-
5GB for 20min 1080p video?
Is this normal? If not, what should I be aiming for? I have this exact same video rendered at a lower bit rate (8MBs) but have been experimenting for best Youtube quality:render/upload time ratio.
I have slow upload speed (1MBs) so uploading such a file takes considerable time which I may need for other things.
I’ve tried using Handbrake to compress with a previous video but the picture was horrid and blurry. Looked like 320p.
Rendered using Movie Studio 13. I decided to go with 50MBs bitrate which in hindsight in probably a bit high for Youtube, but I figured uploading the best quality and letting Youtube do it’s thing would be a better way of going about doing it. Would doing 20MBs a better decision? I’m aware the recommended is 8MBs but the Youtube video just looks too blurry for my tastes.
Also, two pass or single pass? From my understanding two-pass would be better as it encodes better whilst also reducing file size, albeit longer render time.
The main problem I have with all of this is, is because I use the same PC for everything so that longer rendering times reduces opportunity to create new content. Same goes for the uploading.
-
18 Replies
-
HDV 1080 video is 13GB/hour so 5GB for 20 minutes sounds about right to me. File size is controlled by bitrate. The higher the bitrate the larger the file. Quality is affected by frame size, bit rate, and encoding method. 25Mbps 1080 HDV is the same quality as 16Mbps 1080 AVCHD because AVCHD uses more efficient encoding. As you can imagine, the AVCHD file will be smaller at the same quality as HDV because it uses a lower bitrate due to it’s better compression. Two pass encoding with create better quality Variable Bit Rate (VBR) files because it allocates bits better. This helps with low bit rate files but is a waste of time with high bit rate files.
You are correct that you should send YouTube as high a quality file as you think you need but no higher. Since quality differs based on content, you need to experiment with your content to see what bit rate is best for you.
~jr
http://www.johnrofrano.com
http://www.vasst.com -
Thank you for your response.
Since posting this I have uploaded a few more videos and reduced the file-size by about 2GB by playing around with the bit-rate. As luck would have it, however, this is no longer an issue for me as our ISP upgraded our upload speed.
“Two pass encoding with create better quality Variable Bit Rate (VBR) files because it allocates bits better. This helps with low bit rate files but is a waste of time with high bit rate files.”
What would you consider to be the bit-rate threshold between using Two-pass encoding and ignoring that option? I’ve been using two-pass for all renders, varying from 20-50 bit rate. Presumably rendering a video at 50 bit rate as one-pass is sufficient?
-
[Kirk Wood] “Presumably rendering a video at 50 bit rate as one-pass is sufficient?”
It depends on the bit rate of your source and details in your content. If your source was 16Mbs AVCHD then 50Mpbs doesn’t need two pass. If, however, your source was 50Mbps XDCAM you might want to use two pass when rendering back to 50Mbps. Content also plays an important role. More detail requires more bits to represent it.
Here is how it works: Two-Pass makes an initial pass to determine how many bits each frame needs, then it makes a plan to allocate the bits while not exceeding the maximum and keeping the overall rate within the average, then it encodes. If you are already using a bit rate that exceeds what each frame needs, it doesn’t matter. Why waste time with two pass (because it does take twice as long) when your bit rate is already high enough to give every frame all the bits it needs.
This is, of course, depending on the content. Content with a lot of details needs more bits to accurately represent those details. A scene of the ocean in motion needs more bits per frame than mountain scenery with no breeze, and both need more than a blank wall in a room. So there is no magic number. It’s just that at some point two pass isn’t needed anymore when you have enough bits to go around.
~jr
http://www.johnrofrano.com
http://www.vasst.com -
Thank you for the explanation, however I’m still uncertain as to it’s usage in my scenario (and you’ll have to bear with me because this is new to me).
I’m recording with nVidia Shadowplay which saves footage in .avi format, my settings being 1080p, 60FPS at 50MBs.
My project properties are 1080p, field order Progressive Scan, 59.940 frame rate, rendering quality at Best, and deinterlace method as Blend Fields.
When rendering, I select .mp4 format, Mainconcept AVC/AAC, with a custom template based off the existing “Internet HD” template. Profile is set to High, frame rate is set to 59.994, field order is Progressive Scan, aspect ration is at 1.0000, reference frames is 2, deblocking filer is Enabled, Varible bit rate is checked, currently two-pass is checked, and at the moment I am currently experimenting with bit-rate. Progressive download is enabled.
With that in mind, what would your suggestion be in regards to using Two-pass or not? Since it’s Youtube, should I go with a lower bit rate at one-pass, or render at the original bit rate with two-pass? Something else.
Sorry for all the questions, just feel it’s best to ask someone who actually knows about this. 🙂 I appreciate your time.
-
[Kirk Wood] “With that in mind, what would your suggestion be in regards to using Two-pass or not? Since it’s Youtube, should I go with a lower bit rate at one-pass, or render at the original bit rate with two-pass? Something else.”
