RetroTINK 4K test for analog capture?

[Gary34]

Quote:

drones are an emerging technology in a growing market with a very wide range of production choices being sold brand new. As such, market forces continue to drive the price/performance ratio towards lower cost/higher performance.

They won’t be if DJI gets banned from the U.S.. DJI is a Chinese company that basically competes with itself. It totally dominates the drone market. https://youtu.be/aTF1YnqNyP8?si=HfKjkC0ikH5ut_s2 They almost got immediately banned from selling in the U.S. because of the countering CCP drones act. Now it has a year to prove it isn’t a national security threat and if it can’t then it’s banned.

They have been advancing like crazy though. Your video looks a lot better in low light than video from my Mavic Pro. Your camera sensor is almost twice as big my Mavic’s and that allows for more dynamic range and a better low light performance. Your camera has more megapixels. It uses H.265. It also uses a higher bitrate which leads to less compression artifacts. Everything advances soo fast that things become outdated quickly.

I think video gear is a better buy. People can use it and resell it. Drones are fun though.

[RayNotes]

Quote:

Originally Posted by lordsmurf

There's also not really anything "emerging" about drones, GPUs, nor TBCs. It's all decade+ old stuff now.

Wow, are you still using that? I find early Nikon video is too limiting. I bought Z II several years ago, but never used it as intended, lost my muse.

Nice.

Agree, and I'm with you. (These guys don't know it yet, but they just volunteered to teach me about drones. )

I'm not saying the distinction I pointed out between Drones and TBCs encompasses all the psychology surrounding purchase choices, or the whining therein, only that I think it's the biggest one - and I'm just speculating.

I don't really use the Nikon D700 for much anymore, but in a the demo reel, I had the footage so I dropped it. The reel spans 15 years.

I agree that drones as a domain are nothing new, but the rapidly evolving performance-to-price-point ratio is continuously dropping a wider compliment of cinematic tools and nomenclature down on a growing number of consumers. Perhaps VHS/digital technology could see a renaissance resembling that should the craze to save childhood memories reach a fevered pitch. There are a lot of Gen X'ers out there old enough to have disposable income and an intense desire for nostalgia.

[Gary34]

I probably should of let LS have the last word haha.

Quote:

I agree that drones as a domain are nothing new, but the rapidly evolving performance-to-price-point ratio is continuously dropping a wider compliment of cinematic tools and nomenclature down on a growing number of consumers

We are both on the kinda cheap side of the drone hobby. DJI has a lot of incentive because of Americans wanting to buy things that they don’t necessarily need. This thread being about a video game upscaler being used for VHS instead of a new product designed for that purpose points towards there not being financial incentive for a company to make something specifically for that purpose.

Drones are also really safe to buy secondhand. They have diagnostics and logs. Way different than video gear. Someone can get gear then sell it and get a drone or whatever they want with that. It doesn’t work the other way around.

[Shakedown St.]

Quote:

Originally Posted by lordsmurf

It's always interesting to get differing opinions on resolution.

I'm not a purist on resolutions.
1080p/4K/8K, fine, whatever, overkill, unless purpose dictates (and Youtube dictates 1080+)

Smurf has spoken. I'm glad to see you are finally reccomending VHS be output @ 1080P. While definitely overkill, YouTube definitely crunches the bitrate on standard def now. How unfortunate.

I've been watching this thread on RetroTINK. I still don't believe it contains a frame TBC. Line TBCs can reduce dropped frames, but I don't see this as a replacement for the legacy TBCs. Thoughts?

[aramkolt]

I have access to a RT4K for testing, but just like the rest of my hardware for testing, I've not really nailed down what the ideal TBC test would even look like for the results to be considered valid by everyone.

I don't think it is particularly helpful to shoot down various TBC testing methods when they haven't been agreed upon what tests would be useful to look at.

As far as the type of TBC the RT4K has, it's described as a 3-frame-buffer, but I don't know what it does when it runs out of frames in that buffer which would take a fraction of a second to do if video were to drop out completely. An ideal TBC would just replay the last frame in the buffer over and over until new frames arrive, but this is an extremely uncommon feature and I don't believe the recommended TBCs do that either. Mike Chi could easily send out a firmware update that does exactly that though - he could probably code that in 30 minutes or less, could just be no one has brought it to his attention.

