I'll confess I'm not seeing that at the moment, but I'm tired so maybe my eyes are fooling with me. Plus I'm using the PC connected to the TV via VGA at the moment, so maybe that's making a difference.
Other than that, I'm not seeing a loss of detail. I don't have any noise reduction or deblocking enabled in my TVs settings.
How else do you compare the resized version to the original in respect to loss of picture detail if you don't resize it to 1080p again? Of course normally I'd open a script with MPC-HC and run it fullscreen, emulating my normal playback conditions. For the sake of the exercise and convenience I resized the 720p pic back to 1080p before uploading it here.
The sounds a little like a penny dropping. That's correct. Some 1080p doesn't have anywhere near 1080p worth of picture detail. You can call it crap if you like.
How much work did you put into making the pic look blocky?
For the record, I can't exactly remember the hows and whys of those screenshots being created. They were still sitting on my hard drive, but according to the file dates they're a year and a half old. I think I saved the 1080p version, then resized that pic down to 720p and back using Irfanview (it was originally created as an example of picture detail and posted in an old thread). Irfanview would have been using Lanczos resizing and chances are I let it apply a tiny bit of sharpening after resizing back to 1080p, but I can't remember. The way they were created combined with lossy compression probably explains any slight gamma shift etc. They were never intended to show anything other than a difference in picture detail, or lack thereof, and weren't supposed to be a resizing masterpiece.
I probably took the screen-shots from the encodes, not the original. For the purpose of the exercise that was sufficient at the time. You've only got to mention re-encoding here and there's a rush to see who can state the obvious, so we all know you potentially lose a bit of quality when you re-encode. Therefore when comparing the picture detail after encoding it'd be unfair if I compared the 720p version to the source rather than the 1080p encode. I compare the resized version to the source before it's encoded.
+ Reply to Thread
Results 31 to 60 of 84
-
Last edited by hello_hello; 25th Jun 2015 at 05:30.
-
I've managed to get a few old B/W movies down to 1024x576 with almost no loss of picture detail, but mostly I use 900p and 720p.
I do tend to try to remove as much noise as I can when encoding old sources, so if you someone who likes noise or can't distinguish it from the picture detail, resizing down mightn't always work as well. -
Three months ago most people were rather disappointed by x265 and a poll showed almost nobody was using it for actual backups. Reading the x265 thread at doom9 it don't look like the situation has improved much meanwhile so no, I heavily doubt it's ready, not now and much likely not within the next 1-2 years if development pace don't improve significantly.
-
No one have mentioned this, but x265 is made with HD and esp. UHD in mind. I think I've read on Daala web site that HEVC doesn't support 4x4 macroblocks, also it was said that current MPEG4-based tech have reached a similar limit as modern CPUs, so don't expect significant increase in quality to be seen very soon (work on x264 started ~2004, but you can see decent encodes starting from ~2008--2010, although I guess that at that time main contributors of x264 were relatively young/inexperienced). Anyhow:
- x264 Hi10 produces decent encodes;
- x265 is/will be good mainly for non-SD, more so, it may never be so good so that to convert encodes produced by x264;
- even x265 won't produce good results when fed bad video (which is the majority of SD/DVD; also considering that you plan to encode it 2nd time by yourself and 3rd time overall...).
Update
Seems that time span I specified begun after PSY were introduced... Now that I tested its impact I can say w/o any doubt that that increase of quality were caused by PSY. Because of that x265 may evolve faster than expected, esp. considering that it borrowed some of the code from x264. Anywho, it's still may be less usable or on par with x264 for SD content.Last edited by 8day; 2nd Jul 2015 at 09:00. Reason: Added note about psy
-
This makes little, if any, sense. Consider: 1920x1080 has 2073600 pixels, 1280x720 has 961600 pixels, meaning that 1080p is over 100% bigger or 720p has 46% less pixels.
What you are claiming is that the resizing algorithms used to scale down and scale up (I'm assuming different algorithms were used for each process), are over 90% lossless and that regardless of algorithm used the scaling process is better than 90% reversible with no detail loss.
Part of me thinks this is impossible, the only thing that gives me pause is that VP9 has a feature called "spatial resampling" which allows the codec to compress a lower resolution version of the frame which is then upscaled by the encoder to the correct presentation resolution (explanation taken straight from Selur's Hybrid). This is used to increase quality in low bit rate situations.
Now I have to do a bunch of tests. -
This makes sense if the original 1920x1080 didn't utilize 100% of the detail that it could be capable of (not all "unique" pixels). Due to optics & sensor issues, and especially due to compression elements, this is not that uncommon. Particularly since 1080-sized elements are not as often oversampled as 720-sized elements (which oversampling helps retain that utilization of detail).
