How can I losslessly capture video from an 8mm tape?

[Drew Neilson]

How can I losslessly capture NTSC video from an 8mm tape that was recorded at SP speed?
I'm new to this site. I'm considering buying a new hard drive for my computer; one reason is so that I can losslessly capture video from my old 8mm (not Hi8 nor digital 8) camcorder video tapes, edit the video, and then share the video with other people (perhaps by burning to DVD, perhaps by posting it to YouTube, or perhaps by sharing it via other methods). Regardless of the method of sharing, I want my digital master copy to be lossless. My first thought was that in order to make an educated guess about how much storage space I'd need for the video, I'd need to know what resolution the analog video is recorded in, since I want to transfer it as perfectly as possible. I also need to know whether USB 3.1 Gen. 1 is fast enough to capture this video losslessly. I know very little about analog video. I'm a bit more familiar with digital video (digital video resolutions, frame rates, progressive vs interlaced, the concepts of lossless compression and lossy compression, and when it comes to screens, I've heard about brightness dynamic range and color space). The phrase 'heard about' in that last sentence is the operative phrase--I'm certainly not an expert in any of this stuff.
Wikipedia's page about 8mm video tapes doesn't currently specify their resolution. I saw the following:
"Kodak released first 8mm camcorder in 1984. 255 lines of resolution. Maximum tape length is 120 minutes in SP mode or 240 minutes in LP mode." Source: http://tech-notes.tv/Standards-Practices/TVTapeformats.htm
"Then in the 1980's 8mm video tape cassettes (also known as Video 8) came out. Following that came Hi8. This was a higher resolution from 8mm or Video 8 cassettes. Hi 8 had 400 lines of resolution compared to 240 that 8mm/Video 8 had." Source: https://www.videoconversionexperts.com/8mm-to-dvd.html
I also realized that it might be more complicated than a simple resolution, and if that is true, then the information that I stated above might be incomplete. The luma resolution might be different from the chroma resolution… and I barely know what those things are. In the world of digital video, I have heard of chroma sub-sampling.
I don't know what, if any, impact the speed of the tape while recording has on the resolution. For VHS, I read that "For NTSC, LP and EP/SLP doubles and triples the recording time accordingly, but these speed reductions cause a reduction in video quality – from the normal 250 lines in SP, to 230 analog lines horizontal in LP and even less in EP/SLP." Source: https://en.wikipedia.org/wiki/VHS
I haven't even begun to think about audio yet (for example: what bit depth and sampling rate to use to capture the audio in it's fullest fidelity). Most or all of the audio on those tapes is mono; I doubt I recorded any of the video with stereo audio.
My computer does not have Firewire.

[MiniDVLover]

If you're looking for as lossless as possible, I would definitely say FireWire is the way to go. If your computer has an ExpressCard, PCMCIA, or open PCI slot, they make FireWire cards that plug into those ports that you can plug a FireWire cable into. They do make USB-to-FireWire adapters, but I have no experience with them.

If you have a Digital8 camcorder that can play analog tapes, you can just plug that directly into said card.

That's what I have always done. You CAN use an analog capture card for this, but make sure it supports DV-AVI. That's the best format for analog video capture, since it has the same resolution and frame rate that analog video has, and supports interlacing (you will definitely see interlacing artifacts with MPEG-4). DV was designed to work with analog video. It's also ideal for editing as it is intraframe compressed.

If you have an analog camcorder, you can use an analog capture card or plug the camcorder into a device that converts A/V to DV (Sony's Digital8, MiniDV, and Canon's MiniDV camcorders usually do). If said camcorder has S-Video, use that. Otherwise, you can use standard composite (yellow connector). But S-Video will yield better video.

As far as capture software goes, my favorite is Scenalyzer, since you can edit and check the video right there. It supports both DV and analog capture cards.

Hope that helps.

[JVRaines]

If you're looking for as lossless as possible, I would definitely say FireWire is the way to go.

I'm afraid this simply isn't true. The IEEE1394 (FireWire) port will give you DV-encoded video, which uses lossy intraframe compression and (for NTSC) suffers from what many find to be subpar 4:1:1 color decimation. A decent USB dongle or PCIe card (search the forum) will provide uncompressed digital video that can be stored as-is or in a losslessly-compressed format, depending on the speed of your processor and storage device.

