Best hardware workflow for tape transfers?
 

I purchased a video transfer business over 2 years ago and have since transferred thousands of VHS, Hi8, and MiniDV tapes. The previous owners had all of their VHS, Hi8, and MiniDV transfers revolve around a Panasonic DMR-EH59 DVD burner. VHS would come in via svideo, and Hi8 and MiniDV source would come in from Sony GV-D800s and Sony GV-HD700s via firewire. This system was reliable and the hard-drive intermediate on the DVD burner enabled trimming and content removal from the transfers before burning to the DVD. When clients wanted a digital file (ie MP4) version, we use a Nimbie DVD autoloader (which I added to the workflow to replace manual DVD loading) to rip the contents of the DVDs and then use Handbrake to re-encode the VOB files into MP4 (H264) files. I also use the Decomb filter in Handbrake to de-interlace the footage. (PS. I know that de-interlacing is destructive but I feel that providing clients with interlaced files is a dis-service. Interlaced video is obsolete, most people don't even know what it is, and interlaced content will only cause issues when they want to edit their content down the road in the modern progressive world, where they would probably choose a poorer de-interlacing method than I provide. Handbrake seemed to produce better de-interlacing results than a number of other options I explored. I may explore using the FieldKit plugin in Premiere in the future, but currently it's just too slow on even a very fast PC)

I have valued the advice that moderators such as LordSmurf have provided on this forum. I have used recommendations from here when purchasing 'new' VHS machines to replace broken or failing equipment that I inherited (I've focused on getting the Panasonic AG-1980s with a few JVC units sprinkled in).

I'm now in the process of switching from the DVD-centric transfers to using old Windows 7 computers to go directly to digital files via DV workflow. Windows 7 because the firewire drivers seem more stable in this OS. I've read in a number of places, here included, that DV transfers are not the best due to the 4:1:1 sampling rates, but this all feels a bit overkill when looking at the quality of content I transfer. Most is poor and I think this level of concern is only valid if I was regularly transferring commercially recorded VHS tapes. Even then, I think you've have to pixel-peep to notice the difference between 4:1:1 and 4:2:2 average VHS tape transfers. I'm connecting my Sony GV-D800s and Sony GV-HD700 via firewire to these old computers, and using WinDV to capture, which is working very well. But with VHS transfers I'm forced to convert the analog VHS signal to digital with other hardware. The previous owners had some Canopus ADVC-300 units, so I've tried using them with good results. Now I'm reading here that I should turn off as much of it's filtering as possible, which I will.

I did purchase 4 different analog to digital converters as comparison to the Canopus ADVC-300 ranging from $20 to $120 but none of them produced the same quality. Most forced a direct MP4 transfer with very low bit-rates that produced unacceptable artifacting. I'm sure there are better analog to digital converters than the 'top-end' Elgato unit I bought, but, to be honest, I couldn't find any being sold in 2020 that didn't involve some sort of HD up-res, which I want to avoid. And with the Canopus units I can have the same DV workflow as the Hi8 and MiniDV machines.... generating DV AVI files that I then convert to MP4 files via Handbrake as I did before from the DVD VOB files.

I've also read here that I should be using a separate TBC unit, even if I go with a DV workflow, but again, this seems like overkill for the quality of content I'm transferring from the average client. The TBC on the Panasonic AG-1980s seems to provide some help in certain situations, but usually, when a tape plays poorly, there seems to be no magic bullet to avoid tearing or breakup during some sections of the tape. Tape path movement in VHS machines seems to vary greatly and often the best solution for a problem tape is just to try the transfer in a different machine, even an identical model. And it's very unclear to me what value a separate TBC unit would provide on Hi8 and MiniDV tapes, which either often provide a clear steady video signal or don't. The biggest issue I have is with old Video8 tapes that shed oxide onto the heads, forcing me to clean the player heads a number of times during these problem tapes.

Anyways, I'm looking for opinions on my VHS/Hi8/MiniDV to MP4 workflow. I want to create at least 10 stations capable of VHS/Hi8/MiniDV to MP4 transfers, so adding expensive equipment for very marginal benefits is not desirable. I want to find that balance between cost, ease of transfer, and quality. Maybe I spend more money on one or two stations for those difficult tapes?

