis]

I think in this case I can just capture it in MPEG-2 and do a few removal on Premiere since it's easily done. Then I would just export without tampering much with the file settings.

So what should I do with those big avi files? delete them? Just think the quality of the video compared to the size it's a bit of a waste.

Also, what are the best settings for it?

http://img.photobucket.com/albums/v439/pharao/videosettings.jpg

[sanlyn]

Can't advise about Premiere settings, I use After Effects for image mods and color correction. The cost of Premiere is a bit high for edits and masking that one can do cheaper (or for free) in other software. It shouldn't be complicated for Premiere.

Most users would losslessly compress AVI with something like Lagarith, which considerably reduces AVI size without damage. External USB drives for storage are relatively inexpensive these days. They're popular for archiving and for temp storage of work files. The only original captures I've deleted were those that were too crappy to work with or not worth the trouble. I learned my lesson some years back about getting rid of captures that were still workable.

[Richard_G]

Also, what are the best settings for it?

Presumably you are asking what are the best settings for the mpeg capture? So you don't have to re-encode again when you create your DVD, you want to encode to DVD compliant mpeg2. Which for PAL is 720x576, 25fps (you've probably got no choice on that as your source will be 25fps) and something like 8,000kps bitrate. The DVD standard will allow you to go up to 9,800kps but that leaves you with no overhead for audio. Then when you come to author to DVD the file will already be compliant so won't need to be encoded (and degraded) a second time.

[sanlyn]

True. But the owner wants to apply border masks and other image modifications. When making edit cuts, Premiere re-renders only a few cut key frames, not the whole video. But masking the entire video presents another scenario. The whole thing would have to re-rendered, as far as I know. High bitrates are probably the only way around it.

[Ranubis]

sanlyn: if you can recommend a freeware to do the modifications I would be grateful. Premiere goes slow on this laptop.

So compressing with Lagarith and then editing would be your suggestion?

I really just want to get rid of the head switching noise and the border stains...

Richard_G: yes, exactly, that's what I wanted to know.

My quandary is: Should I capture with uncompressed AVI, compress with Lagarith and then edit? Or should I capture with MPEG-2, straight to the editing software and keep it digital or burn to a dvd?

Maybe I'm not being totally clear. I want to know wich route should I follow for keeping them just for storage (editing possible) and the route for burning to dvd. If I can capture them to MPEG-2 edit and still get decent quality that's fine, I'm totally fine with that!

[sanlyn]

If your captured AVI's are uncompressed (i.e, they're not DivX, MJPG, WMV, XVid, or any other lossy compressor/encoder) and are in a YUV colorspace (YV12, YUY2, etc,), you can archive them as-is to an external drive. If they are already encoded/compressed with something like the codecs just mentioned, you're better off leaving them as-is and archiving them that way. I see that they are described in post #1 as "AVI High Quality Profile". But AVI is a container, not a specific codec or compressor.

You can use the free MediaInfo utility to get the specs on a sample AVI, and post the text here. It's easy to use and doesn't modify files in any way. It's fast installer shows an ad for other software, but just ignore it if you wish. Without that basic info, no one could give more specific recommendations.

[Ranubis]

Thanks for the reminder, here are the details from MediaInfo:

General
Complete name : C:\Users\MATM\Documents\CyberLink\PowerDirector\10.0\2007.avi
Format : AVI
Format/Info : Audio Video Interleave
Format profile : OpenDML
File size : 113 GiB
Duration : 2h 8mn
Overall bit rate : 126 Mbps

Video

ID : 0
Format : YUV
Codec ID : YUY2
Codec ID/Info : YUV 4:2:2 as for UYVY but with different component ordering within the u_int32 macropixel
Duration : 2h 8mn
Bit rate : 124 Mbps
Width : 640 pixels
Height : 480 pixels
Display aspect ratio : 4:3
Frame rate : 25.453 fps
Color space : YUV
Chroma subsampling : 4:2:2
Compression mode : Lossless
Bits/(Pixel*Frame) : 15.906
Stream size : 111 GiB (99%)

Audio

ID : 1
Format : PCM
Format settings, Endianness : Little
Format settings, Sign : Signed
Codec ID : 1
Duration : 2h 8mn
Bit rate mode : Constant
Bit rate : 1 536 Kbps
Channel(s) : 2 channels
Sampling rate : 48.0 KHz
Bit depth : 16 bits
Stream size : 1.38 GiB (1%)
Alignment : Aligned on interleaves
Interleave, duration : 10 ms (0.25 video frame)
Interleave, preload duration : 130 ms

[sanlyn]

MediaInfo says that your 2-hour capture is uncompressed, with a version of YUY2. The original would be interlaced. You can reduce its size to about half with lossless compression in standard YUY2. I would suggest Lagarith, which is more widely compatible than huffyuv or some of the other lossless compressors. You must have Lagarith installed on your system. Get it here if you don't have it: http://lags.leetcode.net/codec.html.