[MiniDVLover]

Okay, yes, DV is compressed. However, 8mm's resolution is quite limited already, so DV won't make it look any worse. And I haven't found 4:1:1 compression to be terrible.

[amaipaipai]

Okay, yes, DV is compressed. However, 8mm's resolution is quite limited already, so DV won't make it look any worse. And I haven't found 4:1:1 compression to be terrible.

It's awful actually:
http://www.kenstone.net/fcp_homepage/chroma_investigation_nattress.html

[stunner21]

Not quite sure why you would want to capture firewire in this day and age if hes going for lossless. DV for DV tape would be alright as thats what it was created for but anything analogue would be handicapped from the outset if your limiting what your input comes in at. The resolution of 8mm is limited yes but as JVRaines said by compressing to 4:1:1 you'd also be limiting colour information as well which I would avoid doing for a master unless the footage was captured like that originally. Personally I use the blackmagic ultrastudio which works well enough for me coming in via thunderbolt either uncompressed avi or dpx image sequence. I know aja cards are decent on the higher end plus there are a load of older cards that guys on here have recommended in the past which are supposed to do a decent job for analogue sources but are somewhat harder to operate in this day and age due to driver compatibility.

[MiniDVLover]

I'm a DV enthusiast, in case you can't tell by my username. I wouldn't notice the differences shown in that link with an analog video capture unless you broke them up (and the color resolution is really reduced on 8mm, much more so than DV).

I guess I'm just a bit behind on the times. I've always done 8mm transfers through FireWire and have had no DV-related quality issues whatsoever. If there's a better way to do it, go for it. I just don't want to introduce any artifacts such as dot crawl or color bleeding that you'd get from composite video.

[JVRaines]

I did many, many transfers with a Canopus DV box and thought I had no quality issues until I started using a different converter. Then I noticed that DV colors seemed pale and sloppy in comparison to 4:2:2 subsampling. Outside a few horrid artifacts (try chain-link fence), the difference is not very noticeable to the casual eye.

When comparing analog to digital resolution, a simple fact is often overlooked: analog doesn't have fixed vertical resolution boundaries. While Video8 may provide only 240 lines of luminance (and even fewer for color), the change from one level to another can occur at any one of an infinite number of points along the scan line. The higher the sampling resolution, the less aliasing will occur.

Finally, I would point out that serious practitioners avoid composite video output whenever possible, regardless of the digital standard in use. S-video (Y/C) and component are much preferred and are supported by many hardware devices.

[KarMa]

I guess I'm just a bit behind on the times. I've always done 8mm transfers through FireWire and have had no DV-related quality issues whatsoever. If there's a better way to do it, go for it. I just don't want to introduce any artifacts such as dot crawl or color bleeding that you'd get from composite video.

Color bleeding would be built into the source material (8mm in this case) but might be cleaned up a bit by hardware Noise Reducers. As for dot crawl, that can be reduced or removed by outputting the 8mm video with S-Video and having a capture card with a S-Video port. If you can't output S-Video but only Composite, then having a capture device with a 3D combfilter will help a lot.

[MiniDVLover]

I did many, many transfers with a Canopus DV box and thought I had no quality issues until I started using a different converter. Then I noticed that DV colors seemed pale and sloppy in comparison to 4:2:2 subsampling. Outside a few horrid artifacts (try chain-link fence), the difference is not very noticeable to the casual eye.

When comparing analog to digital resolution, a simple fact is often overlooked: analog doesn't have fixed vertical resolution boundaries. While Video8 may provide only 240 lines of luminance (and even fewer for color), the change from one level to another can occur at any one of an infinite number of points along the scan line. The higher the sampling resolution, the less aliasing will occur.

Finally, I would point out that serious practitioners avoid composite video output whenever possible, regardless of the digital standard in use. S-video (Y/C) and component are much preferred and are supported by many hardware devices.

Very interesting. I always thought FireWire was the way to go. Obviously, it is for MiniDV and Digital8.

I do notice some artifacts when I convert an HD video to DV with After Effects, but I have to be really zoomed in to see them. Of course, with analog video, it's hard to tell which artifacts are DV-based and which are analog tape-based.

[LMotlow]

If you can't tell the difference between a lossy encode and lossless quality, that's no excuse for imposing your lower standards onto those with a keener eye and more stringent requirements.

