Canopus ADVC-110 for digitizing Hi8?
 

Hello, im planning on digitizing 30 old hi8 tapes. And i also want to record on hi8 from time to time and trasfer it to my computer because I like the look of it.

My problem is that im very confused about which device i should buy.

I already read alot about this online and some people state that you defenitely need to invest like 1k for a good setup.

I read that things like time base correction or y/c separation are important or necessary to get a good looking result.

I found an offer for the canopus advc 110 online and found out that this device is said to be good, even though it only has a lite version of the time base correction and no y/c separation. I read that it records in the dv format that compresses color but i want to record pal so it is 4:2:0 that is said to be ok.

My question is if anyone can tell me whether or not i will get ok results with this card because Im new to all this and cant spend 1k on the Equipment.

And I also want to alrady Thank you for answering💪

How about starting with a good Hi8 or D8 camcorder with S-Video out that plays back analog tapes and built in line TBC and a good USB capture device and start experimenting, DV boxes are not recommended

Welcome. :)

Correct, for any degree of quality, you need the proper tools. Good tools costs money, like any tool for a DIY task, or hobby, or profession. For analog video ingest, TBCs (both line and frame) are those tools. It's a boring item like a lawnmower, and yet essentially to successful completion of the task.

Y/C is s-video, also correct.

Whatever you read, or watched, was 100% false. The Canopus ADVC-50/55/100/112 boxes do not contain any TBC of any kind. None, zilch, nada. Does not exist. (The ADVC-300 has a weak line TBC that does a horrible job, either entirely inert, or making the video worse. It's based a very primitive Panasonic chip.)

DV is a 1990s technology, and had a minimum computer spec of Pentium II, with a Pentium III required. It's ancient technology that was obsolete by the 2000s, and should almost never be used. It destroys color information (reduces, "cooks"/alters color), and adds blocks. It's often worse than MPEG (DVDs).

It's only "ok" when you're unable to do better. That's a rare instance, often users stuck on Macs (usually old OS X), no access to Windows.

That box is often overpriced, and a waste of money, compared to better lossless cards. The choice isn't binary, between €1k and overpriced ADVC boxes. There are other better options, wiser use of funds, better quality gear.

It really is a great analog format, and I consider it the best that we had as consumers. Sadly, I didn't realize that until the 2000s, long after we'd used VHS and VHS-C for our family memories.

Yes, good camera required. :congrats:

Hello again and sorry for the late respond.

In the meantime I experimented with the "Elgato Video Capture" in Virtual Dub and got very bad results.
I have the problem that from time to time the video randomly freezes (tried almost all Virtual dub settings: Buffer, Audio/Video sync etc.).

I also read a lot about TBCs and found out, that my hi8 camera has a built-in line TBC (Sony CCD-TRV69E)

I do not know if this is because of the lack of an full frame TBC but I assume it could be the reason.

Because Im on a budget I still consider a dv box like the canopus ADVC 110, but I dont know if it would really fix the issue.

I would really like to know if there are other options avaliable in that price range that would do a sufficient job.

merry (late) christmas to everyone

Lack of frame TBC could be part of the problem but I think you should address the elephant in the room first, get a working capture device, Elgato is not a good one despite some YT reviews that suggest it is a good one.

It's both. Lack of frame TBC, lack of quality capture card.

I always hate seeing when people are still struggling to get a good capture after months, or even years. The enemy here is always yourself, unwilling to spent funds to buy the needed tools. It's like refusing to buy a stove, and trying to cook all food with a lighter. You're just going to burn yourself, the food is still frozen, and you're still hungry.

"only 30" tapes doesn't really matter. Buy it, use it, resell it. It's a project cost, not a permanent cost. Whatever is not recovered on resale is the rental/project fee.