And you actually charged money for this? Really? And the victims paid you? Holy crap!

At least you've given our readers a detailed course in how it should not be done.

Saying that the source "isn't that good anyway" is an excuse, not a reason, regarding NTSC DV transfer. In fact, the worse the source is, the more damaging that the color loss is.

In the 1990s, back when Pentium II and Pentium III existed, DV was the crutch to allow digital video. It adds blocks, and it reduces color. It is compression, and in fact more lossy than H.264. The 90s had hardware appliances and DV, the 2000s had lossless and MPEG, and the 2010s (HD) had ProRes, DNxHD, and AVCHD (H.264). DV < MPEG < AVC/H.264.

But the problem is that SD video is era locked. The software, hardware, OS, codecs, methods. But you should lock it to the 2000s, not the 90s. The 2010s saw no real improvements to capture, and in fact saw a steep decline. The only advances to SD video was in restoration methods (Avisynth, mostly). Not capture.

Pixel-peeing is not needed. All you need is to view it on a now-common/standard large HDTV set. It's obvious even to tech-phobic grandmas.

You need to specific what "MP4" means.
What exact codec, what processes are done to the video, what settings are used.

You cannot just buy any random capture cards, and expect quality. The ADVC-300 is terrible device, and I can easily outperform it with other cards. Even the much-maligned (and deservedly so) Blackmagic cards are better. That's a statement! (I'm not entirely anti-DV, but it is on a case-by-case basis. Using DV for commercial work would not be a situation that gets my approval.)

I'm starting to think your disconnect with quality is with the software being used. Software that forces capturing to "MP4" (random compressed video) is never suggested.

Handbrake has so many flaws and problems. Seriously, learn to adopt Hybrid. It's a Swiss knife of encoding, including thes amazing amount of included Avisynth and Vapoursynth (and more). I cannot bow down to selur (the author) enough for his great contribution to the video community. It is a serious masterpiece of video software.

External TBC is not overkill. Internal line/fields TBCs clean the image, not the signal. The signal causes frame drops, audio sync, and other negative issues with capture and quality. You must have both. Capture workflows require some form of TBC. Some novices try to get by with only a TBC(ish) device like the ES10/15 alone, but many tapes still often fail to capture correctly, or at all. As a professional service, that will never work. At least not if you're trying to provide quality.

Tearing is easy to fix. ES10/15, and done, about 95% of the time.

Hi8/Video8 loves to drop frames. Many, many frames. If you've been using DV methods, that info has been hidden from you. Quality was lost, whether you know it or not.

NJRoadfan has discussed Hi8 oxide shedders before, in more details. Why it happens, tips and tricks, prevention, etc. Not fun, been there, done that. Just a week ago, in fact.

Interlace is not obsolete. That's nonsense. Turn on your TV right now. Almost any channel you see is interlaced HD. Interlacing is not bad or old, it's just a new concept to many people. In fact, interlacing is used in other areas of science, such as astronomy, to reduce bandwidth as the images are sent back to Earth. The entire reasoning that interlace exists is for bandwidth reduction.

When deinterlacing, only use QTGMC. Any else is a disservice to clients. And QTGMC is no longer Avisynth-only, and is easy to use in Hybrid when encoding out a file copy. (Tip: use "faster" preset.) Handbrake uses Yadif, which wasn't bad 10+ years ago (pre-QTGMC), but really lousy now. Premiere, any NLE, truly sucks at deinterlacing. Never use NLEs for that task.

Also read this: What’s in a Professional Video Workflow to Convert Analog Videotapes?

And this series:
- Video Hobby vs. Video Profession, Part 1: What’s the Difference?
- Video Hobby vs. Video Profession, Part 2: Make Money Converting Tapes to DVDs?
- Video Hobby vs. Video Profession, Part 3: The Interview

RE: Sanlyn.... Wasn't expecting such an unfriendly response here, but shouldn't be surprised by anything on the internet these days. Anyways, as I explained in my post, the initial setup was inherited from the previous owners and I'm looking for ways to improve.

Thanks LordSmurf for your kind and thorough response, even if some of my workflow horrified you. :) I'll digest and respond to your comments later tonight when I have more time.

sanlyn has grown weary of fake "professionals" (folks do low-quality work) over the years, as have I.