[MiniDVLover]

I'm not trying to impose anything - I didn't know that they were considered low and was just trying to help Drew out.

So going back to his original question. How would you capture 8mm video?

[Eric-jan]

my guess.... a 8mm-dvd recorder combo, you transfer 8mm directly to dvd, dvd has a higher resolution than 8mm video anyway, and you don't have any composite/s-video artefacts, because you don't use them as passthrough.
8mm video is a very old format, there wasn't much any other hardware then camcorders, players and combo'are even more rare, but these are the only options, spare parts for repairs come only from defect equipment.

[davexnet]

I used to use an AverMedia M780 PCIe card. Had a choice of hardware encoded mpeg-2 (great if you want to go to dvd with only light editing)
Or uncompressed using Huffyuv or similar. Lots of used stuff in Ebay. Whether these devices work in Windows 10 is another question

[lordsmurf]

It's awful actually:
http://www.kenstone.net/fcp_homepage/chroma_investigation_nattress.html

With VHS source, it's far worse than that. He used BetacamSP.

I guess I'm just a bit behind on the times. I've always done 8mm transfers through FireWire and have had no DV-related quality issues whatsoever. If there's a better way to do it, go for it. I just don't want to introduce any artifacts such as dot crawl or color bleeding that you'd get from composite video.

Instead, you intruduce macroblock noise at the capture, and lose 50%+ color data.

Lossless was a compromise in quality for computer of its generation -- Pentium III, small 4200rpm IDE drives, 500Mhz max CPU, 1gb max RAM. In other words, 1990s specs, not even 2000s, much less 2010s.

and you don't have any composite/s-video artefacts, because you don't use them as passthrough..

Internal connections/signal are essentially s-video. Separated chroma and luma carrier.

[Eric-jan]

maybe the very first tape, original was recorded with different sync signals ? and the copies where faulty after that ?

[Drew Neilson]

Is my 2017 Surface Pro an appropriate device for capturing and editing video from 8mm camcorder tapes?

Specs:
CPU: 7th-generation Intel Core i5 7300U (a dual-core mobile, not desktop, chip)
GPU: Intel HD 620
USB 3.1 Gen. 1
- 1 port on the tablet, and a few (four, I think) on the Surface Dock I purchased.
RAM: 8 GB
Storage:
- 256 GB SSD: Samsung kus030202m-b000
- microSDXC card slot
- I'm planning on buying a Blu-ray burner drive
Display:
- I'm using its sRGB mode, not Enhanced sRGB mode.
- https://www.anandtech.com/show/11538/the-microsoft-surface-pro-2017-review-evolution/6 --> note how much they praise this tablet's screen.

[LMotlow]

As near as I've ever been able to gather, fans of analog-to-DV have never been able to make a lossless capture and most have never seen losslessly processed video.This debate has gone on for decades, with the pros still saying that analog to lossless beats analog to DV for cleaner results every time. It also never fails that every time someone comes to this forum and asks how to get lossless captures, the DV crowd always shows up to poobah lossless. And every time a DV fan shows up and says how great their analog-to-DV looks the lossless crowd shows up to poobah analog-to-DV. You also get the HDMI crowd and the h.264 and h.265 crowds, who always post some godawful multiple-encode trash sample to YouTube as if YT were some kind of high quality standard. We've seen lossless project posts without no compression artifacts to fix in the samples, and we've seen DV posts with obvious compression effectys, fried color, plastic detail, and gone through thousands of lines of wild, Avisynth hairy scripts to clean up and get output that still looks like over- processed DV at the end.

And here we are again. Different decade, different year, different day, same crap.

Hopefully the O.P. will decide and we can all help clean up the mess, whatever it turns out to be.

[lordsmurf]

Is my 2017 Surface Pro an appropriate device for capturing and editing video f
Display:
- I'm using its sRGB mode, not Enhanced sRGB mode.

Not with that display. It's about as inaccurate as can be. Video you correct on there will wildly differ from what you see on TV screen, or even another computer.

Trying to capture on a tablet isn't a great experience, not much different that using an old P4 from eons ago. I've done it, for testing only, and there are conflicts and slowdowns. A tablet is a weak laptop, and a laptop is a weak desktop. Wrong tool for the task.

