>donwloading upscaled 1080p anime Hello /a/, I wanted to

File: Pre-Upscaled-1080p.png (1MB, 1280x720px) Image search: [Google]

1MB, 1280x720px

So what is going on here? How can an upscale be better looking than the original resolution? The answer is because both images are upscaled, just that one is pre-upscaled, and the other is upscaled by your computer. There are advantages to pre-upscaling an image.

The one I am going to talk about is bitrate. A pre-upscaled, 1080p release almost certainly has a higher bitrate than the 720p release. In this case, the 1080p release has a bitrate of 2075kbps, while the 720p release is at 1296kbps. This is a difference of 60% larger on the 1080p side, which is common in, say, Horriblesubs releases.

File: Shit-Quality-720p-Upscale.png (1MB, 1280x720px) Image search: [Google]

1MB, 1280x720px

Here is the same image at 720p, upscaled in Media Player Classic, using the Enhanced Video Renderer (custom presenter) (you should not be using this...)

everybody who cares probably already knows to be quite honest family

File: Pre upscale vs original vs madVR.webm (368KB, 1280x720px) Image search: [Google]

368KB, 1280x720px

>But what about madVR! You're not using madVR, noob!
madVR cannot magically compensate for lower bitrates of re-encodes. As you can see here, madVR is not identical to the original 720p file I used to create the 720p and 1080p re-encodes.

>>135022952
>>135023143
>>135023217
>>135023321
Just want to clarify, was this recorded and aired at 720p or was it higher than 720p?

>>135023143
And you would see vastly more benefit from just encoding the 720p version at the bitrate of the 1080p version.

File: 720p madVR.png (1MB, 1280x720px) Image search: [Google]

1MB, 1280x720px

And a comparison image of 720p madVR

>>135023477
In this particular case, I downsampled the blu-ray to a 720p encode. We'll call this, say, the original airing.
I then took that encode, and did a 720p re-encode and an upscaled 1080p re-encode off of it.

Ideally, we'd like the original file, the file they give to TV stations/internet anime providers to air, but we don't always get that.

How is Diaz not here yet?

>>135023630
I'm sure he'll arrive shortly.

>>135023477
Haruhi S2 was certainly produced at higher than 720p, though I think the channel it was on probably only aired 720p. I wouldn't be surprised if the backgrounds and OP/ED are 1080p as well. The ANIMATION is not 1080p, which is why people who quote that anibin blog are fucking retards because he uses the line widths on the foreground, while ignoring everything else.

I don't agree with OP 100%, because it depends on the year and the show, but he is right in that for most modern anime 1080p is certainly worth it. The only reason I would download 720p for something made after ~2010 is if I didn't give a shit and didn't wany large filesizes.

File: Summoning Daiz.png (1009KB, 1280x720px) Image search: [Google]

1009KB, 1280x720px

>>135023630
Cause you're not using the correct ritual.

So which file has the absolute best quality? Which is always the best choice? Original encodes are the best bet 100% of the time, if you are certain to have access to them. If your 1080p is a re-encode of the 720p re-encode, then the 720p will be superior, because a re-encode is never better than the original, ever.

If you have two files both re-encoded off an original, the higher bitrate file will usually be superior.

>>135023591
This is probably true, however, show me a sub group that does this. Also, upscaling from the original is always preferred in a re-encode. Any filtering is best done on the original file, before encoding occurs.

>>135023591
Not really. It's still a matter of upscaling algorithms vs actually having the pixel data to begin with. Throwing bitrate at it won't do much.

>>135023805
>Haruhi S2
This is the Disappearance, FYI. I used the original blu-ray encode for this.

>>135023630
Perhaps he will show up and offer input.

>>135023625
Hmm, interesting.

Excuse my lack of technical knowledge, but I don't understand what the difference is between the higher bitrate in a 1080p reencode of a 720p video, and an upscale by say, MadVR. Is it that a good reencode just performs better than upscaling during playback? Basically, the part I don't understand is "why would there be a benefit in a higher bitrate in a 1080p reencode when those extra bits were created artificially and not any extra information that was in the 720p file to begin with?" Is it because quality is lost in the 720p reencode of the original 720p file?

>>135023954
>This is the Disappearance, FYI.
Shit, I didn't recognize. Well in that case the disappearance is drawn in 1080p as well. Kyoani does not take short cuts on animation in its movies.

