[Norman Black]

[jenny gold] “high bit rate he went for in his screen shot max 24,000 and average 12,000”

One very important difference here, that is a 1080p30 template and NOT 720p30. Those settings are the same as the MC AVC Internet 1080p defaults. We have been talking about 720p files. 1080 has roughly twice as many pixels to encode as 720 has. Therefore 12Mbps in 1080 is actually a little lower relative bitrate than 8Mbps in 720. It is only a 50% higher bitrate for twice as many pixels.

For a lot of source material you don’t need twice the bitrate, in 1080 vs 720, to maintain similar quality but it is something of a common convention. My GoPro cameras do this. Smugmug online video does this.

[Norman Black]

[Graham Bernard] “Question1: Have you tried the Upload to YouTube option, directly from within VegasPro? Have you any idea just what Vegas is using? What would YT then be using to re-encode – if it does?”

I have never looked at that option in Vegas. Some online have said it uses Mainconcept AVC as the encoder, which I have every reason to believe. I have no idea what encoding settings it uses. I would bet they are probably conservative, much like the “Internet” templates.

Youtube uses x264 to re-encode. They always re-encode. All online services are like this. They do not officially state the encoded bitrate target for their result. They only give suggested upload specs, and these are higher than their result file. I have used capture programs to DL the streamed file and use MediaInfo to dig into the encode specs of the online host.

Since they always re-encode I like to upload something that will be a bit better quality wise that what I get back. My view is to let the encoder have enough information so that it makes the best compromise decisions to encode the video within the allowed bitrate.

[Graham Bernard] “Question2: I’ve been using the NEW automatic/One-Click Vegas >> Frameserver >> Handbrake >> Vegas that Marco has provided for us. I like it lots! Any thoughts on using your recipes to intervene to make a BETTER YT option?”

I to have also been frameserving from Vegas recently. But I have been serving to ffmpeg.

When using Handbrake (aka x264) to encode for online uploads I feel VERY safe getting my upload bitrate close to the result bitrate. I would not upload at the host result bitrate as I would like to be conservative. Maybe 25-33% above the online service bitrate. I know Smugmug. I have only experimented with Youtube a little while ago, but the bitrate was similar with YT being a little lower than Smugmug, but at better encode settings. Smugmug has recently improved their encode settings.

Online service 720p30 bitrates are around 3Mbps. When they do not spec the encode result it takes a little testing to figure this value out. They typically encode with two pass average bitrate mode. I would feel safe/conservative uploading something 4-4.5Mbps average, High profile, from Handbrake.

When I used Handbrake to encode uploads I used two pass at 4.2Mbps, high profile, medium preset. I later tried the single pass CRF mode (constant perceptual quality). CRF 27-29 gave about the same bitrate as my two pass in a few test encodes.

So my recommendation would be to experiment with CRF modes to get something close to a 4-4.5Mbps average bitrate, for 720p30, across a few test encodes of your source material. Or use two pass to be exact. This assumes an approx 3Mbps host result average bitrate. Adjust as necessary.

Use High profile. Main profile is good too. It encodes faster than High, but not as small. Never baseline profile. High is “safe” when you don’t know what the host is doing. I know Smugmug now supports High profile in their 1080p files, but still only supports Main in their 720p files. I do not know what Youtube currently does. I know they have used Main profile at 480 and up. When trying to minimize your upload size, I think it important to have your settings as good or better than the host online encodes.

I don’t think the “tune” options are useful with lower bitrate online encodes.

With CRF mode, using better/slower encode settings does not change the quality. Only the result bitrate (aka file size). The opposite is true for average bitrate modes. Fast, Medium (default) and Slow presets are all good middle of the road choices. I think it best to avoid the advanced panel. The x264 developers are smarter than everyone else and they have provided a wide range of presets and tuning options that intelligently select all those advanced controls for us.

There are a couple of schools of thought here with CRF encoding, and there is nothing right or wrong about either. Fast encode settings and bigger files sizes. Slower encode settings and smaller file sizes. It is a balance of the speed of your computer versus the speed of your upload.

This small file encode optimization only matters if you want to minimize your upload time. Otherwise upload something big and sleep well at night. I think with too much effort we can be more efficient.

For files I have on my media PC and watch on my 55″ TV. I encode with slow encode settings. High profile, Slow preset, Film tune setting and a CRF appropriate for the camera output bitrate. Only with a few test encodes can you figure out what the “appropriate” CRF value is.

Those would be my guidelines, for you to do your own tests, with your own eyes on your own source material.

[Norman Black]

[jenny gold] “I have attached screen shot of my render adjusted bit rate settings…so are they correct to use final render to upload to youtube?
are there some other settings here also I should do – if so please let me know”

No, they are not correct, but more on that later in this post.