Scott -
No point in arguing sophisticles!
Some people are simply convinced that 720 is almost 1080 "so downsizing does not make much of a difference". Of course they show utter crap 1080 examples to make their point and not surprisingly those same people are likely to find SD video "good quality" and you will find they are "against" 4K.
It's like teaching a pig to sing!
-
If the discussion is about movies why is bitrate not mentioned? Is it because it has no bearing on movie quality?
-
http://m.youtube.com/watch?v=AtA-FpTZOQw
@newpball, Which people are these that believe this? Name them. Thought not.
Have you not learned anything? Just because video (and peoples' expectations of it) doesn't always fit neatly into your expectation pigeonholes doesn't give you the right to make blatantly crass and grossly wild "stereotypical" denunciations. Seems you just can't abide those exceptions & equivocating complications that exist with media rules.
***********************
@shag00, bitrate ALWAYS has bearing on (digital) movie quality. That's one reason I mentioned that bit about compression elements.
Scott -
Perhaps it has the most significant effect on quality, provided you don't use utterly crappy encoding parameters.
The reason nobody is mentioning it is because, the whole point of compression is to get better quality at the same bitrate, or same quality at lower bitrates. That's the whole idea - maximum quality per filesize. ie, maximum quality at a certain bitrate.
So on this thread, the question being asked is whether X265 can produce much better quality than X264, at the same bitrate. Or equivalently, whether X265 can give the same quality at X264 at a lower bitrate.
If you pump in enough bits/sec, any compression format can give good results. -
Not all HEVC encoders consume a lot of power and take a lot of time. Advantech's entry level HVC series encoder encodes in real time and operates under 15W
-
That sounds like the VP9 codec must contain some super quality, retain all the detail resizing method that an Avisynth resizer can't match.
How are you not able to understand that if downscaling decreased the quality, then I wouldn't do it? I've kept full 1080p resolution in the past. Sometimes I've only downscaled to 900p, but such a simple concept is lost to you amongst your meaningless generalisations. Maybe you could offer an example of a 1080p video that shows loss of detail when downscaled to 720p. Just one. I've seen examples where there is, although even then it's generally not massive. Once again I linked to a post containing a higher quality 1080p sample downscaled and upscaled to different resolutions and and even using different resizing, but you've ignored it again in preference to posting silly generalisations. It's a bit sad. Please show us a small sample of a 1080p video I can't downscale at all without obvious quality loss.
When will you make something other than oink noises?
From the Bluray video I'm playing with at the moment. Uploaded as PNG. Nothing spectacularly detailed, but certainly not "utter crap". I couldn't find a single shot I could downscale to 720p and back and see a loss of detail. The top pic is the video opened with the following script:
LoadPlugin("C:\Program Files\MeGUI\tools\lsmash\LSMASHSource.dll")
LWLibavVideoSource("D:\T1_Video.lwi")
crop(0, 102, 0, -102)
LoadPlugin("C:\Program Files\MeGUI\tools\lsmash\LSMASHSource.dll")
LWLibavVideoSource("D:\T1_Video.lwi")
crop(0, 102, 0, -102)
Spline36Resize(1280,584)
Spline36Resize(1920,876)
Last edited by hello_hello; 9th Jul 2015 at 06:10.
-
Another one for newpball to pretend isn't there. 1080p down to 540p and back this time. Note the difference in file size.
1080p = 2.0MB
540p and back = 1.3MB.
Can you see the missing 0.7MB?
LoadPlugin("C:\Program Files\MeGUI\tools\lsmash\LSMASHSource.dll")
LWLibavVideoSource("E:\1080p.mkv")
Spline36Resize(960,540)
Spline36Resize(1920,1080)
According to Irfanview I removed 168 unique colours by resizing to 540p and back. Bummer....
Last edited by hello_hello; 9th Jul 2015 at 10:24.
-
-
You have a 1080p blu-ray video and you could not find a single shot where a resize to 720p would show a loss of detail?
That means to me that your bluray was butchered already, probably sharpened. You do realize that sharpening a video reduces effective resolution? Sharpening only makes sense if the pixel density and viewing distance of your viewing environment is too large to show the certain details.
There is no point in showing 1080p pictures that are already tinkered with, that's like saying that putting ketchup on bad smelling caviar won't make it worse.
And eh, you only watch blu-ray scenes with shallow depth of field?
Perhaps I do not understand you.