But the key difference here is that you're aware of weaknesses in your workflow (though still some room to grow that knowledge, based on some of your comments). As mentioned, you inherited some. You're attempting to improve it. Get feedback. Get help. Likely suggestions on what to buy, where, and on realistic budgeting. So for that, kudos are in order. Learn, buy the good gear to do a good job.

I'm 99% sure he didn't make it past the first paragraph, screamed into the monitor, and hit quote+reply. I can't blame him. He just needs to re-read it, the whole post, and realize your intention.

Some folks over at VH recently have put on their "big boy pants", and start spouting nonsense, confusing newbies, because they knew not what they were saying. Several of us had to respond to that nonsense (I was PM'd multiple times). I'm certain that's where this reaction came from. Several chuckleheads in a short timeframe have put us on edge. But this site isn't that one. We tend to attract more professionals and serious hobbyists, in terms of tape conversion users and conversations.

So don't be too insulted. :2cents:

@sanlyn: Read it again. He desires to improve. Let's help him do it. :cool:

I'll have to review the resources you've listed, so forgive me if some of my questions below are answered there.

Yes, just because most of the content I transfer is poor, I don't want to degrade it further. The reason for starting with a DV workflow in my current transition from my old DVD-centric transfers is because it's the equipment I had on hand. I also found plenty of advice online advocating firewire and WinDV to capture DV content direct from DV players/cameras. It's not trivial to sort the hobbyists from the professionals on the internet. I genuinely want quality for my clients and labour many hours on some jobs far beyond the financial-sense point because I want their memories to be preserved well.

Do you have any examples that compare a DV 4:1:1 workflow output vs a 4:2:2 workflow on consumer VHS video? I did a test with 3 friends in the film industry of the analog to digital converters I bought, and they didn't notice many of the issues I saw in the tests, Canopus ADVC-300 included.

By MP4 I mean the MP4 container with H264 codec at a variable bit-rate in Handbrake focused on achieving a certain level of quality (setting of 18). Typically this results in a average bitrate of between 4200 and 5000 kbps. I've been thinking of switching to H265.

If DV is so poor, which cards/hardware for analog to digital conversion do you recommend (through svideo ports I'm assuming)? I honestly looked and it seems you are correct in saying that the 2000's have the best hardware. Everything you can buy now seems like junk.... or close to. The Elgato convertor I bought had polished software but still forced you into a highly compressed MP4 (H264) format. I guess I'm forced to buy expensive used equipment for the digital to analog process?

I'll have to check out the software that you list. I had tried to use avisynth, but it couldn't handle the type I AVIs I was outputting from WinDV. I had to use type I because type II was giving me audio drift. It's important that whatever software I use for encoding has queuing features. Handbrake has great queuing support.

Regarding the de-interlacing. I use the 'Decomb' setting in Handbrake, not Yadif.... but maybe Decomb is just another variation of Yadif? I can't find technical information on what exactly 'Decomb' is, only that it visually outperformed all the programs I tested. You are certainly right that Premiere does a crappy job (it just doubles up one of the fields. I'll look into QTGMC and Hybrid.

Regarding your comment about de-interlacing not being obsolete... I have to partially disagree, certainly in consumer video acquisition and in the filmmaking community interlacing hasn't existed for almost a decade. I last bought a interlacing camera in 2005. All our films that we shot since 2008 have been progressive on DSLRs, and now progressive on Red cameras in 6k+. Sure, interlacing may still be used in television broadcasts, but most people stream content from Netflix, Disney, Amazon, Hulu, HBO, etc, and while my Amazon Firestick can display interlaced content (by first deinterlacing it), the majority is progressive. Less and less people have cable. It seems we live in different content worlds if most of your content is interlaced. Sports may be the only interlaced video I watch, aside from a little news, although with this Covid-19 crisis I watch less and less of that and no sports unfortunately :(.

And interlacing is bad if you have to display it on a progressive display (like many displays these days.... smart phones, tablets....) It just will never look quite right. Eventually, all displays will stop deinterlacing interlaced content. I'm sure there will be uses for interlacing for years to come as you suggest, but less and less for the average content viewer in my opinion. I long for the days when sports will be fully progressive at 60fps+ with less compression than the highly compressed content we mostly watch these days.