I've had difficulty finding tapes that do not play well in recommended/refurbished VCRs, so to me that says that these tapes are quite uncommon to begin with and therefore the the average user is unlikely to run into them if the VCR in use has a line TBC in particular. We need a test protocol that is reproducible without specific tapes, or at least a way to create fresh tapes that have these errors.

Seems that some of the most "reliable" timebase errors occur is during hand filming with camcorders if the camera was being moved around rapidly, so I almost wonder if something can be done to the VCR during recording, such as manually slowing the video drum down - (with light finger pressure from above) randomly as it records or plays back maybe?

The other thing that causes a lot of TBCs to struggle is pushing the tracking knob way off in one direction (though this may only apply to VCRs that have manual tracking knob). Digital tracking VCRs with the volume button tracking control I don't know if they'll go to the same extremes. I've seen it particularly reproducible with U-Matic and Betamax which usually do have physical tracking knobs. Not that there's wide access to those machines either, but at least it's a reproducible test that can be fed to all TBCs from a real analog source.

As far as testing for flagging/line TBC correction, some Sony VCRs like to do that reliably - My SLV-R1000 I can put a commercially recorded tape and see flagging almost every time on a basic LCD monitor and in many capture card outputs including the most basic ones like the Elgato. I'm not suggesting using the Elgato for captures, but it's a valid screening tool since it can preserve flagging if present. The flagging on the SLV-R seems much more prevalent than more ubiquitous run-of-the-mill Panasonic Omnivisions and such, or could just be my units I suppose.


Sooooo anyone want to propose a comprehensive TBC test that doesn't require specific already-existing tapes?

[Gary34]

Quote:

I have access to a RT4K for testing, but just like the rest of my hardware for testing, I've not really nailed down what the ideal TBC test would even look like for the results to be considered valid by everyone.

It’s tough to have a test that carries more weight than tests that were done over a huge sample size with a large variety of gear when it was newer. I think the Retrotink will be tripped up pretty easily. I’m just going on what I’ve read though. The Retrotink is probably a lot easier to test than one TBC vs another.

[tocayodj]

are we ever going to get the test results? Thinking of buying a retrotink4k today and wanna know before i spend $700 on the wrong piece of gear. FYI i own a TBC. would sending signal through that before the retrotink make sense and fix the issues everyone is talking about with tape transfers?

[NJRoadfan]

One may have to re-test the RetroTINK 4K due to firmware changes over time. Experimental firmware change logs are showing an added TBC option under the sync options now. See under 1.9.3.

The new CE model maybe more attractive and cost effective for these workflows, although you lose the 3D comb filter from the composite input. Heck even the 5X model might become suitable as its experimental firmware is showing sync control updates.

[briansemerick]

Quote:

Originally Posted by tocayodj

are we ever going to get the test results? Thinking of buying a retrotink4k today and wanna know before i spend $700 on the wrong piece of gear. FYI i own a TBC. would sending signal through that before the retrotink make sense and fix the issues everyone is talking about with tape transfers?

I can’t say if it was better or worse with the RT4k but I loved the results. Examples:

https://youtube.com/playlist?list=PL...uzC4jmT9ltHi3C

[latreche34]

It's hard to judge the samples in that link without comparing them to a reference capture, They could be better or worse, you can't just tell by looking at them.

[briansemerick]

Quote:

Originally Posted by latreche34

It's hard to judge the samples in that link without comparing them to a reference capture, They could be better or worse, you can't just tell by looking at them.

yup, that's what I said.

I can say they look great. These used to be coated in mold.

I can't say it's because of the RT4K or if it's worth the $700.

[aramkolt]

In the couple of reference captures I looked at, you can tell on the edges that there either isn't a line TBC in use or it's not quite as good as it could be. That doesn't really say anything about the RT4K though as I don't think it is claimed to work like a line TBC, probably just is no line TBC in the chain.

[lordsmurf]

Quote:

Originally Posted by aramkolt

In the couple of reference captures I looked at, you can tell on the edges that there either isn't a line TBC in use or it's not quite as good as it could be. That doesn't really say anything about the RT4K though as I don't think it is claimed to work like a line TBC, probably just is no line TBC in the chain.

Remember: the term "TBC" is used far to liberally/generously, and can mean almost anything. As I often quip, my toaster probably has a TBC!

This is why we must define what TBC is for our tape-to-digital purpose. Lots of "TBCs" do nothing at all, especially anything that is old, was cheap, or is an "also does" feature inside other non-VCR devices. At best, many are weak.