Is your point perhaps that since butchering a 1080p with sharpening and other processing and re-coding to insanely low bitrates takes all the life out of a picture you might as well downsample it as well as after all that butchering it does not make any difference anyway?
Last edited by newpball; 9th Jul 2015 at 14:28.
-
I don't want to sound accusatory or insulting to the OP, but it does matter to those that download movies. I am not one of them. However, people that I know that bother (seriously it just isn't worth the hassle to me) to download blu-ray content (I am avoiding certain terms on purpose) from the web, much prefer a 1.5 GB encode over anything bigger. IOW, there is a bit of an arms race online to offer the best encode at the smallest file size. But I think I have said too much. I only offer this as context of why the question gets repeatedly asked. For me, I personally am always looking at how to maximize my bitrate, but my motivations are very different. Back to encoding my personally shot footage.
-
To me they are all hackers, cutting quality just because they don't like large files. Their idea of quality is getting away with the minimum and then they expect a pat on the back for being brilliant transcoders. Go figure!
I compare them with McDonalds managers who call victory when their monthly food cost is lower because they were able to get away with shaving off a small edge from their hamburgers or a quarter inch of their tomatoes.
-
To make 720p out of 1080p (Blu-Ray), then upscale it again to 1080p and to say that it is almost the same as original 1080p because of details are hard to notice is very weird. Perhaps Blu-Ray contents do not have real BD resolution at all.
And secondly, do you remember in the past, how DivX or Xvid encodes were compared to DVD, but all done on CRT? It was almost "a miracle", unless we test it and watch it now on some HDTV. Not saying that we are going to watch movies on some wall projection etc., but logically, those things are not the same. Business of encoding is to deceive our eyes/brain anyway, but as soon you start to blow up that camouflage, more details start to come forward ...
And also I'd like to say on the other hand, that to chase quality , to try to watch only 1080p because supposedly I enjoy that movie more makes not much sense. One can watch an old DVD, non Hollywood and having such a blast because of DVD content. To be almost a snob to refuse 720p for example etc., they are just stupid movies , anyway. One should have decent TV that upscales SD resolution, for example I use a 50" plasma HDTV etc. Another example, Netflix has Terminator 1 on menu, so you can load it and watch it (especially if you are going to see latest Terminator, it is in the same time line, some scenes they just copied). Then you watch it as it comes latest Terminator on BD or go to the movie theater. Does that mean that Terminator 1 is a garbage because it cannot compete with latest Terminator real resolution? It is just movies, content is very important ... Movies have a feel, they do not have to be realistic at all ...
The point of encoding is to not to mess it up. To give it a proper bitrate for given resolution. So for example 1080p and 2GB for movie is not smart etc. -
Apparently, that's what I said.
I really don't care what that means to you, because what it means in the real world is I can reduce it to 720p without losing detail.
I've uploaded examples from two different 1080p videos, I've asked you to supply a 1080p sample that can't be downscaled without losing detail, and what's your response? Ignore the specifics and waffle about sharpening instead. Typical.
I'm showing typical 1080p images. Not some special, 1080p worth of fine picture detail, lovingly mastered just for newpball, non-typical picture. I'm waiting for you to show me one of those, but so far all I'm seeing is talk, talk, talk.....
Perhaps you're waffling to avoid the facts. Some 1080p video can be downscaled without a noticeable loss of detail. Much of it the 1080p video you'd buy on Bluray. It's that simple, no matter how much you wander off on red herring tangents.
Not that I'm defending the whole downloading thing, but it's called bandwidth. -
A lot of them don't, no matter how tightly newpball wants to wear his blinkers.
There's definitely some 1080p that can be downscaled without losing detail, but plenty that can. Quite a lot, in fact. I've offered several examples here. I don't know how many screenshots I'll have to post before newpball can no longer deny it, but that's just the way it is. -
-
First of all I must question if those originals are actually uncooked frames from a blu-ray or frames that are "improved"
In a prior topic you showed a picture that supposedly came from a blu-ray however in my opinion it was clearly processed.
So you will surely understand I rather be on-guard for being potentially deceived when presented by supposedly uncooked 1080p images.
Second, the fact that you are using only shallow DOF pictures indicates to me you possibly only select images that help you show your point.
Third, no single picture that you showed to demonstrate your point had any film grain, also that indicates to me the originals are perhaps not as original as we are led to believe.
Here is a magnification of the picture you claimed came straight from a blu-ray:
Last edited by newpball; 9th Jul 2015 at 20:03.
-
Of course you do. That'd be because it's much easier to question my samples than to produce ones proving your own point as requested.