You'll get quite a few different opinions, but if you want to just see what DV looks like and if it'll meet your needs, really all you need is a Digital8 camcorder that can play back Hi8 tapes (not all of them do) be sure it also has stereo output for sound, and an internal line TBC. That'll have firewire output by default and will internally convert to firewire at the output. Alternatively, you can use it to play back the tapes via S-Video and convert via the usual analog means and still have the line TBC active.

In your situation, I would first ask any family members if they have any older camcorders you could borrow - most MiniDV and Digital8 camcorders did have passthrough to go from analog to firewire/DV and even if it doesn't have a TBC or stereo out, you can get an idea if that's the "look" you're going for.

This video probably shows one of the better direct comparisons between DV and a traditional capture:
https://www.youtube.com/watch?v=SVXp7cjJT8Q

Their conclusion was that video levels were a bit too light and some of the light areas were probably clipped. For the added hassle of a traditional capture vs DV (no audio sync issues and you can probably get away without a separate TBC) this may or may not be enough of an issue to avoid DV.

Color accuracy is probably also lacking a bit with at least some DV converting chips as can be seen here:
https://www.youtube.com/watch?v=XPtYzD6kNNk&t=385s

Seems that at least in this example, color has some added red throughout. This is probably correctable or could at least be improved in post-capture editing, so that depends if that matters enough to you or not as to whether that is worth avoiding. I haven't seen a comparison that shows if any of the DV conversion devices do a better job than others in terms of color accuracy and avoiding clipping of bright areas. That'd be a comparison between say a Sony DV camcorder passthrough vs Canopus device vs other brands of camera passthrough.

I would argue that if you go the DV method, your captures will be better than 95% of what most people are doing with devices like the Elgato or anything that you could purchase off of Amazon right now etc. Could it be better? Sure. Pretty much any other method will be more complicated and expensive though, and that's up to you to decide what is important. DV basically takes the simplicity of Elgato and vastly improves the quality while preserving interlacing at the cost of a much larger file size and needing a firewire device for the capture part. You still will need to deinterlace, so there is some post-processing involved before you have your shareable file. Files will also be muuuuch larger in the neighborhood of 12GB per hour as opposed to Elgato's 2.5GB per hour. Traditional lossless captures are larger yet, something like 30GB or more per hour.

As for color compression - that's an interesting one. most consumer SD video never really had more than 40 effective lines of horizontal color resolution to begin with, so if you sample 1 in every 4 horizontal pixels for color and the horizontal resolution is 720, you are sampling something like 160 lines of color from something that only had 40 lines if you're using 4:1:1 subsampling. 4x oversampling is generally considered to be a good approximation of an analog signal, and DV does give you that. In PAL land however, you're just throwing away every other horizontal line of color data which is probably more detrimental for color detail than 4:1:1 is for low resolution sources. There, you'll be sampling 360 times per horizontal line (in something that only had 40 values to begin with), but capturing none of the color in half of the lines. It is true that YouTube and most modern formats do 4:2:0, but that's a bigger problem when your low resolution source is missing half of the lines of color data. If you were to say upscale the 4:1:1 content to 720p for YouTube and then it throws away every other line of color AFTER the upscale, that'd be far less noticeable. So I'd argue 4:1:1 is better than 4:2:0 for low horizontal color resolution SD, but that's just my logic, could be wrong of course.

As for whether DV has any sort of TBC effect, the test I'd propose for that is timecode generator put right before the input of the DV capture device that'll burn in a unique time code visibly to each frame and also output LTC audio which can be then captured by the DV converter. You'd know if frames were dropped due to the total length of the video file captured based on subtracting the starting and ending time code that is visibly burnt into the first and last frame and see if the concurrently captured timecode audio agrees or not. You'd also want to check the total number of frames in the file to see if it agrees with the LTC audio and burnt in time code.

What is well established is that DV capture devices never have audio sync issues which is the main reason for a frame TBC (well, that and vertical jumpiness). I do also believe there's a line-TBC-like effect to most DV hardware converters as well where they'll fix *most* horizontal wobbling assuming it is a first generation/original recording tape. If anyone has links to a DV capture that has horizontal wobbling present, I'd like to see it.

