Handbrake now has an AppleTV 3 preset

blueroom said:

Web optimized moves the metadata to the front of the file instead of the end. Files should load faster.

Also for Blu-Ray move the RF slider to 22, or 23 (else you get some pretty large files)

Click to expand...

Do you think moving the metadata should have THAT large of an impact? It's currently taking ~2hours to buffer the movie. Do you think that could reduce it to minutes?

My movies are currently only about 2.5-4GB, which seems pretty reasonable for a 1080p movie, so I'm pretty happy with the size/quality right now. Though if this would impact the streaming performance, I would certainly try it.

Thanks for the help.

Something else is wrong. On my setup movie playback is very quick, both Home Sharing and AirPlay.

It's more likely your network setup or slow computer (Windows is mostly awful with iTunes).

My friend in on a PC and iTunes is bog slow, I'm on a Mac and it's smooth as silk.

rshock said:

Do you think moving the metadata should have THAT large of an impact? It's currently taking ~2hours to buffer the movie. Do you think that could reduce it to minutes?

My movies are currently only about 2.5-4GB, which seems pretty reasonable for a 1080p movie, so I'm pretty happy with the size/quality right now. Though if this would impact the streaming performance, I would certainly try it.

Thanks for the help.

Click to expand...

You are not moving just the obvious metadata (title, actors, artwork), you are moving a small packet of information called the moov atom from the end of the file to the beginning. The aTV needs that information contained in that atom to start playing the movie. Due to large file size (compared to a dvd) BR's are more sensitive to this than dvd's. I have experienced what you are having when I first started using the aTV. Since using the optimize function, I have not had one slow file starting problem.

If you want, you can take your problem file, open it in Subler (free) and tell Subler to optimize the file. Save it as a different file name. Then compare the starting performance of both files. This is not a guarantee, but lack of optimization is often a source of this type of problem. Many program's, like mp4tools and iFlicks, have the optimize function programed in to their transcode process. The trade off to optimization, is that if you modify the metadata, the whole file is re-written, which can be slow with some programs.

As Blueroom said, it could also be network related, but it's a lot easier (and cheaper) to fix a optimization issue than a network issue

Great! Thanks for the reply. I will try it tonight.

Just out of curiosity, what's the advantage to having the metadata at the end of the file? (i.e. Why isn't 'web optimized' the default for the ATV3 preset?)

rshock said:

Great! Thanks for the reply. I will try it tonight.

Just out of curiosity, what's the advantage to having the metadata at the end of the file? (i.e. Why isn't 'web optimized' the default for the ATV3 preset?)

Click to expand...

If the metadata is at the end of the file, if it's modified, it will be a rapid save. If it is at the beginning, and it's modified, all the rest of the data after the metadata in file will need to be re-written, which is slower. Subler does the re-save pretty quickly, but I have found iTunes to be a lot slower. So I usually do my metadata changes in Subler, not in iTunes. If you do it that way, you just have to go into iTunes, select the file that has had the metadata modification, do a Get Info, and it will quickly update the displayed metadata.

Like I said, many transcoding apps do this "optimization" automatically. HB added the option to enhance web streaming through optimization, long before the aTV was around. Not everyone who uses HB cares about enhanced streaming, so they make it an option.

There are 2 modifications I recommend for the aTV3 preset:
1) Picture: anamorphic strict
2) Web optimize: checked

You can then save the preset w modifications, give it a name and if you wish, make it your default preset. When you open HB it will be ready to go. (Although I do always check video, audio and subtitle settings before hitting start.)

I have found that I can tell a big difference in quality when I change the quality setting from 20 down to what I typically use which is 16.5. I do have a 60" Pioneer Kuro that I am viewing on so maybe that makes the quality difference stand out that much more to me. One other thing that I noticed when I move from 16.5 up to around 20 is that the video becomes noticably darker. I did a test on the beginning chapters of The Fifth Element and I could really see a visible difference both in the picture detail but also the brightness of the picture (with all other factors being the same - only changing the quality number - Maybe Dynaflash can comment on why this may be).

Now obviously encoding at 16.5 and 1080 transfers created huge files - almost the same as the source. To this end I would have the same issue as a previous poster where it would take forever for a movie to start over my Wirelss N network. However, I have found that the same thing is true when trying to Airplay movies taken with the iPhone 4s (a 5 minute clip can take at least a few minutes to buffer). 1080p video is just a beast to push across a Wireless N network.