I'll have to investigate TBC and your past recommendations. It would be great to see examples of what kind of artifacts full frame TBC can clean up before investing in them. It seems I have many hours ahead of me trudging through Ebay posts and hunting down working used TBC equipment. I'm not much of a fan of Ebay and buying used equipment there, especially when I need items shipped to Canada and sellers don't properly test what they are selling.

With WinDV capture, I wasn't getting many dropped frames and they seemed mostly to occur around recording ends and starts, which are completely negligible. Maybe there is more going on than I think or WinDV is not reporting all the dropped frames? Regardless, I'll look into a non-DV capture workflow.

Thanks for the mention of someone with oxide-shedding workarounds. I'll look into it. Those tapes are terrible to transfer.

And thanks for the suggestions and help.

Do not adopt H.265 yet. If you think clients will not understand interlace, just wait until you give them new formats. And BTW, the jury is still out on whether H.265/HEVC will displace H.264. It will surely exist in professional settings, but possibly never consumer ones. After 7 years, it's still not happened. H.264 was dev'd in '03, and was in broad use by the end of the 00s. More likely whatever comes after H.265/HEVC will get broad consumer appeal. Between MPEG and H.264, we saw lots of flopped formats.

"Expensive" is relative. In 2001, my ATI AIW card was $350 + tax. And I really wanted a Matrox RT2500 or the new X100, but could not afford the $1k+ price. So the ATI AIW was a steal, almost as good ("almost" because it didn't integrate with the NLE, a feature I thought I needed/wanted, but of course did not in hindsight). The fact that you can get great cards for $150 is amazing. Depending on card and model, under $100 at times. Consider 20 years of inflation (~1.5x+), and it's double-cheap.

Progressive and interlace are different, for different uses. If anything, technically, progressive is older. And 23.976/29.97 is massively outdated. A huge problem with interlace material is that it's actually 59.94fps, alternating. (Another longer discussion on whether 29.97 vs. 59.94 is adequate, and I can argue it both ways, meaning there is no best answer to deinterlace.) It gets worse when you inject telecined film with interlaced domain graphics. So, again, different. Consumer analog is interlaced. If you deinterlace it, damage happens somewhere, somehow. You must mitigate the damage, and QTGMC is the way it's done (until something better comes along). Not NLEs, not Yadif (Handbrake, VirtualDub, etc).

I deinterlace often, for progressive viewing (streaming). That was a major part of my studio work before health forced me to move and quit. The key is, again, QTGMC, mitigate the damage of deinterlace.

I dislike eBay as well, due to the extreme amount of lying and shysters ("tested" = not actually/truly tested, the person saw any quality image come out of it; "working" = guessing the item works, because he/she saw lights flashing). That's why the forum has a marketplace subforum. ;)

I have lots of TBC and DV research, many clips. But I lack time, especially right now.

The issue is how dropped frames are reported, and what causes the report. However, with DV, drops can still happen without reporting, internal to the encoding. These are silent, but can be seen on scrubbing the captures, especially at edit time (often weeks, months, or even years after the capture was done).

A comparison is here. Notice how in the DV images, the reds lose detail compared to the uncompressed image.

I have some good captures from tapes, that show the detail loss. Still no time yet to share those.

But probably one of the most realistic samples that I've seen in recent times is at the RED site.
https://www.red.com/red-101/video-chroma-subsampling

Notice the ceiling mosaic.
Ignore the "original", and instead understand that the 4:2:2 is the original in analog terms.
Compared 4:2:2 to 4:2:0 and 4:1:1.
There is a very drastic loss in color saturation, most notable in reds.
Colors like blue further "cook" and change hues.

DV also adds mosquito noise and macroblocks.

Ignore the roof image. That is a cherry-picked image to show exceptions to the rules, as it's specifically playing against both 4:2:0 and 4:1:1 (and even 4:2:2). Interesting, academic, but not overly useful for general understanding of colorspace.

I also have graphs and analogies, when I do share.