>>135023805
>>135024035
what OP did was take the bluray release and reencode it into a 720p file, which we'll pretend is the "source file". Then he took that 720p "source" and reencoded it into both another 720p and a 1080p (pretend those are horriblesubs' 720p and 1080p releases), and we're comparing those two.

>>135024005
>"why would there be a benefit in a higher bitrate in a 1080p re-encode when those extra bits were created artificially and not any extra information that was in the 720p file to begin with?"
This is fundamentally wrong. There is more information of the original file in the upscaled version because the bitrate is higher. I can put it another way, if I upscale, then save an uncompressed copy of the upscale, that upscale contains all of the information on the original. If I knew how the upscale was done, I should be able to recreate the original, flawlessly, from an upscale.

madVR only extrapolates data, as does the upscale. The upscale, however, had the original data to work off of, while madVR is upscaling the artifacts from the 720p re-encode.

>>135024117
Right, I don't have Haruhi sitting around in the original airing so I made do with creating my own.

>>135024117
>>135024326
Wait what? Why would you not do this comparison to the actual source, the BDMV files or at least a remux?

>>135023810
>If you have two files both re-encoded off an original, the higher bitrate file will usually be superior.

There is no guarantee for this if the resolutions are drastically different. When you increase the resolution, you need more bitrate to obtain the same level of visual quality, so if bitrate isn't high enough to compensate for the increase in resolution, the result could easily end up worse.

>>135023930
His argument is that more bits are thrown at a 1080p encode and therefore it looks better.
This is no better and in fact literally worse than just having the 720p at that same higher bitrate.

>>135022952
720p always
1080p for classics

Also, on the topic of resolution in general, upscaling to 1080p will usually happen on production level already. The main mastering is done at 720p, 900p, 800p or whatever, then the final result is upscaled to 1920x1080, usually with an algorithm like bilinear or some kind of bicubic.

There are solutions like debilinear and debicubic that can be used to essentially revert that last step of the process if you can figure out (or make a very educated guess) what the "true" original resolution of the production was, and from that point whatever you do will generally get better results than whatever "official" sources do.

>>135024417
>There is no guarantee for this if the resolutions are drastically different
This is true, but HS usually releases their 1080p files at 60% greater bitrates. Also, if the upscale is done correctly (not simple bicubic, but something more robust) the higher bitrate will always be superior, as they are just doing what madVR does, only from source with higher bitrates.

>>135024436
>His argument is that more bits are thrown at a 1080p encode and therefore it looks better.
Lol no. The 1080p looks better because the Disappearance of Suzumiya Haruhi WAS MADE IN 1080p. He is right in general that its usually best to grab 1080p if the anime has been made in the last 5-6 years. Because despite what Daiz may say about this, anime is not just drawings, and backgrounds and other digital effects can be done in 1080p at virtually no cost to the studio compared to doing them in 720p.

>>135024547
> if you can figure out (or make a very educated guess) what the "true" original resolution of the production was
If it's been upscaled with something normal (from a signal processing perspective, i.e. not waifu2x) then it is totally trivial to do this but so far I haven't seen anyone else notice.

>>135022952
>donwloading 1080p anime
Stopped reading there

>caring about file size in the name of our Lord 2015.
>I'd "upscale" it to fit the 1080p screen anyway, might aswell let someone do it who presumably knows what he's doing.

That is all.

No one gives a fuck about your blog.

720p for airing shows.
1080p for BD's.

>>135024548
>not simple bicubic, but something more robust

Never going to happen with official sources. It's basically always guaranteed to be bilinear or bicubic, and CR/FUNi/etc get the result of this and only downscale from there.

>>135024584
>backgrounds and other digital effects can be done in 1080p

Sure, they could, but the digital mastering resolution isn't usually full 1080p, it's somewhere in-between like 900p or such (where do you get such odd numbers from? it basically comes down to sheet size and how stuff is setup in animation software like Retas Pro). Upscaling that to 1080p will obviously give you more detail than downscaling to 720p, though with most anime the visual difference isn't going to be that huge since the lineart in anime tends to be often quite soft and large.

Downloading from HS, 1080 has a higher bitrate so it looks a bit less shit. Most anime now is above 720 but not quite 1080, so I usually go with 1080. 35mm film transfers, 1080. 16mm, it depends.