All that said, with everything I've acquired hardware wise, I wouldn't be too shocked if the difference between using an ideal setup vs DV was relatively minor in most viewing scenarios (ie on a cell phone screen), and I'll eventually be doing that comparison. New years resolution is to stop buying stuff and start actually testing it. Probably doesn't help that I decided to get into U-Matic players too here just recently haha. For an actual transfer business or my own memories that I'm rather picky about, I'd use an "ideal setup", but if recommending something to family and friends that aren't as techy and don't have the budget, I'd have little hesitation about recommending DV over whatever else they'd have likely come up with on their own that they could still buy on Amazon.

The only complaint I have about capturing analog tapes to DV is lack of control over levels, In an analog capture device you can adjust levels to minimize the damage done during shooting, but with DV the file is finalized and you would have to decode it and do a post adjustments which is kind of too late already since the wrong levels are baked in.

Agree that I don't like the idea of clipping brights, though I think the workaround is probably a simple/low cost variable resistor as a voltage divider to ground on luma only which a few other users have posted about. Blacks really won't get clipped, especially if something like the ADVC-110 is set to black=0IRE. I also suspect that there are some DV converters out there that have better AGC where they won't clip, but I haven't seen too many direct comparisons against different DV hardware converters in that aspect specifically. I know Sony made a few standalone hardware converters - I bought one for testing, but haven't tested it yet. I was kind of surprised to see that there were ZERO electrolytic capacitors inside of the Sony, whereas the ADVC family has something like 15 or more SMT electrolytic caps. I've had to recap every ADVC-110 to get them to work, but they worked fine after. I do kind of wonder if the clipping could be bad cap related maybe? Since I haven't heard of anyone else recapping theirs, could be contributing in some way maybe, or maybe that's responsible for the "blockiness" some report? Datavideo also made some DV to hard drive/flash storage devices as well, could be they don't have the clipping issue.

Interestingly, ADVC must have know that certain capacitors inside were prone to fail, because they did use some organic polymer SMTs for specific caps which you didn't see a lot of back then - those should really never be replaced since they don't tend to go bad and have an extremely low ESR to begin with. You can typically tell them apart because they'll have red, purple, or green markings on top as opposed to black, and they'll test very low ESR in circuit as well.

Thanks to all of you for answering and the detailed descriptions.
I think I will reconsider my approach. And then decide what to buy.

Does control over levels really matter with a capture card. I understand it matters with a proc amp because you can change the signal before it gets processed and becomes the picture but making adjustments on the card doesn’t seem like it would be any different than making adjustments in software after capture.

Not all DV devices have proc-amp controls, so once the DV file is created it's too late. Proc amp adjustment are done at the ADC stage, As long as the file kept lossless, doing those adjustment in post should not make a big difference, but if the files are encoded to DV or other codecs it is too late.

All capture cards DV or not make adjustments too late. They aren’t changing an analogue signal at that point. They are working in the digital realm. I don’t think there is a difference between using the adjustments in a capture card and making adjustments in post after capture.

I don't think you understand the concept of DSP (digital signal processing) in video, The worst you can do to the signal is process it in the analog domain, get an analog proc-amp and try it for yourself and see how shitty the quality is, Digital processing can achieve better results due to the nature of lossless processing, the analog signal gets sampled into digital and processed digitally with minimum loss. Not just proc amps in capture cards work this way, TBC, DNR, mixing consoles, character generators even the VCR menus all work the same way. But once that digital signal gets encoded into DV, MPEG-2 or any other codec, it has incurred a loss already, decoding it and processing it makes it even worse.