But, for the moment, I can at least share the analogy:

Consider a 4-day eating schedule, for the rest of your life.
- 4:4:4 = the rich gets 1 meal per day
- 4:2:2 = the middle class gets 1 meal every other day
- 4:2:0 = the poor get half a meal every other day
- 4:1:1 = the poor get 1 meal every 4 days, and starve those other 3

Quality matters:
- 4:4:4 and 4:2:2 are premium food.
- 4:2:0 quality depends on who makes it, and it can vary from good to bad. (ie, formats where used, be it MPEG-2 or H.264)
- 4:1:1 DV-branded food is stale, rough and chalky, often like a dog biscuit for humans. (ie macroblocks, noise, hue changes)

Nobody converting analog tapes is rich. But you can choose to be middle-class or poor. And if poor, you can choose 4:2:0 or 4:1:1.

So, two main questions:
1. Why would you choose to be poor?
2. And if you choose to be poor, why would you choose to starve for 3 days?

But don't answer those.
Both are rhetorical questions.

Here's the only acceptable answers:

1. Because you need to
-- (A) use a format that specifically calls for that compression. DVD-Video, for example, is 4:2:0, and that is immutable. DV cameras shoot 4:1:1, and that is immutable.
-- (B) conserve bandwidth/space, and that is generally only done for broadcasting and streaming; and it's also why formats/specification (like DVD-Video, with 4:2:0) exist

2. This answer is not so kind.
-- (A) Newbies don't know what they don't know. But now you know. So quit making dumb choices.
-- (B) You're in denial, insisting 4:1:1 is fine. Stubborn. A head-up-ass DV fanboy. A sadist who enjoys starving.

And when you share DV 4:1:1 converted video with others, you foist your bad choices onto them. Sharing 4:1:1 is the most damning part of the DV format. You're not just screwing things up for yourself, but for others.

I'm just going to share my hardware experience to give you another option for your business to consider and in the way it expends your ability to capture more formats such as Betacam analog and digital formats.

Over the years I experimented with numerous capture hardware and software starting with the Edirol VMC-1 a DV box that blows off the ADVC300 socks, to the more modern junk as well as the pro gear, and it's obvious that modern OS's don't give a damn about capturing ancient media materials, So I started to look closely and study the pro gear development over the years and I learned that their stuff is always supported by any OS old or new, So for the last few years I started to gather information on how to build a hardware workflow that stands the test of time and I pretty much narrow it down to the following but alternatives are always an option, So here is my suggestions:

1- Start with good VCR's (S-VHS, S-Beta) with line TBC, DNR and good known drop out compensation circuit - See the VCR guide here.

2- Find a pro capture device that was designed for both consumer and pro analog formats with SDI out not USB (I will explain why later), with built in TBC, proc amp and a frame synchronizer, it's all done in the digitization process in one step. I use the BrightEye 75. It captures natively in 720x486 for NTSC and the benifit of those 6 extra lines is that when you trim them off you get the benefit of getting rid of most of the head switching noise at the bottom of the video frame while keeping the legal 720x480 (that wasn't the intended purpose of that resolution though). I contacted the engineer who leads the design team and he assured me that the built TBC is one of the best in the business (12bit vs 10bit for other stand alone TBC's) and I can tel from using it (He sent me an e-mail about how it works if you are interested).

3- Once you have a clean, stable digital video in AVI PAL/NTSC/SECAM 4:2:2 10/8 bit selectable and a PCM 24bit/48Khz digital audio right at the tip of an SDI connector, from there use a SDI to USB 3.0 adapter if using a laptop or a SDI PCI card for a desktop, It will work for both PC and MAC, and works with either the adapter's software or Vdub alike because all what it does is transfer the raw data to hard drive, the capturing has already been done prior to SDI. The benefit of the SDI is it has been and it is well supported by any OS.

4- Once you have the lossless files on your hard drive, it depends on what you or the customer wants to do with them, You can edit, restore, de-interlace, encode, author to disc ...etc. From my experience vdub is good for quick line cropping, QTGMC is good for de-interlacing, FFMPEG is good for encoding, It becomes routine once you start using them and you can leave desk while running these tasks, Or multi task if your computer is powerful, USB 3.0 and SSD work well in my system.

Here is another benefit of the SDI port, If you want to expend your business to pro format, most of the pro decks are equipped with SDI output, So you can just connect the SDI cable to the SDI/USB adapter and you're good to go.