>>135024215
The part I'm not getting here is why would a 1080p upscaled (uncompressed) reencode of a 720p source have more information than a 720p reencode of the source? The 1080p upscale would have a higher filesize than the 720p source, but that doesn't mean it has any more "information". Basically, what's the harm in a 720p reencode of a 720p source, as opposed to a 1080p reencode, in terms of information storage, is what I'm asking.

>>135024721
>Sure, they could, but the digital mastering resolution isn't usually full 1080p
It is though, unless the studio is working on toasters. The actually scanned in drawings are not done in 1080p, I will give you that. But rendering digital effects costs literally nothing. And if the backgrounds are digital they will also be in 1080p. Because the time it takes to render them in 1080p vs whatever resolution the drawings are is pennies. And since the original master is going to be for the BD anyway they might as well save themselves some work and render shit in 1080p.

File: daiz summoning ritual.webm (3MB, 1280x720px) Image search: [Google]

3MB, 1280x720px

>>135023808
You're not doing it correctly either

>>135024547
Interesting. I wonder if some blu-ray encoding groups do just this?

>>135024675
>No one gives a fuck about your blog.
>14 unique IPs
I disagree, shitposter.

>>135024721
>Never going to happen with official sources.
True. So assuming they get the file at 1080p already, then, despite being an upscale,m it will always be better than their 720p release, since that's a downscale of an upscale.

>>135024810
>why would a 1080p upscaled (uncompressed) reencode of a 720p source have more information than a 720p reencode of the source?
It won't. He doesn't know what he's talking about and his comparison is stupid. He should have done his encodes from the actual source, or at least just manually scaled the images from two different encodes.

>>135024730
>1080 has a higher bitrate so it looks a bit less shit.
1080p has over double the amount of pixels, so of course it has a higher bitrate. That doesn't mean it looks less shit. It means CR isn't COMPLETELY bitstarving their releases.

>>135024810
As Diaz has said, you could un-bicubic the 1080p file and get the original back. The 720p re-encode is different, since lossy-encoding destroys information.

Is non-lossy encoding even encoding?

>>135024828
>It is though, unless the studio is working on toasters.

I'm pretty sure that's actually the case in quite a lot of studios in Japan.

>>135024874
>Interesting. I wonder if some blu-ray encoding groups do just this?
This is why >>135023477 asked, because the argument of "do I download the 720p or 1080p off nyaa?" (bluray or not) always ends in "check what resolution the original was produced in here http://anibin.blogspot.ca/ ", but if we're dealing with a 720p source, things get interesting.

>>135024954
>He should have done his encodes from the actual source, or at least just manually scaled the images from two different encodes.
This is exactly what I did. I scaled the images from 2 different encodes. The fact they are my own encodes is better, because I was able to ensure they came from the same source.

>>135024980
You can do 1080p video editing/rendering on midrange hardware from 2010. If we're gonna get specific and talk about Kyoani then there is no way they don't have the money.

File: 1449509672066.gif (3MB, 536x520px) Image search: [Google]

3MB, 536x520px

>>135022952
Nope, 1080p looks better. Kys.

Captcha: 420

>>135025063
>The fact they are my own encodes is better
You just admitted to transcoding them from a 720p encode. At the very least you need to encode from the SOURCE, aka the .m2ts files or a reumx.

>>135024961
So what you're saying is, a 720p reencode of a 720p source stores lower fidelity information than the source, whereas a 1080p reencode uses upscaling algorithms and, if the bitrate its stored at is high enough, is able to retain all the information from the source? I understand your original argument then and would obviously agree.

My question now is why would you even do a 720p reencode of a 720p source then, as opposed to just releasing the 720p source? I guess this goes to a more fundamental question: why does reencoding of video need to be done when simply adding subs to a file?

Shut the fuck up and give me an upscaled image that you want to know the original resolution of.

>>135025214
There is no such thing as a "720p" source. OP took an already encoded file and transcoded it. If an anime has been released on BD, it has been released in 1080p. Period. That doesn't mean the actual video is 1080p quality, it means that it was upscaled to 1080p before it was even put on the disc.

File: blu-ray vs my reencode.webm (199KB, 1280x720px) Image search: [Google]

199KB, 1280x720px

>>135025137
The 720p file I created was from the .m2ts file. It is essentially identical to the original blu-ray in every respect. I created a 720p blu-ray quality encode first.

