What about BrightEye TBCs?

[lordsmurf]

That's a terrible sample clip. Too much going on, nothing but fleeting 1-2 second clips. Your eyes don't have the ability to examine anything, so you miss everything, and falsely think "the quality is fine". Sure, you can look at stills in the editor, but this isn't a photograph, it's video (moving photographs).

It's also a retail "big name movie" release, so you mostly have duplicate frames, not true interlaced content. But the deinterlacer still found a way to screw it up, as the dupe frame dance/vibrate.

Overall, the video has butchered deinterlacing. It's a mess of vibration, comb artifacts, and mush. Some scenes have this weird error where the deinterlace causes part of the image to movie, and not the entire image. The sources wasn't that unstable.

OBS is such an awful tool for use with videotapes.

I don't have this movie to do a direct comparison.

I also lack time, in midst of other capturing tasks that take priority.

[Aya_Rei]

I can agree, does seem like OBS captures tend to look like they have compression artifacts compared to properly restored video, even if the video source isn't as crisp as others. That and I can just tell the video has duplicated frames, noticing that since it's from a movie, the overall video isn't as smooth as if the was footage was recorded in 29.97 fields per second and then de-interlaced properly to 59.94, such as footage recorded on standard camcorders or TV recordings depending on the content shown.

As an example, here is a clip from a VHS tape containing footage from some awards ceremony in 1992 hosted by a bank company that one of my family members use to work for, deinterlaced using QTGMC as well as some restoration work done in Hybrid.

[lordsmurf]

Quote:

Originally Posted by Aya_Rei

I can agree, does seem like OBS captures tend to look like they have compression artifacts

And in this case, it's dual damage. OBS = bad, Elgato = bad.

Quote:

That and I can just tell the video has duplicated frames, noticing that since it's from a movie, the overall video isn't as smooth as if the was footage was recorded in 29.97 fields per second and then de-interlaced properly to 59.94, such as footage recorded on standard camcorders or TV recordings depending on the content shown.

In many cases, actually IVTC (inverse telecine) to 23.976 fps, from the actual 24 fps film sources. You only deinterlace if issues, mixed source output (ie, broadcasted documentaries sourced from VHS), etc.

Quote:

As an example, here is a clip from a VHS tape containing footage from some awards ceremony in 1992 hosted by a bank company that one of my family members use to work for, deinterlaced using QTGMC as well as some restoration work done in Hybrid.

Very nice!

In that exact clip, we can nitpick softness, upscale quality. Lots of factors are play here, discussion for another time. But getting to "nitpick" should be a basic goal. Get a baseline quality, and then feel free to quibble over minutia. Congrats, you are there. And getting there was simply a matter of (a) good gear, (b) good software. The real fun, the real rabbit hole, is all the advanced post-processing to restore/upscale/etc.

[Aya_Rei]

Thanks for the compliment. Yeah I know I could redo the restoration part. But I'd say that my goal of getting the capture part looking all good was achieved, if I need to redo the restoration by tweaking some settings in Hybrid, then all I need to do is just reuse the original capture file instead of needing to replay the tape from the beginning, redoing the capture part, which shouldn't be warranted imo.

Of course when it comes to the softness of the image, part of it can be the quality of the recording itself. The film camera used, and the fact that since this was an edited together awards show, there is a possibility that this was edited using a computer and then recorded onto tape, or perhaps contact duplicated.

I've digitzed worse quality recordings (just judging the quality of the image, not the stability), such as amateur home movie footage recorded onto VHS-C, yet I've also digitized other home movie VHS-C tapes from a different source that look much better, so the camera itself can certainly be a factor. A cheap camera won't look as sharp and nice as a more costly camera, even if they both film on the same tape format.

[lordsmurf]

Softness is a combination of source/recording, playback, capture. Lots of Q&A to isolate every softness hit to quality. And then more discussion on detail revival via restoration. That's the fun (or "fun") part, where actual video art shines. You are there! Everything becomes nuance.

It can easily get worse --- which is what was posted earlier.

Getting to quality is often simple fixes, get better hardware and software. People usually just make it hard on themselves, unwilling to spend funds on the tools needed.