You can open the file directly in VirtualDub.
1. in the top VirtualDUb menu, Click "Video" -> "color depth..." -> "4:2:2 YCbCr (YUY2)".
2. Click "Video" -> "compression..." -> "Lagarith Lossless Codec"
3. In the Lagarith dialog window, click "Configure" -> "YUY2" -> "OK", then "OK" again to close the remaining dialog.
4. Click "Video" -> "fast recompress".
5. Click "File" -> "Save as AVI...", give a filename and location.

The recompress should run at 100 fps or so (4 times the source frame rate, depending on your system), would take perhaps half an hour, and the results would be about half the size of your 2-hour capture.
You can also use Avisynth to open the file as YUY2/interlaced=true, but the instructions above should be OK with the latest 32-bit version of VirtualDub.

You can modify the AVI's bottom border in Avisynth with the Crop() and AddBorders() commands in YUY2 with Avisynth. Frankly, this sort of cleanup is best done with the original AVI rather than done in an NLE to maintain the proper colorspace for interlaced YUY2. Many NLE's and VirtualDub would convert this to RGB for that task, which you should avoid unless you do it properly with something like Avisynth. But fixing the borders would come later anyway and converting to RGB would nearly double the size again.

If you need disk space, I suggest that you get an external drive. It's really difficult to work a lot of video conversions without more storage space. Copy the original uncompressed captures to the external drive for safe keeping until your recompressions are complete and you have viewed the new compressions to ensure there are no errors. With Lagarith installed you should be able to play the videos in most media players, including WMP and VLC Player.

You can capture lossless AVI directly to YUY2 and specify Lagarith as the compressor. It is fast enough on modern PC's for real-time compression during capture, and was designed for that purpose. This saves you the extra re-compression step.

[Richard_G]

The resolution is wrong too, it should be 720x576 and the frame rate should be 25 fps and not 25.453.

[sanlyn]

It can be encoded later as PAL 720x576 but captured as 4:3 640x480 (the camera is not playing the video as 720x576). In Avisynth one could specify AssumeFPS(25) to set it on spec. Maybe the capture device or something in the capture chain was outputting that frame rate.

[Ranubis]

It worked like a charm! Thanks a bunch sanlyn! The only gripe I have is when I playback the newly converted/captured videos on VLC. They have a strange strobing light to them in changing colours like this: http://img.photobucket.com/albums/v439/pharao/vlcsnap-2013-06-21-12h54m55s183.png

On WMP that doesn't happen and the playback is flawless...

For some reason I can't find the AviSynth exe, maybe I'm doing something wrong I don't know. Just need some help on that.

Michael_G unfortunately I didn't have that resolution and decided to go for full 640x40 as you can see here:

http://img.photobucket.com/albums/v439/pharao/videosettings.jpg

The framerate was something that baffled me, go figure

[sanlyn]

Can't answer for the playback problem, as I'm not familiar with Power Director (I see many complaints about that software elsewhere). You referred to the video as "converted/captured", so I'm not sure what steps you've taken. Need a small chunk of that video to see what's happening. The Power Director capture you posted says the resolution is 720x576, not 640x480, and it's PAL. So I'm not sure which system you're working with, PAL or NTSC. In post #14 you said you're working with captured AVI's. Not certain which project you're discussing here.

You can't run Avisynth with Avisynth.exe (it doesn't exist). You write a script in Notepad and save that text file with an .avs file ending -- for example, "script.avs". Then open that avs file in VirtualDub. What the script would read and manipulate is a .d2v project file created with the free DGIndex utility. Below is a samplescript that would read a d2v file, deinterlace it, crop and resize. The resize in the sample is for NTSC. The pixel numbers to be cropped and the resize dimensions would be customized for the video you're working with:

MPEG2Source("drive"\path to d2v file\filename.d2v")
QTGMC("very fast").SelectEven()
Crop(0,46,0,-46) 
Spline36Resize(854,480)

There are probably 50 ways to do this, and ways to re-intelace the video if you want. But basically you have to join the VOB's without re-encoding, without quality loss, without re-processing the original lossy compressed audio and thus corrupting audio as well, without a lot of disturbing colorspace conversions, and finally resizing for a full-screen 16x9 image. If you're working with a captured AVI you don't need a .d2v, and the first line of the script would be different. You can do all the border cutting you want in TMPGEnc, but it won't resize the image properly for full screen 16x9.

[Ranubis]

Yeah, it's not the best there is that's for sure...

I've taken the steps you described. Be it converted video with lagarith or captured with it, when it plays back on VLC.

I will upload a video later so you can see!

Ok, I'll explore AviSynth and post back with results.

Thanks

[Ranubis]

Here's the video: http://netload.in/dateiDcvA5zalHu/Outubro2005aMaio2006sample.avi.ht

I really need more time to explore AviSyinth. Btw is there something similar to AviSynth but easier to handle?

There's something happening with VirtualDub that really annoys me. I can't change the framerate when I capture. It's currently capturing at 29 fps and I don't get any video, just sound. It was all of a sudden, because previously it was ok. Now it's always at 29 fps...

Any ideas? Thanks!

