I am digitizing analog PAL video from VHS, VHS-C, Hi8 and Mini DV cassettes. I recently learned that analog to digital conversion is done by sampling and quantizing the continuous analog signal into discrete snapshots and values. See for instance: https://www.monolithicpower.com/en/learning/mpscholar/analog-to-digital-converters/introduction-to-adcs/fundamental-concepts

In light of this, I decided to open up my video ADC in order to get the name of the IC that does the job, because there was just no way that I'd be able to read the IC's resolution off of Linux, considering that the signal had gone through a capture card and ffmpeg's filters after the conversion.

I learned that the IC is an MS1858E by Macro Silicon, which quantizes with an 8-bit resolution (not to be confused with image size resolution!) and samples at 54 MHz sampling rate, although my Chinese is a bit rusty, so take this with a grain of salt... Here: https://www.lcsc.com/datasheet/C42374496.pdf

Question: the 8-bit quantizing resolution means that my little ADC can output at most 256 (2^8) video values. That probably includes not only the color space, but all other information too... If I just let this all go and ask myself whether I'm satisfied by the end results of my digitizing, then yes, I am satisfied. But it also makes me wonder whether paying $200 for a professional ADC is worth it (as opposed to $20 for my current one) if it comes with greater resolution? Like 16-bit? I can't even know what I get before buying and opening up the product, because the retailers don't exactly write that in the product description...

Edit: I’m asking whether investing in an ADC that can work with more bits would make a significant improvement in subjective video quality, considering that the analog signal is at most early 2000’s PAL.

I'm not sure exactly what you're asking, but Quantizing just rounds off the floating point value, essentially, so in this case it's working as a sort of passthrough buffer for the analog signal. Having more bandwidth isn't going to improve the quality of your transfers, if that's what you're asking.

In this case, an IC that can "round off" the continuous ("floating point") analog signal with a higher resolution would mean that instead of having to choose between, say, light green or dark green, it could also choose very light green. Which is to say, I'm asking whether investing in an ADC that can work with more bits would make a significant improvement in subjective video quality, considering that the analog signal is at most early 2000's PAL.

Maybe there's a way to build a filter in ffmpeg to simulate it and see what you think?

Now we're talking! Smort 🤯 I'll do some more reading and see what parameters to fiddle with! 🔧🔩

If you're willing to tap into the RF input of your various tape decks, the best way to digitize any analog video for long term preservation would be to use something like a modded CX card (or cards, depending on how many signals you need to capture at once), detailed in this wiki: https://github.com/oyvindln/vhs-decode/wiki

Software defined decoding from direct RF captures can result in quite a noticable improvement in quality compared to other solutions, and keeping the direct RF captures would be what you would want to archive, so that if/when future decoders get released, you can just feed the captures into the new decoder to potentially get an even higher quality output.

If you don't want to go through all that, though, then you need to consider what you think would be "good enough" in your eyes for what you are attempting to archive.

This is amazing! Thank you!