[916Area52]

Good morning,

Thank you for the input.

Guess I don’t have an eye for it. The clip that Aya_Rei provided and the clip that is attached looks no better or worse to me. Just saying that I can’t justify/rationalize spending any more of my already, mostly exhausted, “beer budget”. At this place and time, I can only be satisfied and work with what I have available to me. As time permits, I will continue to dabble with this hobby and seek out advice and constructive criticism.

Thanks in advance.

Please see attached

[Aya_Rei]

No offense but I can tell that it looks worse and more blurry/compressed, even compared to your OBS capture, part of the reason why is the low bitrate of the mp4 file, around 2300 kb/s. Guess this was done with Elgato's capture software? Sure seems like it, especially since it's only 29.97 FPS. From what I remember, Elgato's deinterlace method is quite frankly, very bad.

[lordsmurf]

Quote:

Originally Posted by 916Area52

. The clip that Aya_Rei provided and the clip that is attached looks no better or worse to me.

It's vastly different, and those differences will be huge when it's viewed large (not in a teeny tiny computer preview window). If we had the same tape source, it'd be easy to show. This isn't some small difference, but an elephant stumbling drunk, essentially impossible to not see.

Rather than give up, I suggest agreeing to a uniform source tape. What movies do you have?

I find it daft how this thread when from semi-overkill BE SDI boxes, to essentially Elgato/OBS crap method. What happened between now and then? It's only been a few days?

[916Area52]

Good morning/afternoon Aya_Rei,

No offence taken…

Quote:

“Aya_Rei
No offense but I can tell that it looks worse and more blurry/compressed, even compared to your OBS capture, part of the reason why is the low bitrate of the mp4 file, around 2300 kb/s. Guess this was done with Elgato's capture software? Sure seems like it, especially since it's only 29.97 FPS. From what I remember, Elgato's deinterlace method is quite frankly, very bad.”

It is significantly more compressed. The full OBS 007.mp4 file is over 7GB, whereas the recent uploaded clip full version Elagato file (E-007.mp4) is a mere 1.84GB. The OBS 007.mp4 file was created using this YT link https://www.youtube.com/watch?v=XzQt2pOgRE0 .

Great feedback here,
Thank you for the input.

-- merged --

Quote:

“Lordsmurf
Rather than give up, I suggest agreeing to a uniform source tape. What movies do you have?”

That would be awesome…
Please see attached.
VHS-1 = Files with 02 in front would require less opening of boxes. If we can’t/don’t have one of these same movies, I attached my complete list.
VHS-2 = Complete list, I’ll be opening boxes for awhile

Quote:

“Lordsmurf
I find it daft how this thread when from semi-overkill BE SDI boxes, to essentially Elgato/OBS crap method. What happened between now and then? It's only been a few days?”

Crazy, I know… I do value the feedback.

The few days has given me pause to reexamine my budget.

Thanks in advance.

[lordsmurf]

Ah, you have some good tastes!

I see some on the VHS-2 list. Let me see where I am this weekend, perhaps I can squeeze in a few test captures.

Young Frankenstein is actually one of my fodder tapes for gear testing. It has specific calibration uses. We may, or may not, use it as a test.

I want to squeeze in Back to the Future, as I never tire of seeing it. But I'll need to find key scenes. A lot of movies open molasses slow, so it's not always easy to say "capture the first 5 minutes".

[916Area52]

"lordsmurf
Ah, you have some good tastes!

Thank you.

I see some on the VHS-2 list. Let me see where I am this weekend, perhaps I can squeeze in a few test captures.

Young Frankenstein is actually one of my fodder tapes for gear testing. It has specific calibration uses. We may, or may not, use it as a test.

I want to squeeze in Back to the Future, as I never tire of seeing it. But I'll need to find key scenes. A lot of movies open molasses slow, so it's not always easy to say "capture the first 5 minutes"."

Thank you for taking the time for this...
Attached is a five minute Back To The Future Clip. I'll do another post for a Young Frankenstein clip. Also let me know if I need the VHS tape for this.