RetroTINK uses the nonsense term "double buffering"(or "triple buffering") -- not much different than "reverse the polarity of the neutron flow". It's technobabble BS.

The "TBC" was a "feature" that was initially claimed by the developers, but they quickly backed off when the videeotape/analog/capture communities called BS on it. Then it languished in "firmware update hell", with no clear usability.

All samples to date have been in a vacuum, using the nonsense non-sequitur "looks good to me" by users. Too many people have beer goggles for video quality.

I will easily state that you can do far worse than a RetroTINK. But you can also do better, especially for the price. For our needs, the price is ridiculous. But it wasn't made for us. It's a video gamer scaler, and does quite well at that task. For that purpose, I wouldn't balk too much. Good tool costs money (video, games, or otherwise). Crap is cheap or free.

If I ever get back into console games, I'll give it a look. I recently came across a Youtube video showing how easy it is to hack PS2 now. Some of my favorite arcade ports are on PS2, and I still have a console, so that may happen next year. Also some neat non-USA games like Smurfs and Macross!

RetroTINK has a place, but VHS conversion really isn't it. Wrong tool. Not as bad as screwdriver vs. hammer, but more like using a flathead in a Phillips or Torx. Yeah, it "works", but not great.

[RayNotes]

Well, VHS digital conversion is a visual medium. Words can say anything, like "there are square circles.", etc. Conversion results can't do that - they are what they are. If the RetroTINK truly gets its butt kicked by other methods, this could be demonstrated... visually. I.e., the proof of the pudding is in the eating. It might be worth putting this gizmo to the test. Maybe I'll do that this winter if I come across the funds acquire the gear and do a test.

[aramkolt]

I think double/triple buffering refers to 2 or 3 frames of rolling buffer which is typical of frame TBCs. I don't know that it makes any "line TBC-like" claims which would just involve making sure the left most edge of each line starts in the same place on the left, and if you're getting fancy, you'd also stretch or compress the line linearly to make sure it ends on the same place at the right of the displayed image. Since it's holding those lines in the buffer, seems at least the first one wouldn't be too hard to accomplish (since essentially you're just ignoring the horizontal blanking interval and releasing those lines at the correct rate starting at the image portion of the line).

I do have access to a RT5X and a RT4K for testing, but haven't tested them yet nor really any of the other of the variety of devices that I want to test. Designing the test that satisfies everyone is the main problem. I don't really want to test each TBC-like device ten times as it'll get pretty time consuming and keeping track of all of the different captures and finding a way to post them all. My guess is what needs to happen is some sort of preliminary testings that excludes devices from being worth further testing and then focus on the (few?) remaining ones that pass more basic tests.

So far, I think the idea is to do a passthrough of each device from a TPG21 playing the SW2 pattern via S-Video and then Composite through each device. This is a "clean" signal that doesn't need timebase correction, but if the device visibly harms a clean signal, it is probably not worth further testing. This will also show how well any decombing filters work on the composite signal.

I suppose then that the next test would be a VHS recording of the SW2 recording made from the TPG21 and played back on a S-VHS machine that doesn't have any type of internal TBC, or has it disabled. The Sony SLV-R1000 would be a good candidate for this since it doesn't have a line TBC and it likes to create plenty of flagging even on commercially produced tapes, so that'd give an opportunity to see if any of the devices correct line errors as well. I'd also see about adding a visible timecode and LTC audio which might help identify if any frame drops or Duplications are occurring and this would be inserted at the time of the recordign to the VHS tape.

I think these tests could be reproducible for others at home just by making a burnable DVD with a long loop of the SW2 pattern with a visible timecode burnt in as well as the LTC audio being present.

From there, you could capture such a DVD directly from a DVD player with whatever card is being tested vs a VHS recording of the same DVD. You'd then visually look for problems in the SW2 pattern and run a "duplicate frame scanner" on the output file and see if the LTC audio drifts from the "burnt in" (visible) timecode using an editor that can translate the LTC audio into a digital timecode. A chain that isn't dropping any frames, you should not find any duplicate frames and the LTC audio should not drift from the displayed audio. For it to be a really valid test, you'd probably need to capture at least 15 minutes though as frame drops can be infrequent.

Just ideas for now though. Wouldn't be that hard to actually do. It won't satisfy those that say you need to check devices with specific bad tapes though, but it should be a very good screening to see what devices are worth further study I would think.

