One suggestion - as DV is an interlaced format, conversion to AVC/HEVC should perform a double-framerate deinterlace. ffmpeg supports yadif, which handles this quite well, pass it mode=send_field (for "output one frame per field) and parity=bff (as DV, whether PAL or NTSC, is always bottom field first). As all frames should usually be considered interlaced, the deint parameter (which can deinterlace only marked-as-interlaced frames) can be omitted.
I see people converting SD video all the time and dropping half the fields because "well, it says 29.97fps!". The only time this is not an issue with DV is if the tape was recorded with a camera that has either a progressive 30fps or 24fps setting enabled.
Edit: Wow, I don't know how to read. I see the author already did this. I'm just so used to people getting it wrong that I overlooked the fact that you got it right. I'll leave this here to take my karma lumps lol
Alternatively you can just encode them interlaced as that's what the camera recorded the video as - AVC supports this, don't think HEVC does. My DV archives are encoded as interlaced AVC and I let the player's hardware or software deinterlacing take care of it.
I haven’t really explored the interlaced side of AVC. While I do like to preserve a properly interlaced version when possible, I also figured deinterlacing to the field rate sidesteps any weird issues with interlaced content that various players might exhibit.
I see people converting SD video all the time and dropping half the fields because "well, it says 29.97fps!". The only time this is not an issue with DV is if the tape was recorded with a camera that has either a progressive 30fps or 24fps setting enabled.
Edit: Wow, I don't know how to read. I see the author already did this. I'm just so used to people getting it wrong that I overlooked the fact that you got it right. I'll leave this here to take my karma lumps lol