Thanks again,
Roy

-- merged --

Young Frankenstein clip

[lordsmurf]

YF is useless here, so I removed it. All credits. The post-credits to train scene may be better, not sure. May need another section.

BTTS is a good comp clip, it looks bad, lots of blended deinterlace.

Your earlier clip was 1440x720, but not these? Are these using same method as above, Elgato/etc? Let's make sure everything is set before I start comps.

BTTF is a favorite movie, so I have no problems taking a 5-minute break to watch it, and run a test capture.
Processing will take longer, "get in line".

[Aya_Rei]

Another thing I noticed with both clips is the excessive amount of rainbow artifacts, especially noticeable around the text, granted it was more easy to spot on the now since removed opening credits of Young Frankenstein.

I might be wrong but the mp4s should be from the Elgato, I've used it before so I'm aware of the file thumbnails the card's included software generates. The mp4 being in 640x480, deinterlaced to 29.97 and have a low bitrate of around 2200 kbps looks to be in line with using the Elgato's quite frankly extremely crappy software with it's default settings.

So yes LS, the BTTF and YF mp4 clips were made using Elgato's own software unless the OP states otherwise. Using it before gave me an idea on what clues to look out for

[916Area52]

Quote:

Originally Posted by lordsmurf

YF is useless here, so I removed it. All credits. The post-credits to train scene may be better, not sure. May need another section.

BTTS is a good comp clip, it looks bad, lots of blended deinterlace.

Your earlier clip was 1440x720, but not these? Are these using same method as above, Elgato/etc? Let's make sure everything is set before I start comps.

BTTF is a favorite movie, so I have no problems taking a 5-minute break to watch it, and run a test capture.
Processing will take longer, "get in line".

Take your time, I very much apricate

Both YF and BTTF were Elgato, and my very early attempts with this endeavor.

I've attached a 007 clip that I recenly did (30min ago), using AmaRecTV and HuffYUV. If you have any time could you give an opinion?

Quote:

Originally Posted by Aya_Rei

Another thing I noticed with both clips is the excessive amount of rainbow artifacts, especially noticeable around the text, granted it was more easy to spot on the now since removed opening credits of Young Frankenstein.

I might be wrong but the mp4s should be from the Elgato, I've used it before so I'm aware of the file thumbnails the card's included software generates. The mp4 being in 640x480, deinterlaced to 29.97 and have a low bitrate of around 2200 kbps looks to be in line with using the Elgato's quite frankly extremely crappy software with it's default settings.

So yes LS, the BTTF and YF mp4 clips were made using Elgato's own software unless the OP states otherwise. Using it before gave me an idea on what clues to look out for

Thank you for your feed back. You are correct about these are Elgato mp4s.

Could you also comment on the clip that was attached to my reply to LS?

-- merged --

My mistake... It is AmaRecTV with Lagarith (@720x480 fps=59.94) and not HuffYUV.

Converted from AVI to mp4 by VLC

[Aya_Rei]

Well it's 720x480 (3:2 AR) and not flagged with a 4:3 aspect ratio for starters, so it appears stretched than how it is intended. That and the mp4 is extremely compressed with only 1000kbps. It's starving for bitrate and it very shows, it has massive compression artifacts, that and by freeze framing, was able to still see interlaced lines. So it looks like it was incorrectly deinterlaced. Though it should've probably not been deinterlaced to begin with since the source is from a 24 FPS film.

[916Area52]

Quote:

Originally Posted by Aya_Rei

Well it's 720x480 (3:2 AR) and not flagged with a 4:3 aspect ratio for starters, so it appears stretched than how it is intended. That and the mp4 is extremely compressed with only 1000kbps. It's starving for bitrate and it very shows, it has massive compression artifacts, that and by freeze framing, was able to still see interlaced lines. So it looks like it was incorrectly deinterlaced. Though it should've probably not been deinterlaced to begin with since the source is from a 24 FPS film.

Appreciate,

Post #36 is what I did. Was hoping for a better outcome than OBS or Elgato.

Plan on redoing all of my VHS collection.

Definitely need to get better educated with the proper settings for AmaRecTV.