I have a TBC 3000 that I got from the original owner and he used if for a month. It was kept in his office since. It’s possibly a 5th gen. Even LS hasn’t tested the 5th gens. I drove to the guys house to pick it up. It’s never been shipped from person to person. I don’t use the proc amp on my card. I raise and lower the volume with an audio mixer then use the proc amp on my TBC to adjust levels. The proc amp on that is strong. Anyways when you make corrections on a capture card the capture card clips then you make corrections. When you make corrections on a proc amp you push it into legal range then capture. I know I’m not going to tell you anything new but I’m just saying capture cards don’t avoid clipping and a proc amp does. I don’t see the difference in using the proc amp on the card vs using software post capture.

The proc amp in your TBC and the capture card work the same way, the only difference is that in TBC you have no range reading indication, the capture card does give you the reading, Your TBC is before the capture card because it cannot be connected after the capture card, It doesn't mean it is processing in analog, it is digitizing the signal first, then processing it and then converting it back to analog.

Adjusting the proc amp of the capture card to prevent clipping during capture is still noticeable, granted I've mainly see the improvement when I capture footage from my Playstation 3 onto Hi8 tape using a Sony CCD TRV-66 camcorder, as oppose to footage shot from camcorders, probably due to exposure and the values being clipped when the tape was recorded to.

Here is an example, was playing back gameplay footage I recorded of Antonblast using my PS3 to view the video file. I intentionally increased the contrast by 75 afterwards to make the differences more noticeable.
https://imgsli.com/MzM0Mzgy

I originally captured the footage with the brightness decreased by 10 and the contrast increased by 18. That's the first image, second one is those values reversed in order to match the image on what it'd look like if it was captured without any proc amp adjustments by the card to begin with.

I only adjusted the levels using the capture card, do have a Cypress TBC that has it's own proc amp controls, but those are left at default

Yes, nothing reverses a badly shot footage to its original values, there is no question about it, But having some control over levels better than having none.

If you look on the back of a TBC 3000 there is a composite monitor plug in. When you plug that into a CRT and adjust the proc amp the image on the CRT is free of any proc amp adjustments but you can see the adjustments you made on the preview window in Virtualdub. That would lead me to think that the monitor output is after the timebase corrections and before the analog proc amp adjustments.

They don’t look like they work the same way to me when I compare them.

Your assumption is wrong.

Work the same way and produce the same results are two distinctive descriptions, It's like saying cars work the same way, so why they don't have the same horsepower.

I still think the proc amp adjustments on my TBC are being made when the signal is analogue. The monitor composite output is analogue and it doesn’t have the proc amp adjustments on it even though the output from the device does so the composite monitor output is before the proc amp. You can view simultaneously before and after proc amp adjustments that way if you were wanting to view a reference through a CRT. Since the signal goes digital in the TBC why would they put an analogue output there then turn it back to digital to process it in a proc amp?

-- merged --

Here is what I am talking about with that. You can see it has a tint adjustment I made on the proc amp in my TBC but the CRT I have connected through the monitor preview doesn’t pick up any proc amp adjustments which leads me to believe that the proc amp is not making the adjustments when the video is digital.

The monitor output A is to monitor source A input, The same for B, The TBC digitizes the signal and applies proc amp, timing ...etc and convert to analog and output via the output (non monitor output) for recording or editing which is its purpose back in the day, the receiving device should be connected to another monitor so you will see before and after adjustments, But if whatever you believe makes you happy, so be it.

I’m not trying to argue. I’m just trying to figure it out. I just don’t thing that the proc amp in my TBC and my card are doing the same thing and it seems like what the capture card can do could just as easily be done in software later. That’s all I’m saying.

-- merged --

You are probably right. You are really experienced. Okay you’re right they correct the same way as a capture card they are just before the card instead of in the card.

I've stated for years that capture card "proc amps" are not true proc amps (processor/amplifier), and operate in the post-capture digital domain. It's a bad borrowed term.