Your question is stated backward so I want to correct that first. Your options would be:
- low bit rate with two-pass
- vs high bit rate with one-pass
If you want to play with lower bit rates, definitely enable two-pass. It will make better use of the available bits and give you higher quality. If you are OK with generating 50Mbps files then there is no need to use two-pass.
Gameplay is particularly tough on encoders. In the real world, your eyes can’t see all that detail and video cameras don’t record it, so video encoders use this to their advantage. That’s why they drop information about colors that your eyes are not sensitive to. Computer generated graphics, on the other hand, have extreme amounts of unnatural detail and that causes encoders to work harder to preserve it all. You would actually be better off using a slight (0.002) Gaussian Blur on your master video bus to help the encoder by softening the image ever so slightly. The problem is, I know a lot of gamers don’t like this softening. They expect their YouTube videos to look just like their original computer monitors did. So it’s a constant trade-off between details and bit-rate.
~jr
http://www.johnrofrano.com
http://www.vasst.com -
Oops, yes, that’s what I meant.
Well, I’ll play around with it a bit more and see what happens. Thank you very much for your advice.
-
great post Kirk…. that saves me asking lots of same questions… hope you don’t mind me posting on your thread..?
John… My footage is extreme sports action jumping from aircraft with lots of people movement (obviously) shot with Gopro mainly at 720p 60fps but some at 120fps experimenting for slo-mo quality using the ‘Flux’ which comes with GoPro studio 2.5 (makes for much smoother slow motion although it’s more ideal if the camera is fixed and that is rare as I’m part of the action 99.9% of the time, but it still improves on what it’s like without Flux).
As I’ve been working on my project I’ve been rendering to Sony AVC/MVC .MP4 using the ‘Internet 720’ template and not playing with the settings as I don’t yet fully understand them.
1. This gives me a pretty good end result although I’d like to know if I can improve on the quality..?
2. Have you done a tutorial on customising render template settings and what each setting will change and why you might want to change it..?
3. I’m about to add in some 1080 footage at end of project… should I downscale to 720 when I convert to avi cineform codec in GoPro studio..?
cheers
Nick… BASE1268
3…2…1…C ya
-
[Nick McMahon] “1. This gives me a pretty good end result although I’d like to know if I can improve on the quality..?”
Quality can be adjusted by increasing the bit rate and/or using the Two-Pass option which will make better use of the available bits.
[Nick McMahon] “2. Have you done a tutorial on customising render template settings and what each setting will change and why you might want to change it..?”
No, I haven’t and Sony does’t even document what all of the parameters are for. I don’t know why they let you tweak parameters without documentation but I usually leave the templates as they are with the exception of adjusting the bit-rate when needed.
[Nick McMahon] “3. I’m about to add in some 1080 footage at end of project… should I downscale to 720 when I convert to avi cineform codec in GoPro studio..?”
Using CineForm is a great idea. If GoPro Studio does a better job of resizing then I would use that. This will also make your render go faster because Vegas Pro won’t have to resize.
~jr
http://www.johnrofrano.com
http://www.vasst.com -
[Nick McMahon] “3. I’m about to add in some 1080 footage at end of project… should I downscale to 720 when I convert to avi cineform codec in GoPro studio..?”
Using CineForm is a great idea. If GoPro Studio does a better job of resizing then I would use that. This will also make your render go faster because Vegas Pro won’t have to resize.
Ah… so I think I’ve just now realised then that by setting the project properties means that any footage you bring in that doesn’t match will be resized by vegas when rendering to match..?
So what happens when I bring in an SD clip which is 3 seconds worth of this…? (Project properties are set to 720@ 30p)
Is there any way to make this lower res footage look any better..?
Thankfully it’s not on the screen long enough for people to register it’s quality difference but of course I KNOW it’s there… and it bugs me somewhat
cheers
Nick… BASE1268
3…2…1…C ya
-
[Nick McMahon] “Ah… so I think I’ve just now realised then that by setting the project properties means that any footage you bring in that doesn’t match will be resized by vegas when rendering to match..?”
Not Render. Let me restate that correctly.
Any footage you bring in that doesn’t match the project will be resized on-the-fly by vegas when playing back the timeline.
This has nothing to do with rendering. Now… if you have a 720 project and you drop 1080 footage into it, and you render to 720, Vegas will, of course, have to resize the 1080 footage for rendering. If, however, you rendered that same project to 1080, Vegas will upscale the 720 footage to 1080 and leave the 1080 footage alone. In other words, it will not scale the 1080 down to 720 and then back up to 1080. That’s the difference.
[Nick McMahon] “So what happens when I bring in an SD clip which is 3 seconds worth of this…? (Project properties are set to 720@ 30p) Is there any way to make this lower res footage look any better..?”
You can use a plus-in like Boris Continuum Complete UpRez to upscale the footage and sharpen it.
~jr
http://www.johnrofrano.com
http://www.vasst.com
Reply to this Discussion! Login or Sign Up