Now that I'm back home I can share samples of my work flow, The samples are the final lossy product that the customer wants, Lossless looked better before de-interlacing and encoding, They could be cleaned up further if requested:

Betacam SP sample

VHS Sample

Thanks latreche. I'll definitely look into the BrightEye 75 and review your samples.

Please note, that since I've started doing video transfers I've focused on only using and acquiring the VCRs that are recommended on this site.... so I feel I have already taken an important step in improving my workflow from what I was given.

Yes you've accomplished step one but in public forums like this we always lay out all steps involved so when someone else come across the thread he will have full guidance.

Actually I did read all the way thru. I agree, intentions are important but i guess I'm getting impatient with age. It surprises me that so many people don't really look at what they're getting with poor capture the wrong gear and zero processing? Many never notice the noise and errors? Has mass media made us that blind? Some read the advice of real experts from the video business and express skepticism: why after reading about successful gear and methods used for so many years, are they skeptical?

Maybe I was lucky. I bought AIW's way back when they were the newest thing. My first DVD was a complete failure (thanks to Cyberlink), spit back out at me from my DVD player after about 10 seconds. So I hit the internet. Amazingly, the very first link I found was to a tutorial at the SignVideo website about IRE levels and proc amps. The second website I found was digitalfaq and its analog to digital capture guide. How lucky can you get? I'm fortunate not to have started thru the copycat video blogs from the wannabee crowd. The take I had on those guys was, how obviously know-nothing can you get? And copying word for word from each other, at that!

So, yes...I get into a fit of despair these days more than anger. But I understand: bad video and bad audio permeate the market and for decades the public is trained to expect bad stuff as the norm. And truth to tell, when was it ever different? I guess I keep hoping for too much.

Good question! :old:

We were lucky. (Well, I had been doing research for 5+ years, and knew what I wanted. It didn't exist until the AIW.) Lots of kiddies gamers also caused adoption, both for themselves and parents. So lots of people were spoiled with luck, not realizing how ugly the capture card world was.

Me too. My 1st lesson was about "safe area" (overscan) menu design, and thinking DVD-R was like CD-R (more bad media, costly, overburning, 4.38gb vs 4.8gb, 2x burns too fast). I still remember that very vividly, and it was 20 years ago. Fun times? :screwy:

Many blogs and sites just disappeared in the past decade. Some are missed, but most are not. The biggest sources of nonsense these days is Youtube. People that don't know jack constantly post, re-post, and regurgitate (re-re-post) worse yet. Then all the lemmings "like" it. It's like Lord of the Flies for information. Good information is buried ... or worse, disliked. :rolleyes:

That is the medium that is being consumed by MOST people. I myself, before I fell down this rabbit hole, watched a few Youtube videos about capturing video. Luckily, I found this and VH to set me on the correct path.

Maybe DigitalFAQ should start posting on Youtube and try to counter some of the garbage. My vote is for sanlyn to host it!!

People need to see the difference between a good workflow and bad one. Simply talking about it isn't enough, they need to see the best you can get from a good workflow versus the composite-to-HDMI upscaler + HDMI H.264 recorder workflow.

And I'd like to see a comparison between ATI 600/AIW + TBC1000 versus BrightEye 75!

Sorry to contribute to the tangent this thread is taking, but reading it reminds me of how thankful and fortunate I am for having found digitalfaq early in my journey. I am also amazed and perplexed at how so much disinformation and poor information becomes the mainstream "knowledge" that is copy/pasted everywhere.

I am torn at times, because I see so much obsolete or archival video in the world that could have been better preserved/rescued if only the basic knowledge out there wasn't so poor. However, on the other hand, if the majority of folks understood that the gear mentioned on this forum was the way to go, I fear it would be so much more difficult and expensive to acquire this precious gear.

I've seen the same thing on youtube. View count seems inversely proportional to the degree of wisdom/knowledge being presented. How can somebody who was a child or not even born when this technology was around get hundreds of thousands of views, while someone with decades of experience in the industry or an engineering degree get maybe a few thousand or only hundreds of views?! Makes it so difficult to find videos with reliable/useful information. I guess we are a small group, and the masses watch out of casual curiosity with neither technical aptitude nor a desire to learn. I suppose the masses also want simplicity. They tune out when told that something is complicated and requires significant learning.