File: 3902cd8bebab214a4bd1e15107cabfdc.png (2MB, 1920x1080px) Image search: [Google]

2MB, 1920x1080px

>>135025294

>>135025361
Ok then. So after you made that 720p encode, you then proceeded to transcode it to 1080p?

>>135025370
effectively 1920x1080 due to shitty filters

>>135025214
>is able to retain all the information from the source?
Is able to retain more information, yes. I used the extreme example of an uncompressed upscale, but it holds as you add compression to it.

>My question now is why would you even do a 720p reencode of a 720p source then
The source my be in a file format you don't release, say, one that doesn't support subs.

>>135025319
>There is no such thing as a "720p" source
I was referring to airing anime, where most people seem to think that 720p is superior because it matches the "source" resolution.

File: baea4e7df5dd18b7ab7109fd18fdaa73.png (3MB, 1440x1080px) Image search: [Google]

3MB, 1440x1080px

>>135025447

>>135025445
I took the uncompressed 720p (530mbps) and created the 720p and 1080p versions off of it, yes.

>>135025073
Yeah, you're most likely right. I was looking for this and it basically says the same thing:

https://twitter.com/khwan_/status/594894419708379136
https://twitter.com/khwan_/status/594898003699204097

So basically, most of the industry uses Retas for making the digital animation cels, and it has issues with resolutions beyond a certain point, and about the most you can get out of it is ~955.5p. The composition is then done at 1080p or 720p in After Effects or such, and in the process the cels get scaled, most likely with some basic scaling algorithm like bilinear or bicubic (having personal horror stories about terrible scaling in Adobe software I guarantee it's not going to be anything great for sure).

>>135025508
>I was referring to airing anime, where most people seem to think that 720p is superior because it matches the "source" resolution.
If it's a TV source than yeah that is the source resolution. I can't imagine late night anime being aired in 1080i60

>>135024721
>Never going to happen with official sources. It's basically always guaranteed to be bilinear or bicubic, and CR/FUNi/etc get the result of this and only downscale from there.

Wouldn't it be better to give the Streaming sites separate files scaled from the master-source?

>>135024675
>720p for airing shows.
>1080p for BD's.
This.
Unless its Kyoani.

File: smoking-123rf.jpg (302KB, 2508x1672px) Image search: [Google]

302KB, 2508x1672px

>2GB for a 20 minute episode

>>135025624
>Wouldn't it be better to give the Streaming sites separate files scaled from the master-source?
Best would be to simply give them the source. Barring that, you are correct. You are always best scaling from source. Doing anything from source is best. Denoising, sharpening, blurring, etc...

>>135025596
>I can't imagine late night anime being aired in 1080i60

Basically everything involving broadcast in Japan is done via HDCAM, which basically works out to 1440x1080i60. This gets digitized at 1920x1080 and inverse telecined to 24p.

>>135025624
As per >>135025563, the scaling has most likely been done during composition in the mastering process already, so even the studio could only give out a 720p or 1080p video file at most depending on what resolution the composition was done at.

Some shows air in 1080. Try again.

File: basically.png (174KB, 674x609px) Image search: [Google]

174KB, 674x609px

>>135025701
>2GB for a 20 minute episode
Pretty much.

>>135025525
It's 720x480 but I couldn't get exact numbers for some reason even after cropping the subs out.

>>135025808
>Basically everything involving broadcast in Japan is done via HDCAM, which basically works out to 1440x1080i60. This gets digitized at 1920x1080 and inverse telecined to 24p.
Well fuck me. Then why don't we see any 1080i HDTV releases from groups? Do they just not know people who can cap at that resolution?

>>135025808
>the scaling has most likely been done during composition
You mean that they are scaling it to 1080p before they even create the master encode?

And some components might be scanned in at lower resolutions while other effects are done at greater resolutions?

In that case, it seems that 1080p should always be superior, assuming that's what the studios give out to websites.

>>135025916
>In that case, it seems that 1080p should always be superior, assuming that's what the studios give out to websites.
It is. The issue with CR/Funi isn't that they upscale what they get, though. Their issue is that they're fucking awful at encoding and it looks like absolute dogshit.

>>135025525
>Mai Hime
Excellent taste, anon.

