How to Get the Best Quality When Rendering a Video

[kristian221]

Hey there does anyone have experience with Video Rendering is Sony Vegas or video rendering in general? I am using Sony Vegas and I am a bit of a videophile (is that even a thing?). I notice EVERY SINGLE JAGGED PIXEL and it bugs me to no end >.< I want to render a 1080p video it the highest quality possible. The footage I am using is video game footage so there is lots of fine details and motion. Now I am EXTREMELY picky with my video quality. So when I ask if it improves quality, I am talking on a pixel by pixel basis. If it fixes one pixel and adds 1GB onto the file size I still would like to know about it. I am not concerned about file size at all as I have several very large hard drives and don't mind waiting for the video to process. But I still need to edit the footage so it has to be rendered, I can't just use the raw footage from fraps. If there is anything I forgot to say let me know!

1) [SOLVED] Variable bit rate is best yes? I have messed around with it and found the highest I can possible crank it because anything after that seems to lag the video. The bit rate I have set is
Maximum(bps): 240,000,000
Average(bps): 50,000,000

A: The above bitrate is the max and is even recommended as the highest quality by Youtube itself.

2) [SOLVED] Allow the source to adjust the framerate. Is this something I would want? Wouldn't I want it to stay at a consistent 60fps? Or is there something I am missing?

A: Youtube recommends that the source should be the same as the encoded framerate. So yes allow the source to adjust the framerate. It would rather it be at a maximum of 30 however.

3) [SOLVED] The field order. There is upper field first, lower field first, and None (progressive scan). I have it set to None (progressive scan), I read somewhere that results in better quality, but I would love to know why and what it does.

A: Youtube doesn't even support interlacing, it convertes it to progressive when you upload it. So encode it with a progressive scan.

4) [SOLVED] The pixel aspect ration. I think I got this one figured out, 1 is highest quality yes?

A :This is correct, you pretty much always want a 1:1 pixel aspect ration, unless you wanna stretch the video for some reason.

5) [SOLVED] Referance frames. What are these? What would give me the best quality? There is an option next to it that asks if I want to use a deblocking filter, what is this and what would give me the best quality with it?

A: The more the better, it improves quality in moving scenes. Impact to encoding time is minimal, as well as file size increase, so go ahead and crank it up. The max number supported by most codecs is 16.

6) Enabling Progressive Download. I read this has something to do with making it easier to load when it is uploaded, what is it exactly and does it result in a loss of quality?

7) [SOLVED] Probably my biggest question is the Two-Pass option. I understand what it does, it runs through the video once to detirmine the exact needed bit rate and then goes through it again where it actually renders it with that bit rate. But does this result in higher quality footage (even the slightest bit, on a pixel level) with my high bit rate? Seriously if it even fixed those several pixels that make a word look jagged I would like to know.

A: Looking for the BEST of the best? This is it. It is virtualy perfected bit rates. It will fix those itty bitty pixels that are giving you trouble. It is kinda hit and miss on the file size, for higher defenition video it seems to add a bit onto it, but not by enough to worry (several MB). However if you are uploading to Youtube, I wouldn't worry about using this, as any gain that you see here is so small it will be lost to Youtubes video processor. If you really want to use it though, be prepared to wait because it doubles your encoding time.

8) [SOLVED] What are the advantages of different file formats? AVI seems to have a large file size while MPEG4 is smaller. In fact the only reason I am using MPEG4 is because I heard it plays better with youtube. Is there a disadvantage to this file format vs AVI or any others worth mentioning?

A: Youtube's native HD file format is MP4, so it is best to encode it into an MPEG4 so that Youtube does not do a file coversion on your video, resulting in lower quality.

If you actually read that wall of text than congrats! And if you responded than thank you very much!!

 

[kristian221]

Too much text?

 

[bmwbaxter]

Too much text?

Not a ton of hardcore video editors around here. Give it some more time and someone will chime in I am sure.

FWIW, almost sounds like you are asking for more than what is possible. How good is the original Fraps footage? You can almost never get better quality than the original.