To the original poster, thank you for caring about the quality of your product and looking to improve. Like so many things in the world, the quality of the work can be compromised or just plain ignored due to a desire to maximize profit. Most people probably wouldn't even know that their tapes could look good or even great if only proper care were taken. I guess when most people get mediocre to terrible results, it becomes accepted, and it's rare for someone to hear about having a great experience. I'd like to dream that if more video transfer businesses had a healthy balance of quality vs profit, it could somehow displace the bad shops or force them to improve. When someone has a positive experience, you'd like to think it would generate more referrals. Maybe there's just so many people out there with tapes in their closet, and so it's easy to get customers without having to be competitive with those that know what they are doing. So thanks for trying to change things for the better, one customer at a time. :salute:

I agree 100%. Seeing results from different workflows would be the way for all to see the best results.

Thank you Keaton! All I want is to improve my workflow/quality of transfer and educate myself.

BrightEye 75 is no different from any classic USB capture device from the era, they were made at the time where analog video was mass consumed and digital started to dominate the broadcasting business so they felt the need to migrate the contents to digital, It differs from a consumer capture device that it captures to SDI instead of USB because it was used to capture to tape based lossless SD digital formats such as digibeta, The files are exactly the ones you would get from a USB capture device in lossless mode. For a long time SDI data was only possible to be transferred to computer via SDI PCI cards (Impossible thru USB 2.0 and barely thru firewire), but now with USB 3.0 it is possible by using a SDI to USB 3.0 adapter to use such capture equipment with laptops and all in ones.

What worries me is these easycraps, composite to HDMI boxes and the likes of intensity shuttle.

Oh I forgot, the other difference is that it is built in TBC/frame sync already that saves few hundred dollars, here is a copy from the email I received from the manufacturer design engineer:

I wonder if that 24 bits of dynamic range reduces the need to check and adjust sound levels before capture?

If would be good to see how well it can compensate for stretched tapes.

I wonder if this reduces the need to check your histogram before capture? A good test would be a tape with too much dynamic range in a single scene (shooting indoors out through a window, etc.), then try to pull information from the bright and dark areas at the same time.

So there's potential to make video capture a little more idiot-proof, and if true, I think there's a lot of value in that.

The VHS footage should have been detelecined instead of deinterlaced, and why is the ABCNewsWeather footage 30p with every frame duplicated into 60p? Is it a limitation of the BrightEye, similar to the HDMI solution?

The BE 75 has an audio gain adjustment for both audio channels anyway, I will post screen shots of all its settings when I get home. I don't know how to telecine, but the footage plays fine, as to 30p and 60p I don't know what are you viewing the spec with, But I used QTGMC to deinterlace and ffmpeg to convert to mp4, The bright eye has nothing to do with de-interlacing, it just captures lossless interlaced footage.

Here is the conversion command:
And here is screen shots of the BrightEye75 control menu:

http://www.digitalfaq.com/forum/atta...1&d=1589442690

http://www.digitalfaq.com/forum/atta...1&d=1589442690

http://www.digitalfaq.com/forum/atta...1&d=1589442712

http://www.digitalfaq.com/forum/atta...1&d=1589442712

http://www.digitalfaq.com/forum/atta...1&d=1589442712

http://www.digitalfaq.com/forum/atta...1&d=1589442712

Since it outputs 486 lines instead of 480, when you capture VHS, do you get 6 lines of actual video picture at the top, or 6 lines of black, or some combination of the two?

All, neither, or a mix. It varies by tape, and by VCR.

There is always extra lines of head switch, blanking signals and other data if available as LS mentioned and they are always greater than 6 lines, So trimming those extra lines off helps getting rid of most of that junk without cutting into the active video area. Consumer capture cards however do capture at 480 so you will have to live with that noise or mask it with black bands because you can't crop it.

I've also started to crop down the horizontal resolution to 704 since it's considered legal, that also took care of the useless padded 16 pixels (8 on each side) that was added in the ITU-Rec 601 (CCIR-601 or D1) standards and was adopted by all capture cards, That leaves a clean frame of 704x480 with 95% or so of active video, then I just add the 4:3 aspect ratio flag to the encoded file and wala.

Here is a snap shot (no aspect ratio set):

http://www.digitalfaq.com/forum/atta...1&d=1589781476

It's spelled "viola". Pronounced "vwala". It's French ("voilà").

