Video quality feedback - let's hear your thoughts!

Status
Not open for further replies.

Hanners

Anime Limited Representative
Industry Representative
Hi folks!

I know some of you are aware that for some of our recent projects (Castlevania and Silver Spoon being two recent examples) we've employed a different UK-based authoring house to work on them for us. While I know people overall are happy with the video quality of the releases, people have flagged that the gamma/brightness if higher than they'd like, and so I've been working closely with said authoring house to try and improve the quality of future encodes to find a baseline configuration that people will be satisfied with.

And so, after a lot of hard work, here I am with the results! The below is a simple, single-frame comparison from Castlevania Season 1, where I present you with five different encode settings, and I'd be interesting in hearing your thoughts on them. So, take a look at the comparison link (you can click through the various images), and rate your preferences from best to worst.

I will say right now, one of these images is from the original ProRes master as my "control" sample, but I'm not telling you which for obvious reasons.

Have at it, and have your say!

Castlevania Test Encodes | Slowpoke Pics

If you'd like some more comparisons then let me know, and indeed if you've watched Castlevania and have a particular timestamp you'd like to see compared I'd be open to that too.

Looking forward to more incessant arguments about picture quality hearing your thoughts! ;)
 
I've been occupied with the Taiko Vita game throughout the day but from a quick glance at the screenshots #3 and #5 are definitely the weakest of the bunch, so it's basically between #1, #2 and #4 at the moment.
 
I'm no expert, so I'm just judging how things look in general, so I'd say:

4 > 2 > 1 > 5 > 3

The positioning of 1 is debatable, because more is visible (like the beams in the ceiling), but might it be too bright? I'd say anything but 5 or 3 would be acceptable though.
 
1 has the most detail
2/4 seem to be decent encodes of 1
3 has crushed blacks and too much detail loss
5 is too bright which introduces banding

For the next comparison please use the same frame.
 
Last edited:
For me, it's a toss-up between image 1 and 4. 2 is decent. 5 is too bright and 3 is like a TV in "torch mode".
 
As a proviso I'm only looking at these on my phone rather than my TV. Also I'm no expert at these visual pick ups that people seem to spot so easily.

3 is horrible to look at quite frankly.
I like that 5 allows more background detail to be seen but the characters look a little washed out.

I'm struggling to really see a difference between 1 and 2. And along with 4 I like them all similarly.

For what it's worth I've only watched the blu ray release of the show and thought it looked great.
 
I prefer 2 and 4, probably 4 more than 2. 1 is Also okay. It just seems more natural to me and I know I could watch either of those and be happy with it.
 
3 and 5 are unacceptable. 4 seems to be the best to me, with the proviso that these are static screencaps viewed on my laptop screen and don't really tell me how it will behave in motion.

I would suggest that you consider the services of David Mackenzie / Fidelity in Motion for some future projects, as he is considered the best in the world at what he does and has been used by Arrow, Masters of Cinema, and the BFI amongst others. It's safe in his hands and you wouldn't require one of these threads ever again!
 
Perhaps it's my setup, but I can see no discernible difference between 1, 2 and 4. I tried altering my brightness and picture mode, even going so far as to download them all and sit them side by side but still can't tell them apart. It's confounding me a little since I know I have particularly good eyesight and score perfectly on colour blindness tests. What can people see that I can't?

3 however is definitely awful and 5 looks too bright and washed out.

Thanks for asking for input though Hanners, it's always nice when companies ask for community feedback - Things like this can serve to help everybody out in the long run.
 
Last edited:
3 looks really bad, as if the contrast bar was ramped up.
5 seems too bright, but seems to retain details as opposed to 3.
1 has this strange white border. Is that some boxing or something? ?_? But other than that my eyes on my 31" HD Monitor don't see a diffrence.
 
Jesus... just tell the encoder to keep the original and intended level color (Gamma/brightness) as the master.
No need to choose between others options, if the staff behind the serie wanted the image to look like what is on their master, why the hell would anyone want to change that just for aesethic?
 
Also, any encoding profile that may looks good for a series may result in a complete disaster for another. Just try to match your source material…
 
Also why is this only on this forum. Shouldn’t it be wider. But yes it should be the best that matches the creators intent and one size will not fit all.
 
Also why is this only on this forum. Shouldn’t it be wider. But yes it should be the best that matches the creators intent and one size will not fit all.
I've forwarded it to the Blu-ray forum, and it's got a bit of a reaction, most notably from Dynit's Mp3dom.
 
And entirely correct, of course. "Baseline configuration" also concerns me, like it is to decide on a setting to apply across-the-board for all future releases rather than treat each on a case-by-case basis as they should be.
 
Why ar enough even asking? This is a f**king stupid question. Just use the Japanese BD (if there is one) or your main source as a reference and match it in terms of brightness levels. If you do the gamma wrong, you raise the chances for artifacts.

Here’s my suggestion- stop using your sh*t authoring house in France (Com'On screen) and use someone go actually knows what they’re doing.

I have 3 suggestions, and I’m sure anyone who knows me knows who they are:
Justin Sevakis (MediaOCD)
mp3dom (Dynit)
David MacKenzie (Fidelity in Motion)
 
Status
Not open for further replies.
Back
Top