I remember running extensive tests on true proc amps (SignVideo, others), weenie proc amps (inside TBCs such as AVT-8710), and in VirtualDub/etc software connected to capture card drivers (AIW, various USB cards, etc). This was long ago, 10+ years now, so I also tested some broadcast gear I had access to at the time. There was a stark difference on how effective adjustments were.

From a very obvious mechanical POV, the board parts required (and space on a board) do not exist for analog adjustment on input on capture cards. Not even on larger ATI AIW boards, certainly not on cramped USB boards. Nor DVD recorders, nor Canopus DV boxes, etc.

What you get on capture cards is digital domain post-ingest controls, with limitations and baked-in flaws. Some cards suck more than others, namely Canopus DV boxes when it comes to the integrity of color (and that's just raw, without "correcting" it in any way).

The one exception may be luma gain, because it would be needed for AGC (which exists on all cards). That value seems to be analog adjusted on input. Sometimes user adjustable, sometimes not. VirtualDub/software "proc amp" varies highly here on the brightness control, for that reason.

For true proc amp, each input needs a dedicated controller/chip/pathway/whatever (words/jargon escape me at the moment). You can't just cram a raw signal in, and expect it to "figure it out" (automagically). It can't all be mushed together in that way.

The main reason a TBC-3000 board is so large is to accommodate proc amp.

Though very false, the common myth is that software can fix anything. Manufacturers probably knew it was bunk, but went with it anyway. If everybody wants to correct color in software, why bother increasing costs and complexity of a capture card? So just skip it, AGC the luma, and done.

This is false. There is nothing special about DV capture cards. Those can, and do, lose audio sync just as easily as anything else. I've seen this many, many times over the years (decades, actually, as Canopus are ancient technology from the 1990s).

It eludes me. Production had to be extremely low. I believe that I've now used every DataVideo TBC model (and sub-model), except for that one.

The unfortunate aspect for you is that the 3rd/4th gen "interlace error" may be approaching. All of my 3000 units are failed. Something, somewhere, still not located, is causing it. KhAoS182 claimed to have found it, but I've not had time to ask him for more details.

It always comes down to the quality of the design, Serious gear such as pro, semi pro and institutional tend to have larger size because of the bigger number of components used and the quality of the circuit design, consumer stuff is cutting corners and adopting software control over physical knobs, no monitor output and just good enough for the main job which is capturing. I was explaining the principal of DSP which all digital equipment use, pro or consumer alike, Results however, differ obviously based on the quality of the gear.

I’ll keep my fingers crossed. I hope you guys figure it out.

I have at least always used a UPS. I keep a de humidifier in there too. The UPS for sure and maybe the dehumidifier are things I wouldn’t have done without the advice on this forum. I’ve seen a bunch on eBay and they will have some random power supply they got from Amazon on them. I definitely appreciate the forum and what it does to try to prolong the life of gear. I’m thinking the number of bad units on eBay is A LOT higher than I thought when I first got into this.

Since Latrech brought up DSP I am wondering about it now. So the SignVideo proc amp actually operates in analog but an AVT 8710, TBC 3000, etc. makes its corrections digitally while the frames are being stored in a buffer? So the analog monitor output from a TBC 3000 is from before the signal gets stored in a buffer?

Yep. :congrats:

DE-humifier? Watch that. You don't want static to build up. How are you gauging that room for proper humidity levels at the computer? Don't trust the on-unit meters, trust something near the item need to have less humidity. (Even I don't do this much. Needs vary by area, and even home design.)

Fun fact: those cheap units can nuke the device. So $5 PSU from Chine + $K TBC = failed TBC! Only buy OEM and name-branded PSUs. Most people are not aware of what "name branded" means, so learn it. For example: Honor, APD, Jet/ENG, Group West.