[Drew Neilson]

Not with that display. It's about as inaccurate as can be. Video you correct on there will wildly differ from what you see on TV screen, or even another computer.

If my screen is inaccurate, how do you account for AnandTech's positive review of my screen?

Trying to capture on a tablet isn't a great experience, not much different that using an old P4 from eons ago. I've done it, for testing only, and there are conflicts and slowdowns. A tablet is a weak laptop, and a laptop is a weak desktop. Wrong tool for the task.

What setup do you suggest I use? As I said in my original post, I want to capture the video from my 8mm video tapes losslessly. What hardware specifications will I need in order to do that? Or is there a place I can send my tapes to that will capture my video losslessly?

When I said "lossless", I meant it in the literal sense, not just 'it looks really good'. I want perfection, or at least, I want to know if perfection is possible, and if so, how much will it cost. That's why I started out by trying to find out how many lines of resolution 8mm NTSC video recorded at SP speed has, so that I can capture the video at that resolution, neither lower nor higher. I wouldn't want to capture it at a lower resolution and lose quality, and I wouldn't want to capture it at a higher resolution and waste space.

[davexnet]

XThe source is an analog tape, it is converted to digital in the analog to digital converter in the capture device
but the quality of the ADC does play some part, as does everything else before that. The condition of the tape, the playback head,
tracking, the cables, whether a time base corrector or other signal processor is used, etc, etc.
When the output from the ADC is captured as-is and put into a file, this is considered lossless.

If you convert to DVD later, you'll need to convert to mpeg-2 @ 352x480 or 720x480. So I would say your capture should be either
of these. Whether they'll be any difference in perceivable quality, maybe, maybe not. There is no perfection in analog capture,
only compromises

[JimTheGreek]

Hello to everybody!

I have been looking for a while, to find what to choose for a similar dilemma. I have many 8mm and miniDV tapes to digitise (8mm are way more than miniDV ones) and bearing in mind all the factors that can affect final result for each method (cables, head of camcorder, capturing devices etc) I would like to ask if anybody can help me with putting into an hierarchy all the methods, from the best to the worse one. I find it difficult to decide on my own, as I don't know the impact of different methods on the final result. I am an amateur on the video field, and I am lacking of basic knowledge, as I am not sure if upscaling the resolution of the input signal, could give a better quality output video. I know that the camera those tapes were recorded with, gives PAL video signal (it was a Sony Handycam CCD-SC6E). I also know that component cable is better than S-Video, which is better than the Composite one, regarding the way to transfer the signal from camcorder.

My equipment at present time is:
A Sony Handycam Video8 CCD-TRV35E PAL camcorder (has only composite output)
A usb 2.0 video grabber with composite & S-Video input
Α Mac computer

Most common methods are the following, as far as I am aware of:

A. 8mm tape > Camcorder Video8 CCD-TRV35E PAL (its the one I own) > USB 2.0 (or 3.0) video grabber device through composite cable or S-Video cable > Computer with appropriate Software.
B. 8mm tape > Camcorder Video8 CCD-TRV35E PAL (its the one I own) > (via S-Video or RCA) to DV camcorder with passthrough capability, producing DV AVI signal > (via FireWire connection) Computer and appropriate software.
C. 8mm tape > Digital8 Camcorder (more difficult to find and a bit pricey) with output DV AVI signal > (via FireWire connection) Computer and appropriate software.
D. 8mm tape > Camcorder Video8 CCD-TRV35E PAL with composite output > a quality capture device like Blackmagic Design Intensity Shuttle with Thunderbolt> (via Thunderbolt connection) Computer and appropriate software.

Budgetwise, if difference in quality among methods is reasonable, I would consider buying an old Digital8 camcorder, the suitable cables, or an expensive capture device like the above.

Thanks in advance. Hope this is not too complicated

[Eric-jan]

The IS needs a clean video signal, the CCD-TRV35E can't provide that.
Don't know if there's good DV "tansfer" software for MAC... or which interface dongle will work for MAC...

[dellsam34]

If you are capturing analog tapes stay away from DV, DV is an obsolete tape format not a capture codec. If CCD-TRV35E does not have line TBC get one that does even if it's Digital8 as long as it plays analog tapes, has S-Video out and line TBC and capture with method A.