In order to compromise, and since I feel that if I really want the best experience I will just plop in the Blu-ray itself, I decided to do all my blu-ray transfers at 720p. This allows me to keep the quality setting at 16.5 and get a file size that is reasonable (typically around 6-7GB for a 2 hour movie) while still producing an image that looks good on my 60" screen. At least to my eyes, encoding at 720p with a qulaity value of 16.5 looks better than a 1080p image encoded with a qulaity value of 20 (your milage may vary). And like i said, if I really want the BEST quality, I will just put in the original Blu-ray.

Mac Eagle said:

I have found that I can tell a big difference in quality when I change the quality setting from 20 down to what I typically use which is 16.5. I do have a 60" Pioneer Kuro that I am viewing on so maybe that makes the quality difference stand out that much more to me. One other thing that I noticed when I move from 16.5 up to around 20 is that the video becomes noticably darker. I did a test on the beginning chapters of The Fifth Element and I could really see a visible difference both in the picture detail but also the brightness of the picture (with all other factors being the same - only changing the quality number - Maybe Dynaflash can comment on why this may be).

Now obviously encoding at 16.5 and 1080 transfers created huge files - almost the same as the source. To this end I would have the same issue as a previous poster where it would take forever for a movie to start over my Wirelss N network. However, I have found that the same thing is true when trying to Airplay movies taken with the iPhone 4s (a 5 minute clip can take at least a few minutes to buffer). 1080p video is just a beast to push across a Wireless N network.

In order to compromise, and since I feel that if I really want the best experience I will just plop in the Blu-ray itself, I decided to do all my blu-ray transfers at 720p. This allows me to keep the quality setting at 16.5 and get a file size that is reasonable (typically around 6-7GB for a 2 hour movie) while still producing an image that looks good on my 60" screen. At least to my eyes, encoding at 720p with a qulaity value of 16.5 looks better than a 1080p image encoded with a qulaity value of 20 (your milage may vary). And like i said, if I really want the BEST quality, I will just put in the original Blu-ray.

Click to expand...

Why not just remux your BR to mp4?

Because I do want the space savings, especially for putting the kids movies on the iDevices (I do a 480p version as well for this purpose). For me it was a personal balance between quality and size. Before I started doing a large selection of my collection I ripped at various settings and picked the one that met my criteria.

If I just wanted the best quality and convienince of having everything instantly available (i'm not that lazy... yet) then yeah, I would have just done what you suggested. Besides I still like some of the interactivity and extras of the Blu-ray format so I will typically watch it that way first.

Mac Eagle said:

Because I do want the space savings, especially for putting the kids movies on the iDevices (I do a 480p version as well for this purpose). For me it was a personal balance between quality and size. Before I started doing a large selection of my collection I ripped at various settings and picked the one that met my criteria.

If I just wanted the best quality and convienince of having everything instantly available (i'm not that lazy... yet) then yeah, I would have just done what you suggested. Besides I still like some of the interactivity and extras of the Blu-ray format so I will typically watch it that way first.

Click to expand...

What you say makes sense for the multi-iDevice need. I was more questioning why use HB at such a high quality setting, yielding files around the same or bigger than the original, when you could remux very quickly and leave the original video/audio quality unchanged.

mic j said:

What you say makes sense for the multi-iDevice need. I was more questioning why use HB at such a high quality setting, yielding files around the same or bigger than the original, when you could remux very quickly and leave the original video/audio quality unchanged.

Click to expand...

But i'm reducing to a 720p image and even at a quality setting of 16.5 I'm getting between a 50-75% file size reduction over the original.

Mac Eagle said:

I have found that I can tell a big difference in quality when I change the quality setting from 20 down to what I typically use which is 16.5. I do have a 60" Pioneer Kuro that I am viewing on so maybe that makes the quality difference stand out that much more to me. One other thing that I noticed when I move from 16.5 up to around 20 is that the video becomes noticably darker. I did a test on the beginning chapters of The Fifth Element and I could really see a visible difference both in the picture detail but also the brightness of the picture (with all other factors being the same - only changing the quality number - Maybe Dynaflash can comment on why this may be).

Click to expand...

imo rf 16.5 is ridiculous for blu ray or even 720p hd sources. Of course while the rf goes "down" the "quality" goes up. However to speak to what you asked me specifically about the "brightness" that really does not make sense, however possibly the lower quality might be hitting by chance less bright spots in the source. It really shouldn't show any difference in "brightness" though. Is it possible you are perceiving the lack of sharpness and detail in the lower quality encode (20) vs the higher (16.5) encode ? I really don't know tbh. There is no empiracle evidence this would be the case.