Shockingly so. I wish eBay would report when items were returned/refunded. It's much higher than people seem to realize. In actuality, functional electronics gear is huge risk to both buy or sell there. Far too many sellers are just recyclers and resellers, and have zero clue about how the gear should function. I recently saw some VCRs and TBCs listed, and I can 100% guarantee the buyers were in for a disappointment. All they saw was a cheap bid/buy price, but I saw red flags that reveal it to not work. I can only wonder if the buyer will find his/her way here, or if it will take them more than 30 days to realize it's a lemon.

Get schematics, or trace the boards (as best as is possible). :wink2:

So what would be the ideal test for proc amp performance?

I think if the signal remains truly analog, levels should never clip on bright areas for one. Black areas should be clippable on both analog and digital proc amps since the lowest the signal can go is zero volts/zero IRE and you can't bring it back if you try to amplify a zero value later. It would make more sense for a TBC to do proc amp changes digitally, but since the values are all truncated to whatever color bit depth that the unit processes at, the question again becomes about whether they will clip incoming or outgoing values.

The main advantage to analog processing is that there is no "bit depth" (8 vs 10 vs 12 bit), so you won't get any "banding" on scenes where the scene is mostly the same color like sky/sunsets or just very dark scenes where most of the scene is roughly black. The issue being when everything I almost the same color, all of the newly create pixels will get "rounded" to the same nearest value and that's where the banding or possible blockiness comes in. The recommended TBCs are all 8 bit which for the standard capture process doesn't matter since the captures are also 8 bit, but if you were capturing in 10 bit or higher, a TBC that does 8 bit or less conversions is going to negate any advantage of doing a 10 bit capture, so that's something to consider. This is one reason why I like ProRes422 - it is 10 bit and the format can preserve illegal values at least up until a point which could be brought back down in post processing if desired without losing that "normally clipped" information.

The DPS-475 (rackmount TBC) is notorious for "Pre-clipping" any IRE values higher than 115 - so you could feed it hot luma and even if you adjust levels down later, it'll show as very flat/clipped on anything that was higher than that on the input. I'm not totally clear on whether this is just with composite or if also with S-Video though since I haven't tested it personally yet.

Other TBCs are perfectly happy accepting boradcast-illegal levels and outputting them as illegal levels as well without clipping - at least to a point. The TBC-1000 I am pretty sure will accept and output illegal levels, but that doesn't have a proc amp to really do proc amp testing.

The basic test I was going to do in my broader testing is feed a known hot signal from a pattern generator or still frame which is boosted by a known analog proc amp to say 130IRE and see what the capture card or TBC can bring the levels back down without clipping. If they can, then it's more or less acting like an analog proc amp in that way. This same test also would be a way to check if automatic gain control is particularly effective to get levels to legal values before they leave the VCR or get assigned digital values on say a capture card.

This is the area where pro gear shine, 12 to 14bit processing vs consumer 8 to10bit processing, This is also what distinguishes proc-amp processing during capture vs post, Because once the file is created it will be 10bit at most (8bit consumer) so the opportunity to process at higher sample rate is missed.

Yes a de-humidifier not a humidifier. It’s in a 10 by 10 non corner 1st floor room with basically no sunlight due to a hill blocking it. The whole room stays pretty well the same. I keep the humidity at 45. I have a AcuRite Indoor Digital Thermometer & Hygrometer. I’ll throw a thin sheet over my gear and unplug everything if I’m not using it for two weeks or if there’s a storm coming.

I saw a TBC 3000 on eBay that said power cord not included but it is tested and working. Out of curiosity I asked him how did you test it if you don’t have the power cord and this is a copy and paste of his reply “Hello, It was tested with a cord that I have. It is not the original cord and I keep it for testing other units. They sell cords that fit it for about $15 on Amazon.” He sold it to somebody.

I’m trying to watch what I say this thread since I was blatantly wrong for half of it but I’ll just say I don’t see any clipping. My card clips until I push values into legal range and I’m not having to use close to the full capabilities of my TBCs proc amp. My tapes were shot on cheap gear inside and the values are pretty off.