I learned that from Archie comics about 30 years ago. :laugh: I kept wondering what the characters were saying, so I looked it up. Still no idea why it was used so much in the books back then. Seriously, it seemed like at least once per digest issue. I forget which character used it most. Maybe Mr. Lodge?.

And now you know. :P

I do fully speak French and while it's spelled voila in French it's totally pronounced wala in the above purpose which means "there it is or there you go"

Well, viola! :laugh:

(It's actually a somewhat interesting linguistics discussion on a few sites, something I never knew. This was interesting: "My best guess on the v > w change is that the w in the French [vwala] weakens the v to the point where it may be more like a beta, and then the process continues to drop the v entirely." So both "vwala" and "wala" are accepted, and likely regional dialect based.)

Didn't mean to sidetrack the conversation... :o

Yes in most cases is Vwala such as: Le Voila - There he is, or: La Voila - There she is, but when you want to say like "here", or "here you go" it is just pronounced wala.

I just want to add a quick note about where the 486 lines come from for NTSC standard.

The NTSC TV standard uses 525 scan lines per frame for both the odd and even fields, Vertical retrace and sync use 20 lines for each field. So 40 lines per frame are reserved for such purpose, The remaining 485 lines per frame are used for active video signal, But since the last line is scanned from the edge of the screen and disappears in the center and resumes at the top center of the screen and finishes at the opposite edge of the screen both halves are considered as full lines by the pro capture cards with half black and half video for each line, so 486 it is.

I have a tape whose clean frame area varies between episodes on the tape, comprising between 694 and 705 clean pixels across, with between 10 and 21 pixels of useless area on the left and between 5 and 8 on the right. Maybe some of it is my VCR or capture card, but some is also the tape itself. I can't crop a fixed number of pixels on the sides to remove all of the useless picture area without also removing some of the clean picture area. Since the left and right edges aren't distracting, I just center as well as I can and don't crop the left and right sides at all.

That seems to be common with TV recordings, especially for shows with footage from mixed sources like documentaries. It would all have been hidden by overscan anyway so some footage having more black borders than other footage wouldn't matter to the end user. TBCs, video processors, cameras etc seems to have have varied a bit in where in the overscan area the transition between black and actual video data would be.

Yes, extra pixels in the border area is fine. Missing pixels in the picture area is not. The possibility of increased vertical picture area is one thing that makes the BrightEye interesting.

Vdub lets you customize what number of pixels you can crop on all 4 sides, for the vertical pixels make sure you crop in pairs to avoid reversing the field order, On all tapes I've done I haven't come across a tape with more than 704 pixels of H resolution because the standard of the first digital format (D1) was set to 704 for each analog scan line.

Overscan is only on CRT monitors and some of the old plasma's and LCD's, most of nowadays flat panels are set to just scan and the widescreen makes all that junk visible on the sides of the frame that's why I decided to go with 704x480, I played this captured clip on my OLED TV and I was impressed.

Hi Guys,

Thanks LordSmurf for the capture hardware... it came yesterday. Haven't had a chance to test it yet.

In the meantime I'm getting a weird problem with WinDV capture of some MiniDV (and some Hi8 tapes). The resultant AVI (Type I chosen to avoid audio drift I was getting with Type II) seems corrupt in some way and Premiere chokes on it (hangs permanently trying render frames from the last few seconds). When I export these AVIs as MP4s in Handbrake there is audio drift. If I trim the last 2 seconds in Premiere (without trying to view them), the resultant trimmed AVI exports properly in Handbrake. ffprobe reports interleaving problems with these trouble AVIs.

I've researched online and some people are saying that WinDV has audio output issues if you use the device control option, but I get these issues regardless if device control is on or off. I believe the problem lies in recordings that switch between LP and SP, with bit rates for audio changing between 12 and 16bit.

Does anyone know of these problems and possible workarounds?

Try ScenelyzerLive, If you still have problems then you have to capture segments with different audio parameters individually.

Capturing DV over analog is a workaround, and quality different will be minimal at worst. Sometimes DV transfer simply will not cooperate, and you'll spend a lot of time and frustration attempting to "do it the right way" (transfer). But analog capture is fine. The biggest quality loss is the DV itself, and the original consumer camcorder low-quality optics.