[jenny gold] “from my understanding what you are saying Youtube adjusts the bit rate right down anyways and when it was originally set in vegas at 8mbps it was to high and overkill “

Youtube re-encodes everything we upload to their specifications. I have not tested in a while but in the past it has been around 3Mbps, main profile. 8Mbps out of Mainconcept AVC or Sony AVC is just fine for upload, but many people want to minimize their upload time. The Vegas encoders do not need 2.6 times the bitrate to keep a similar quality to what YT will give. I have stated multiple times I have been using 2x with Sony AVC which is still a little overkill, IMO, but I am fine with that.

[jenny gold] “My project property settings in vegas pro 13 has the option of:
pixel format 8 bit
or 32 bit floating point either (full range) or 32 (video levels)
which should I use?”

Let Vegas match your project setting to your media settings. It will be 8-bit I am sure.

[jenny gold] “1) you suggested as an option if at all I was going to use handbrake then have high settings set for render in vegas and bring that into handbrake. Because handbrake will reduce this down and bit rate etc as final rendered version .

and my understanding what you were saying is that handbrake version ends up being a better quality then had I just used sony vegas pro main concept render setting and the 3mbps I may have even used?
have I got that correct in what my understanding you were saying ?

Which begs the question:
Is the final quality using handbrake final version noticeable quite bit compared to say only using vegas to render ?

I am not sure what you are asking here.
What I said was that if you want to use Handbrake for your renders then give Handbrake the best quality file you can. I gave a couple of encode options. XDCAM EX 35Mbps and XAVC-S (not XAVC).
Using Handbrake for your final renders is a common practice and I do it myself for some things.

[jenny gold] “(at the above settings and screen shot I have and 3mbps you suggested for youtube)”

I NEVER suggested a 3Mbps encode setting for Youtube. You have taken my 3Mbps comment WAY out of context. Which brings me back to the start. I am sorry I have failed trying to explain my encoding views. Apparently I have only confused you and I do not know how to state things any differently. Just because I understand things a bit and have tested things does not mean I can explain it. I was probably too pedantic which probably clouded what I was saying.

As for your settings screen shot. IF you want to go for an approx doubling of the typical video service bitrate which is what I have said that I normally do, then you can try these settings for MC AVC.

High profile, average bitrate 6Mbps, Max bitrate 12Mbps. I am not sure why you set both values to the same value in your screen shot. The default templates typically set the max bitrate to about 2x the average average. One would think that if you change the average you keep the same ratio between the two. You can set slices to 1. It can result in better compression and thus quality. Two pass should give a bit better quality but at twice the encode time. I normally skipped that. As per Youtube/Vimeo recommendations I would set the audio bitrate to something higher than the Vegas default of 128K. I personally use 192K. The YT page recommends 384k.

The MC AVC encoder will not give you exactly the average bitrate you request in the template. Two pass is closest, and two pass CPU only is very ideal. This is not too critical but I am being pedantic again.

Handbrake in average bitrate mode is always ideal. In CRF mode the resultant bitrate is unspecified and unknown. If encoding with Handbrake your upload file bitrate can be very nearly the resultant video service bitrate. This is because the Handbrake quality will be same or probably a little better than the video service encode. Not true for the Vegas AVC encoders at the same bitrate.

I upload 5-10 minute videos. It is quicker for me to encode once in Vegas and upload something a bigger, than encode in Vegas and then encode that result in Handbrake and then upload the file.

Your eyes are the one and only true judge as to what is good enough quality.

[Norman Black]

Let me lay out how I came to my conclusions. The the sake of argument lets say the online service encodes to 3Mbps AVC.

I encode to XDCAM EX 35Mbps and this is my intermediate for Handbrake.

I encode using Sony AVC 720p30 to High profile 3Mbps. Save that.
I encode using handbrake to High profile 3Mbps, two pass.

Visually compare those. The Sony AVC will not be near the quality of the Handbrake (x264) encode. Both are using AVC/H.264 encodes. We all know that some brands of cars are better than others. The same is true for encoders, that are encoding to the same encoding “language”. In this case AVC (H.264). x264 AVC goes farther on a gallon of gas (aka bitrate) than Sony AVC does.

Now I encode Sony AVC at a higher bitrate. Say 50% higher, 4.5Mbps, and compare that to the 3Mbps Handbrake encode. At some point the Sony AVC, with more bitrate (more gas) will look the same (drive as far) as the Handbrake encode.

x264 is the encoder that the online services use. I think Mainconcept is a little better than Sony AVC. x264 better than both. You can always get the same visual quality out of all of them. Some encoders need a little more bitrate than others for a “constant” visual quality.