 

[kristian221]

If you have ever used fraps the first thing you will notice is that the video it records is a massive file. It is several GBs a minute (depending on your resolution of course, this is at 1920x1080). The original looks exactly like you are playing the game yourself, you could never tell fraps footage from live gameplay comparing them side by side. But after rendering it with Sony Vegas you get a much smaller file, and I know I won't be able to get the same quality as the original footage. But I don't know, for example, if one format would have an advantage over another. An AVI file converted to a MPEG4 file will have a much smaller size than the original, and I do not know if there is a downside to that at all. In fact I am adding that to the OP!

And I suppose I should probably be asking this elsewhere, I just trust this community more than any other to provide accurate and reliable information, so I figured I would try.

 

[ebug122]

2) Rendered frame rate should be the same as that of the recorded footage. Unless you want a clip at half speed (slow motion shot). In that case, the recorded footage would be 60FPS while the rendered clip would be 30fps.

3) Upper field means the odd numbered lines will be drawn to the screen first in one pass, then the even numbered lines.

Lower field is the opposite. Even numbered lines are drawn to the screen first in one pass and then odd.

The above is also known as interlacing. 1080i <- i stands for interlaced

Progressive is when the lines are drawn to the screen in order from top to bottom, in one pass. 1080p <- p stands for progressive

See example illustration here.

4) Pixel aspect ratio is fine at 1

If your end goal is to upload these videos to YouTube, here is a guide of DOs and DON'Ts, a list of suggested bitrates as well as supported file formats.

Don't quote me on this but I think YouTube converts most HD videos to MP4. If you use that format, the uploaded product should look like the original rendered file with minimal loss of quality.

 

[M3dley]

I do a lot of editing of ski and skateboard footage and in all honesty the best compression/rendering methods have come from vimeo. They are also quite easy to understand if you are new to it. You can check it out at the link below.

http://vimeo.com/help/compression

Hope that helps a little. This should also work for other websites like youtube etc. most have the same upload size requirements.

 

[kristian221]

Updated the main post with the answers I have gotten so far, I even learned 320kbps is low audio quality for HD video! I only have 3 questions remaining, and two of them are big ones!

 

[jlwmanagement]

Probably too late, but I do alot of Video processing and this forum is Awesome!
http://www.videohelp.com

They have a ton of info, tools, etc

 

[ssjwizard]

Two pass encoding is the video equivalent of anti aliasing in games. The first creates a rough encode within your target bitrate. The second pass reexamines the frame and does the final encode. This process can make edges look better and motion look smoother. It does not always make the file smaller depending on the original source complexity.

Reference frames is the same a Key frames. These are basically how often you want the video to verify sync. More key frames can improve high action scenes, but can cause playback delay if you set it to low.

You may also consider using handbrake to encode its pretty powerful tool with alot of options.

 

[c627627]

Where are the actual answers to questions you marked as solved?


You can somewhat clean things up but you cannot really improve low res material like you see in sci-fi movies where they zoom & enhance out of thin air blurry low res original footage.


Constant, not variable bit rate is of course "better".


My personal notes on using Sony software, they were really written with standard DVD limitations in mind. Unlike Blu-ray, standard DVDs should have 9100 kbps max video bit rate if using AC3 audio and maybe 8100 kbps max video bit rate if using uncompressed LPCM wav audio:

 

[Pinky]

#7, regarding two pass encoding -- yes, this will help the quality of the final output. It's a way of the VBR (variable bitrate) codec to ramp up when needed, and decrease when unneeded. This saves on overall file size but retains good quality output. I wouldn't even bother encoding in a compressed format unless I was using this.

Oops, I see ssj wizard also helped answer this...

Super encoder is also very powerful but has a bit of a learning curve (at this point though you're probably savy enough to be able to understand the settings).

http://www.erightsoft.com/SUPER.html

Download link: http://www.erightsoft.org/GetFile.php?SUPERsetup.exe

 

[kristian221]

Where are the actual answers to questions you marked as solved?


You can somewhat clean things up but you cannot really improve low res material like you see in sci-fi movies where they zoom & enhance out of thin air blurry low res original footage.