>>135025903
Basically all fansubbers encoding from TV use Transport Streams, which are indeed anamorphic 1440x1080i60.

They don't do 1080p encodes from there because the source bitrate will rarely if ever justify that. 1080p simulcasts don't tend to fare much better. BDs are really the only reliably good source quality-wise for doing 1080p encodes.

>>135026033
which has always been the actual issue anyway

>>135025903
Because he's talking out of his ass.

I usually get 1080 because the subtitles look better

>>135026033
>The issue with CR/Funi isn't that they upscale what they get, though. Their issue is that they're fucking awful at encoding and it looks like absolute dogshit.

Pretty much this.

>>135026092
You aren't providing much credibility.

>>135026051
>They don't do 1080p encodes from there because the source bitrate will rarely if ever justify that.
How so? What are the actual bitrates of the streams themselves? Even at 10 Mbps that would be worth a 1080p encode for animation.

>>135026127
Why don't you apply for the job then?

>>135026051
Out of curiosity what source did fansub groups use for Working!! S3?

>>135026173
CR wouldn't hire Daiz after his crusade in the Reddit AMA.

>>135026173
He got offered a job from them even but didn't take it because he wouldn't get a position where he could actually change stuff.

>>135026173
I think he must've explained this shit about 5000 times by now.

File: a1b29c0dd36f5934e2cae18fc4c5d609.png (421KB, 1440x1080px) Image search: [Google]

421KB, 1440x1080px

>>135025882
If it's actually 480p then that is a fucking impressible upscaling.

>>135026152
It varies, but the thing to note is that with JP TV the video will basically always be shitty single-pass MPEG-2. There are some channels that use H.264, but those usually end up looking even worse because the broadcasters think "hey now we have H.264 so we can use a tiny-ass bitrate!" and things end up ridiculous bitstarved.

>>135026223
CR 1080p. The show wasn't particularly busy visually so CR's relatively low 1080p bitrate ended up being pretty decent for it. A lot of fansubbers tend to use simulcast encodes as sources these days because they offer a somewhat better base for further processing than many TV broadcasts.

Ideally the simulcast encodes would be good enough to be used as sources as-is, but the reality is that CR, FUNi et al can't and won't do anything to fix source issues, nor will they care about their encoding process introducing banding either.

File: 1440869070996.jpg (140KB, 1280x720px) Image search: [Google]

140KB, 1280x720px

So if I have a decent enough internet connection and enough HDD space, there is no reason for me to ever get the 720p version over the 1080p one, right?
Worst case scenario, I get the exact same thing for a somewhat larger file size, right?
There will never be the case where the 720p version upscaled by my PC/media player to fit my monitor will be superior to the 1080p "official" upscale release of the originally 720p anime?
Daiz please respond.

>>135026283
>https://www.reddit.com/r/IAmA/comments/2b26ou/im_kun_gao_the_cofounder_and_ceo_of_crunchyroll/cj138o9
thread for reference

>>135026316
It is and they did use a really high quality upscaler.

>>135026451
>So if I have a decent enough internet connection and enough HDD space, there is no reason for me to ever get the 720p version over the 1080p one, right?

~2010-2015: 1080p for BDs
~2006-2010: YMMV, 1080p might be an upscale of 720p master
~2000-2006: 720p AT MOST, the DVDs might even be better
pre-2000: 1080p BD if available.

>>135026173
I actually did send in an application a couple years back when they had an interesting position available, but I never heard of back from them. I heard afterwards that their jobs page was basically a sham at the time - if you didn't have a referral from an existing employee, they basically just tossed any application into the trash right away.

I do actually work in this field nowadays, though. I'm basically in charge of digital distribution at FAKKU - I handle the image processing and encoding for all the digital books and magazines and I have relatively free reign to ensure we do it right. I'm also in charge of the online reader for the site, because besides the actual content being in good quality you should have proper tools for consuming the content as well.

File: 2015-12-20-201532_838x732_scrot.png (87KB, 838x732px) Image search: [Google]

87KB, 838x732px

>>135026473
Holy fuck is this for real? My fucking sides.

>>135026451
>>135026576
In the case of years 2000-2010, the 720p will be better when the encode is very large (visually lossless), and relatively artifact-free. This is because your madVR upscaling will be superior to the bicubic upscaling. If the encode is tiny, though, you want more bits, as the 720p will have more compression artifacts.