[sanlyn]

I can't change the framerate when I capture. It's currently capturing at 29 fps and I don't get any video, just sound. It was all of a sudden, because previously it was ok. Now it's always at 29 fps...

I can't say I've had that problem, but it's been years since I changed anything in VirtualDub capture except the output file name. In the capture window Try "Capture..." -> "Settings...", and in the "Capture Settings" panel set the frame rate to 25.000

The Outubro2005aMaio2006sample.avi had several problems. I fixed some (didn't have time for all of it). Poor color balance (too purple -- no clean blacks, grays, or whites). Bright contrast too high for video, so some brightest details are washed out, but not too serious. I took two steps to clear the right border stain, in Avisynth and then in VirtualDub. There are horizontal hum bars rolling vertically, could be caused by overmodulation (signal too strong) or grounding problems in the power source, etc.. The bars are less obvious in MPC-BE or Media Player. There are interlace problems as well, with split and ragged lines, dot crawl and aliasing on edges. I smoothed some of this, but it could stand more tweaking and sharpening. The repairs are most obvious in the third before-and-after image, below (left=before, right=after).

Original, frame 307

[Attachment 18467 - Click to enlarge]

[Attachment 18468 - Click to enlarge]

[Attachment 18469 - Click to enlarge]

Here is the Avisynth script I ran in VirtualDub. At this point is would make absolutely no sense, but you can pursue it if you want. Likely you don't have these filters....yet. After running this script I used Camcorder Color Denoise (setting=16), Color Mill, and gradation curves in VirtualDub. Everything needs more tweaking. There are also line-sync problems -- there's is shimmer and wiggle along borders and on angular and vertical lines (look at the poles and borders in the original). Some of the broken and ragged horizontal lines are partly caused by this problem.

AviSource("drive:\path to video\Outubro2005aMaio2006sample.avi")
FluxSmoothT()
Santiag()
ConvertToYV12(interlaced=true)
assumeTFF().SeparateFields()
a=last
 
a
chubbyrain2()
smoothuv(radius=7)
# ---- crop all except the right border area ----
crop(620,0,0,0,true)
TweakColor(sat=0.6,startHue=200,endHue=260,maxSat=75,minSat=50)  #green
TweakColor(sat=1.5,startHue=20,endHue=80,maxSat=75,minSat=50,smooth=0)   #magenta
Tweak(coring=false,sat=1.3,Hue=15)
awarpsharp2(depth=30)
ChromaShift(L=4,v=2).AddBorders(0,0,0,4).ChromaShift(L=-4).Crop(0,0,0,-4)
b=last

# ----- overlay the cleaned border (b) onto the original video (a) -----
overlay(a,b,x=620)
weave()
ColorYUV(off_y=-5,cont_v=-20,cont_u=-60,off_u=-2,off_v=-3)
ConvertToRGB32(matrix="Rec601",interlaced=true)
AssumeFPS(25)
Crop(4,0,-14,-9).AddBorders(9,5,9,4)
return last

function ChubbyRain2(clip c, int "th", int "radius", bool "show", int "sft")
{
# ----- based on Mug Funky's ChubbyRain --------
th = default(th,10)
radius = default(radius,10)
show = default(show,false)
sft = default (sft, 10)
u = c.utoy()
v = c.vtoy()
uc = u.mt_convolution(horizontal="1",vertical="1 -2 1",Y=3,U=0,V=0)
vc = v.mt_convolution(horizontal="1",vertical="1 -2 1",Y=3,U=0,V=0)
cc = c.mt_convolution(horizontal="1",vertical="1 2 1",Y=2,U=3,V=3).bifrost(interlaced=false).cnr2().temporalsoften(radius,0,sft,2,2)
rainbow=mt_lutxy(uc,vc,Yexpr=string("x y + "+string(th)+" > 256 0 ?")).pointresize(c.width,c.height).mt_expand(y=3,u=-128,v=-128)#.blur(1.5)
overlay(c,cc,mask=rainbow)
show==true? rainbow : last
}

[sanlyn]

...And with more time and extra help (QTGMC and a NNEDI3 routine to help reduce jaggies) and the dither plugin to help maintain color depth, results can be improved. But it does look as if your AVI sample has either been incorrectly deinterlaced and reinterlaced earlier, or some capture hardware doesn't treat interlace correctly. Something has occurred to make a mess of interlacing and has cost some detail. The attached effort looks a bit better than the above, but little else can be done with bad interlacing.

[Ranubis]

I really can't change it... it keeps forcing 29 fps. Unfortunately I had to get back to the old capturing routine and then converting with VirtualDub... Some things are really weird, like when I capture with the PowerDirector sometimes I get videos with 26 fps or even 33?!

Thanks for all the effort on the videos. Unfortunately I don't have the time to do all those enhancements. At least not for now. But from what you say the problems are hardware related and not inherent to the tape. It appears that the capture system is the culprit but I really couldn't say... I wish I had something better but for now it has to do the trick.


I'm still pretty green to AviSynth and it will take sometime to get into it. Is there no other software I can run for now? to perform the same operations you are performing with it?