As previously stated I am doubling the bitrate my online service encodes to. I use Sony AVC because it is faster on my particular machine. That is very conservative amount, and comes to about 6.5Mbps. If I REALLY wanted to get smaller uploads I could push it to maybe 50%, but I just do not want to bother. 6.5Mbps is a very manageable upload. Conservative and no worries, ever.

I compare the online service result to the video uploaded. The upload is always a bit better than the result. The result is fixed by the encoded bitrate. Uploading something twice as good as the result and 10 times as good will get the same results. The 10 times bigger one will just take longer to upload.

And as stated for my home playback, I do not care about file size and go for quality at CRF 22.

[Norman Black]

Mainconcept AVC Internet 720p30 is 8Mbps average bitrate. This works but is excessive, IMO. You can customize that template and go lower to get a smaller upload, like I do as previous described. Places like Youtube re-encode the file to around 2.8-3.2Mbps.

If you render a low bitrate like the “Internet” templates, then DO NOT re-encode this low bitrate file in Handbrake for something even smaller. You have already lost of detail and will lose more with a re-encode.

Either encode directly to something smaller by changing the bitrate in the template(s) and upload that. Or encode to a high quality, very large file, intermediate to feed Handbrake for you final encode. The encode to an intermediate loses basically nothing and then your final encode in Handbrake is the only place you get any loss you might notice. I gave some options in my previous post about intermediate encodes. Of course, you do get some loss since the online video service re-encodes everything we upload.

The templates in Vegas are good starting points, and the Internet ones are conservative, in my opinion, and you can get some smaller uploads. If that matters to you, then customize them for your specific needs. That is what I do as stated in my previous post. Additionally I bump the audio bitrate up to 192Kbps since the online services are going to re-encode audio as well as video.

Here are some page the online services give for guidelines for encoding.
https://support.google.com/youtube/answer/1722171?hl=en
https://vimeo.com/help/compression

Take everything I say as a guideline for your own judgement and testing. The final and best judge is always your own eyes.

[Norman Black]

Here is a V13 VEG project file showing the mask operation as described.
7948_maskdemo.veg.zip

[Norman Black]

The size of your rendered file is completely dependent on the rendered bitrate chosen. It has nothing to do with the encoder you use.

That said, some encoders are better than others. The main encoder in Handbrake that people are suing is x264. This is an AVC/H.264 encoder just like the Mainconcept AVC and Sony AVC encoders native to Vegas. At lower bitrates the x264 encoder will look better than to encoders in Vegas, when all are outputting the same bitrate. At high bitrates they all look the same. Higher bitrate means higher visual quality within the same encoder.

I would recommend that if you use Handbrake, do not use the Advanced tab. The standard presets designed by the developers of x264 are good. They are smarter than you or I in what all the individual options do.

The standard presets go from faster to slower encoding speed. The standard recommendation is use the slowest you can bear. Does this faster to slower speed mean lesser to better visual quality. Yes and No. If you are using the CRF bitrate mode then NO. If you are using the average bitrate mode then YES.

With CRF bitrate mode you do not control what the resultant bitrate of the file is but are specifying a roughly constant visual quality.

So in CRF mode using a slower encode preset will give a smaller file than a faster preset.

There is so much to say about file encoding, but I’ll stop here. There is stuff online you can read and you can ask questions here.

What do I do?
For uploading to Smugmug (my equal to Youtube), I render a file using Sony AVC at about 2X the bitrate that Smugmug re-encodes to (3.2Mbps main profile). That is excessive but this is still only 6.4-7Mbs bitrate at 720p30.

I have used Handbrake at High profile, fast/faster preset, crf 26/27. I have also used two pass average bitrate at about 5-6Mbps, High profile, medium preset. But these modes require I render a high quality intermediate from Vegas to use in Handbrake. The single render from Vegas is just quicker.

For files I playback on my media PC, I want MAX quality so I do use the x264 encoder and I use, High profile, Medium Preset, CRF 22. This results in high bitrate files.

For high quality intermediate files to use with Handbrake I have used the free DNxHD codec. You don’t want to use “normal” type renders to Handbrake. Others you can use are XAVC-S (not XAVC) or XDCAM EX 35Mbps.

[Norman Black]

Try using a mask. Mask is in the event pan crop window. Two keyframe mask positions should be enough. In your second example, the mask would start out masking basically everything in the first keyframe and the second keyframe would have the mask not masking anything.

[Norman Black]

If you try PNG, then remember to set the pixel to 32bpp from 24bpp.

24bpp is RGB which has no alpha, 32bpp is RGBA.

[Norman Black]

PNG uses lossless compression so it is as good as uncompressed quality wise. It uses the same compression algorithm as ZIP files. PNG will always be smaller than uncompressed. Pictures and video will not compress nearly as well as text. Some of the PNG filter options might increase the compression amount depending on your source material.

By using the QT PNG codec, then anyone you give the file will need Quicktime installed for their NLE.