Diaz since you're here. Whats the shitiest/cheapest CPU I can get for reliable anime viewing?

RETAS!PRO HD
https://www.youtube.com/watch?v=TeprFkEkXEg

I wonder how HD it really is if they have to do composition in Adobe Software. (I wonder how related the Vector Graphics implementation is to Adobe/Macromedia Flash.)

>>135025808
>the studio could only give out a 720p or 1080p video file at most depending on what resolution the composition was

I see.

So according to the Information >>135025563 it seems like RETAS is a bottleneck.

>>135026775
Yep, you missed it, the threads on /a/ back then about it where pretty good too.

>>135026819
Even an i3 can playback 10bit 1080p. Just get the cheapest CPU made in the past 2-3 years. Hell my laptop can play back 1080p 10bit with semi-placebo settings on mpv.

>>135026819
>cheapest CPU
madVR is GPU-bound, unless there's some software setting somewhere. I use it, and you cannot even buy GPUs as old as mine anymore.

>>135026857
>vectors
>48 bit color
What the fuck is the industry doing if it has these tools available?

>>135027126
Vectorizing from scanned images seems to run into problems with bigger resolutions.

Producing gaps in lines.

>a 0.5% pixel difference

oh yes thank you OP my autistic little mind has not been able to watch 720 manual upscaling because it's so obviously jarring and now you have shown me the superior format!!

>>135027789
yeah now I can safely say that without a doubt wasting twice the amount of storage for a single episode is totally worth it for that tiny barely noticeable to the naked eye quality increase

>this entire thread
>literally being this autistic

>>135027789
>>135027919

>anon seriously you downloaded the 1080 version? such a waste of space
>n-no! It's b-better quality lemme show you!

>pulls up a screen shot at 10X zoom comparing between 720 manually upscaled to 1080 pre-upscaled

>s-s-see? See the difference? It's obviously better!!!

>>135027789
>>135027919
>>135027922
>3 comments
>2 more unique IPs added
One of you is a samefag.

Also, why would you use an inferior download when it takes MORE energy to run it by using madVR to upscale it? It literally costs you more money to use 720p.

>>135027998
I dont use madvr because it's stupid

>>135027998
>being so autistic that you had to check the IP for any new people for some fucking reason, like samefagging is gonna make their side of the argument any less true

>>135027992
>such a waste of space
Do you seriously not have room for 500MB on your computer? What kind of shit computer do you have? We're mostly talking seasonal anime, here. Even do, when i download a 15GB anime of old, i watch it once and delete it. Even you should have room for 15GB...

>>135028021
>I dont use madvr because it's stupid
Then the difference between 720p and 1080p is astronomical, unless you have something comparative to madVR. If you cannot readily see it i feel very sorry for you.

everyone here including me is autistic for even bothering to join this autistic little conversation

the only winning move was not to play and all of us lost by clicking this bait-worthy thread

>>135027998
One of them could have already posted in the thread, retard.

>>135028112
I think you're mistaking me for a different poster, therefore you must also be stupid

File: TLR.webm (388KB, 1920x1080px) Image search: [Google]

388KB, 1920x1080px

>i don't use madVR because I'm stupid
Enjoy your overly blurry bicubic shit, then. The difference is pretty jarring, DESU. I don't even have to label them, the better looking on is madVR.

>>135026723
Daiz how is it going the work of killing hentai?

>>135028399
Jesus Christ I'm sick of this webm shit. Do a PROPER comparison using screenshot comparison.

>>135028399
wow how jarring

>>135028399
Holy shit, when people jerk off over madvr I had no idea there was that big an improvement

>>135028526
Why, what's wrong with webM?
>>135028554
>>135028638
Yeah I know. Like watching with/without glasses on.

As I understood you get the horriblesubs 1080p release when it's from funimation, and 720p when it's not. Is that wrong?

>>135030195
As I recall funimation has it's own issues like color correction.

It comes down to one thing. What do you have more faith in? Your madVR configuration or the shitty encodes the streaming services have available.

>>135030195
No. CR and Funi get 1080p. But compared to fansubs they look like shit.

>>135030195
Pretty much any HS release you want 1080p. Any fansub group release you want the highest they give you. If they give you 1080p, it's probably for a reason.