Which prior topic?
Well I guess you can understand why I'm on-guard for being potentially deceived when you offer no proof of what you're saying at all.
Show me ones that prove I'm wrong.
The digital revolution arrived without you noticing?
Yeah, you posted that before and I wrote an essay explaining the picture did not come directly from Bluray and where it did come from (post #31). I never claimed it came from Bluray so please don't put words in my mouth and then accuse me of lying. It was only intended as a simple exercise to show that 1080p video doesn't necessarily have 1080p worth of picture detail and it depends on the source. A concept you seem to find particularly hard to grasp.
Of the examples I posted earlier, the pics in post #42 are from the "making of" documentary, directly from the "Joe ****er, Fire It Up" Bluray. Untouched by me. The following posts contain screenshots taken directly from a 1080p itunes video as that's what I had handy.
You go on about resizing 1080p as though you'll always lose detail and I've proved several times that's not always the case. I've used Bluray video as an example and I've used iTunes video as an example. It doesn't matter where the samples come from, and it doesn't matter how many excuses you decide to offer, some 1080p video doesn't contain anywhere near 1080p worth of picture detail. At least not detail you can see. That's not my fault, it's just the way it is. Pick a Bluray video and show me an example of where I'm wrong. You should be able to. According to you every 1080p video will show quality loss when downscaled, and I've acknowledged you can't downscale every 1080p video without losing detail, but you can downscale much of it.Last edited by hello_hello; 9th Jul 2015 at 21:05.
-
"Making of", "Add-ons", "Extras" etc are often of far worse quality (MPEG-2 based videos made by die-hard stuffy old interlace lovers?) than the main video. Just saying!
Just look at that picture you provided with that great Canon $200,000 lens, just about everything is out of focus!
Seems to me that camera operator needs some lessons on the hyperfocal distance.
Wouldn't you agree?
If they used that for the main show and it did not capture anything close to 1080p resolution the whole production team needs to be fired.
Last edited by newpball; 9th Jul 2015 at 21:11.
-
What, so you're saying not all 1080p Bluray video is the same quality with the same amount of picture detail????
What, so you're saying not all 1080p Bluray video is the same quality with the same amount of picture detail????
These are screenshots of "moving pictures", not photos.
What, so you're saying not all 1080p Bluray video is the same quality with the same amount of picture detail????
Your excuses are getting old.
Okay..... you got me. I looked through the whole disc and found lots of shots where downscaling showed a loss of detail, but I picked a couple of shots from the "making of" that didn't and then I lied about being able to downscale the whole thing without quality loss, or whatever you think I did.
FYI I picked those particular shots because they showed more fine detail than any of the concert footage. While the concert footage looks good it's much harder to find a still close-up where you can see fine detail.
Are you going to provide any examples yourself or are you content with talk, talk, talk, excuse, excuse, talk, talk, excuse, talk, talk.....Last edited by hello_hello; 9th Jul 2015 at 21:53.
-
I find it interesting that everyone is ignoring the elephant in the room.
Five years from now the same discussions will be about how poor the quality of Blu-Ray is compared to whatever is available then. VCD vs SVCD vs DVD vs DIVX vs X264 vx Blu-Ray vs ??? There's always something better, "a few years from now" that will "revolutionize video". *YAWN* I'll take and save (as is) what I have today (DVD, Blu-Ray, MKV) knowing that when TRUE 4K, HDR and whatever display technology comes next, I haven't compromised my "source". -
That was mentioned in post #9
I'm still trying to figure out what this bit was about
Originally Posted by newpball
Is it just that anything he doesn't understand (ie do personally) MUST by definition be silly? -
-
They were butchered long before I got to them. I'm just trying not to make them any worse than necessary.
The alternative would be 3 to 5 stages of lossy intermediates.
Can you do me a favour and actually give me a link to a thread where you contributed something useful?
Similar Threads
-
H.264 vs x264 and H.265 vs x265
By LeoKac in forum Video ConversionReplies: 27Last Post: 18th Jan 2017, 07:59 -
Which is the optimal motion search range in x264 and x265?
By Stears555 in forum Video ConversionReplies: 52Last Post: 6th Feb 2016, 10:48 -
x265 vs x264
By deadrats in forum Video ConversionReplies: 71Last Post: 10th Jan 2016, 06:14 -
X264 vs x265 at sane values
By Valnar in forum Video ConversionReplies: 12Last Post: 15th Oct 2014, 11:56 -
x264 vs x265??
By xenotox in forum Video ConversionReplies: 167Last Post: 12th Jul 2014, 11:16