[Cornucopia]

I did many, many transfers with a Canopus DV box and thought I had no quality issues until I started using a different converter. Then I noticed that DV colors seemed pale and sloppy in comparison to 4:2:2 subsampling. Outside a few horrid artifacts (try chain-link fence), the difference is not very noticeable to the casual eye.

When comparing analog to digital resolution, a simple fact is often overlooked: analog doesn't have fixed vertical resolution boundaries. While Video8 may provide only 240 lines of luminance (and even fewer for color), the change from one level to another can occur at any one of an infinite number of points along the scan line. The higher the sampling resolution, the less aliasing will occur.

Finally, I would point out that serious practitioners avoid composite video output whenever possible, regardless of the digital standard in use. S-video (Y/C) and component are much preferred and are supported by many hardware devices.

@JVRaines, I think you meant to say HORIZONTAL resolution boundaries - the resolution within each scanline.

Analog does have fixed, discreet, exact resolution in the vertical dimension, as that is the # of scanlines (525 total/486visible in NTSC, 625/576 in PAL).

There is still a realistic limit to the rez, based on optical resolving power of the lens & sensor, and the bandwidth available. And the digital figures decided upon fairly accurately represent this.

Scott

[Brad]

When comparing analog to digital resolution, a simple fact is often overlooked: analog doesn't have fixed vertical resolution boundaries. While Video8 may provide only 240 lines of luminance (and even fewer for color), the change from one level to another can occur at any one of an infinite number of points along the scan line. The higher the sampling resolution, the less aliasing will occur.

@JVRaines, I think you meant to say HORIZONTAL resolution boundaries - the resolution within each scanline.

You're replying to a post from 2018, thanks to the necro. But if we're nitpicking, I'm more concerned by the section I bolded. To me, this implies that we should be able to, for example: have an analog scanline where 3/4 of the scan time is spent on one white-to-black-to-white transition and the other 239 can somehow be shoved into the last 1/4.

[Cornucopia]

Doh! Thanks for pointing that out to me. Gotta check that in future. But I don't remember the necro waning coming into play - perhaps it was within the threshold.

..............................


Theorectically, that non-linear scanning might be possible, but is totally impractical ESPECIALLY when one would get to the point of downsampling back to standard, equal-pixel-width, linear scanning, as there either would be major overhead in the interim, or there would be unequal sensitivity, or there would be issues of bitdepth budgeting. And how to downsample appropriately, particularly if the element is fully within a single pixel width and might normally be negated by averaging.
I would assume there could also be positional artifacts due to the nonlinearity (as there are motion artifacts due to the nonlinearity of VFR).

Scott

[JimTheGreek]

The IS needs a clean video signal, the CCD-TRV35E can't provide that.
Don't know if there's good DV "tansfer" software for MAC... or which interface dongle will work for MAC...

What is needed for a clean enough signal then? A Digital8 camcorder, a passthrough capable camcorder to make the signal digital and send it via cable to computer?

As for software, I believe that video capture function in Final Cut Pro, or Adobe Premier would work via a FireWire. Or it won't?

[JimTheGreek]

If you are capturing analog tapes stay away from DV, DV is an obsolete tape format not a capture codec. If CCD-TRV35E does not have line TBC get one that does even if it's Digital8 as long as it plays analog tapes, has S-Video out and line TBC and capture with method A.

I understand that DV is obsolete, but I don't want to save the analog signal of the 8mm tapes to DV tapes (not sure I understood your point ). All I was thinking of, is the usage of a newer analogue camcorder than mine, to use it as a passthrough and give a better signal to my computer.

As for the best method, would the A method deliver a good quality via S-Video? Because I tried to make it work, using my old Video8 camcorder, giving signal to the USB video grabber via composite connection that has available, and then from the grabber, sending the signal via usb to computer. The quality is not good at all. The software coming with the grabber has 720X576 maximum resolution, h.264 compression, and thus, a 1 1/2 hours movie gets 40GBs of storage....I mean, will a Digital8 camcorder and a S-Video connection make any difference?

[Brad]

Attach a small sample of the results you're getting now. Capture a segment under 15 seconds, ideally a relatively still shot containing objects like "vertical columns" (lamp posts, door frames, the letter "I" in the title of a movie, etc.) and attach using the Upload files button on the forum Reply feature.