Constant, not variable bit rate is of course "better".


My personal notes on using Sony software, they were really written with standard DVD limitations in mind. Unlike Blu-ray, standard DVDs should have 9100 kbps max video bit rate if using AC3 audio and maybe 8100 kbps max video bit rate if using uncompressed LPCM wav audio:

It says solved next to each question, with the answer under the question for people who stumble upon this post in the future to get an answer to their questions

And all my source material is 1920x1080 FRAPS footage with lossless RGB, 30fps. So it is good source, I just want it to come out of the rendering process without losing too much of that raw quality it has.

So I did some digging and 16 referance frames is the max you can use, and I see no reason not to as it only added about 30 seconds onto the processing time of a 9 minute video, and nothing to the file size. So that is solved.

Two-pass basicly gives you the highest possible quality in terms of bit rate, but when the source is a high resolution to begin with, and your bit rate is very high, the effects of it are virtualy unnoticable. Don't get me wrong, if you look close you can see them, but we are talking that one pixel I was talking about that was out of place kinda detail. Honestly, considering youtube will not allow that pixel perfection, and add on the fact that this double the encoding time, it is just not worth your time.

That leaves the question of the progressive downloading which for the life of me I can not find out if it applies to Youtube or not. -.-

EDIT: Color coded the questions and the answers
EDIT EDIT: Either I have been at the office too long, or my red blue post looks really trippy and 3D :S Might change if anyone complains about it.

 

[c627627]

Off Topic but here are some nice colors

 

[bmwbaxter]

EDIT: Color coded the questions and the answers
EDIT EDIT: Either I have been at the office too long, or my red blue post looks really trippy and 3D :S Might change if anyone complains about it.

a bit trippy

 

[kristian221]

Changed the color, my hope is that the many people who have these questions (there are a lot of people who like to upload games to youtube) will be able to just get them all answered here in one handy thread. One question left to answer!

 

[c627627]

You are best off if no conversion whatsoever is being done. It makes no sense to spend time getting the file to be in a format different from what will be used for the end product.

So for the purposes of making a standard DVD, there would be limitations for that, for the purposes of a Blu-ray, limitations would be different and if YouTube has a standard and that's where the ultimate file destination is going to be - then whatever the YouTube limitations - that's what you want.

In other words, it's one thing to talk about playing the file directly and another to talk about playing it *FROM* a standard DVD, from Blu-Ray or from YouTube.

 

[kristian221]

You are best off if no conversion whatsoever is being done. It makes no sense to spend time getting the file to be in a format different from what will be used for the end product.

So for the purposes of making a standard DVD, there would be limitations for that, for the purposes of a Blu-ray, limitations would be different and if YouTube has a standard and that's where the ultimate file destination is going to be - then whatever the YouTube limitations - that's what you want.

In other words, it's one thing to talk about playing the file directly and another to talk about playing it *FROM* a standard DVD, from Blu-Ray or from YouTube.

But raw fraps footage is roughly 2-4GB per minute. If you are recording for an hour you could never upload that, it just wouldn't happen. And you could not split it without once again rendering it. Youtube would not allow a file that wouldn't fit on an average hard drive. Plus you would not be able to edit the footage in any way if you were not going to render it again. It is just a must, and there are a lot of variables that go into rendering as you know, so many things you can do to try and get the same quality footage as the original while being able to edit it in your own way, that it can be a bit overwhelming at first. So I guess I didn't just make this thread for me, but for anyone who might be looking for an answer to these questions.

 

[c627627]

You are correct. It's always recommended to post for the benefit of someone hitting your thread through a search, rather than as just a reply.


Regarding your general topics, when I edited for standard DVDs, it was "better" to
1. Have the highest constant rate rather than variable rate.

2. Variable rate was used to "fit" longer material on standard DVD since single layer DVD could hold about an hour of max constant rate material.

3. Field order depended on the source, so, my video card capture was upper field first but camcorder footage was lower field.


Best of the best would require video card and monitor capable of displaying it.

So how much higher than 1080p [1920x1080], do people's cards and monitors go nowadays?