[lordsmurf]

Quote:

Originally Posted by aramkolt

I think double/triple buffering refers to 2 or 3 frames of rolling buffer

Yes, it's self-descriptive.

Quote:

which is typical of frame TBCs. I don't know that it makes any "line TBC-like" claims which would just involve making sure the left most edge of each line starts in the same place on the left, and if you're getting fancy, you'd also stretch or compress the line linearly to make sure it ends on the same place at the right of the displayed image. Since it's holding those lines in the buffer, seems at least the first one wouldn't be too hard to accomplish (since essentially you're just ignoring the horizontal blanking interval and releasing those lines at the correct rate starting at the image portion of the line).

This is a stretch of the term, and takes lots of assumptions to get there. TBCs need RAM buffers, but buffers are not TBCs.

It's this logic fallacy again: John likes bananas. John likes blue. Bananas must be blue.
Or here: TBCs fix video. TBCs use buffers. Buffers fix video.
No.

Quote:

So far, I think the idea is to do a passthrough of each device from a TPG21 playing the SW2 pattern via S-Video and then Composite through each device. This is a "clean" signal that doesn't need timebase correction

I don't agree. Skewed results.

Quote:

The Sony SLV-R1000 would be a good candidate for this since it doesn't have a line TBC and it likes to create plenty of flagging even on commercially produced tapes, so that'd give an opportunity to see if any of the devices correct line errors as well.

I don't agree. You're adding a variable with the Sony deck. A proper test should not involve known-faulty gear (create errors). It's creating a fake error, not natural, due to its own failures.

Quote:

I think these tests could be reproducible for others at home just by making a burnable DVD with a long loop of the SW2 pattern with a visible timecode burnt in as well as the LTC audio being present.

Patterns have limited usefulness in the real world.

Quote:

Just ideas for now though. Wouldn't be that hard to actually do. It won't satisfy those that say you need to check devices with specific bad tapes though, but it should be a very good screening to see what devices are worth further study I would think.

If it were that easy, we would not collect libraries of bad tapes. We'd all be using test patterns.

[Tea Monster]

There are so many interpretations of what a TBC is and what is used by different people. I think the best thing is to just state what you are using and provide a before and after test and we can make up our minds from that.

[Speedy77]

Loved catching up on this thread, thanks for everyone’s contributions and advice!

As someone getting into VHS transfers over the last few months, I’ve found sites like DigitalFaq and YouTube super helpful in determining what the right workflow is for me. LordSmurf played a big role in the start of my research, so thank you for all your help there! I’ve looked at options like the basic VCR to capture card setup, which thankfully I quickly moved past, to other simple methods like the DVD/VCR combos (the Panasonic ES30V and EZ45V) into even better setups like one I’m working on now incorporating what will be a newly-recapped Panasonic AG-1980p.

I own a RetroTink 5X and will be more than likely buying the 4K in the coming months. I'm currently struggling stomaching the purchase of a frame TBC when it seems like I can't quite get a clear picture of where to safely buy, which model to buy, and from what production-time to look for. On top of that a lot of it also seems very relative and nuanced, which makes me want to go for a less variable digital option like the RetroTink. When price tags for non-confirmed working TBC 1000's are going for $650, there's not a ton of room for experimentation for me.

I've been reading all the discourse and I understand that the product isn't for everybody. However if you have anything you feel like you could contribute or criticize about the product, especially if you're interested in improving the devices features for individuals like myself who use the it, feel free to join the Discord. There's even a dedicated Video Transfer channel for those who are looking for advice and would hope to improve features.

You can check out the Discord server here: https://discord.com/invite/jE6deAhjCM

Sadly with VCRs and TBCs slowly succumbing to the test of time, and with TBCs being relatively unknown to the general public, I think the best course of action might be to replace the older technology that we have with something modern where we can (VCRs themselves could never be replaced as a way to read VHS tape, but something that processes signals like a TBC theoretically could). Sadly yes that means using a technology originally meant for retro games (though now can be used with much more) but this also means more consumers can boost the popularity of things like VHS, which could save our hobby. If anyone has any skills or know-how and wants to contribute, there's definitely some ways to do that!

All that to say, has anyone made any recent comparisons or is anyone previously a part of the thread using/not using the Retrotink in their workflows? I'm more than likely going to be picking up the 4K in the next few months to pair it with my AG-1980p, but am curious about your results.