Just wanted to update the thread and say that I tried using the 'Web Optimized' check box in HB, and that did indeed fix the streaming problem. Files that were taking 2+ hours to buffer and start playing, now play in seconds. Thanks for all of your help!

rshock said:

Just wanted to update the thread and say that I tried using the 'Web Optimized' check box in HB, and that did indeed fix the streaming problem. Files that were taking 2+ hours to buffer and start playing, now play in seconds. Thanks for all of your help!

Click to expand...

Glad your problem was fixed...easily.

dynaflash said:

imo rf 16.5 is ridiculous for blu ray or even 720p hd sources. Of course while the rf goes "down" the "quality" goes up. However to speak to what you asked me specifically about the "brightness" that really does not make sense, however possibly the lower quality might be hitting by chance less bright spots in the source. It really shouldn't show any difference in "brightness" though. Is it possible you are perceiving the lack of sharpness and detail in the lower quality encode (20) vs the higher (16.5) encode ? I really don't know tbh. There is no empiracle evidence this would be the case.

Click to expand...

Thanks for your input dynaflash. It has been several versions since I conducted my comparisons so maybe that has had some affect but I will re-run my test on the current nightly build that i'm using. 16.5 may be overkill, but anything over 18 definitely had a lack of detail and poor gradients that were noticeable to me. I guess at the time these were the settings that gave me the best picture quality/file size compromise. I plan to revisit my test in the next few days and see if that still holds true.

fwiw, it looks like we might be making web optimized (which is kind of misleading) default in the next HB in an effort to fix the issues with atv3 streaming of big files. For now just check it in HB to maximize the streaming efficiency.

----------

Mac Eagle said:

Thanks for your input dynaflash. It has been several versions since I conducted my comparisons so maybe that has had some affect but I will re-run my test on the current nightly build that i'm using.

Click to expand...

How many versions ? HB as well as libx264 had been upgraded many times since even the last public release 0.9.8. Try a nightly and it should be much better depending on which version you have been working with.

I have noticed a brightness difference as well in frame grabs from an ATV3 encoded file, and the original mkv file ripped from a Blu Ray disk.

Ok, I have been able to re-run my tests using one of the latest nightly builds and I have to say that my results were much different this time. I have to assume that the encoding of HB has improved since the last time I ran this kind of test (probably about a year ago) as my results are much different this time.

I did a direct comparison primarily using the aTV3 preset and modifying the RF setting. Unlike the last time, I am unable to see much difference between the RF i am currently using (16.5) and RF 20. (dynaflash now saying "see, I told you so"). I can see a slight sharpness increase at 16.5 but to be honest I have to look for this on paused images side-by-side.

I also went back and re-tested my Fifth Element encode where I was seeing a definite brightness difference between RF20 and RF 16.5 and this time I did not get the same result and the image brightness remained consistent between the encodes.

So I guess I will be rethinking my RF setting. I am now thinking that I may go up to at least an RF of 18 at a 720p resolution. For me, I am looking at saving file space while retaining an acceptable picture quality (like I've said, if I want the best I'll put in the original Blu-ray). I looked at both the 720p and 1080p image in full screen mode on my 27" iMac screen and found that I could see very little difference in quality. I would have suspected that I would have seen more artifacts in the 720p image. I have not yet compared these two files on my 60" Kuro yet.

DYNAFLASH - I still have a couple lingering questions that hopefully you can answer.

1) From all that I have read, Blu-ray being fixed pixel count, they are not anamorphically encoded. Therefore, does using the Anamorphic setting (either strict or loose) really do anything? Is using the None option really the correct thing to do?

2) When using the cropping option, as long as you leave the width the same (i.e 1920 or 1280) you do not affect the overall image quality; however, if you have a 4x3 image in HD and you allow HB to crop the image and the output width increases from the original then technically you are adding more information than is really there in the original and compromising quality? For example, say you have a 640x480 image but have cropped 20 pixels off each side but the resulting output size is still 640x480 are you now stretching the image and (further) compromising quality?



Just for info, my results are tabulated below. I am using an 2011 iMac with a 3.4Ghz i7 processor.

HB Version 4728svn x86_64 (2012061401)
RF Resolution File Size (MB) Time to Encode (min)
16.5 1080p 776.1 10
20 1080p 487.8 7
16.5 720p 423.6 5
18 720p 362.1 4
20 720p 307.6 4

The above test was performed using the Avatar Blu-ray and chapter 31 which is 5:19 long. The base setting was the AppleTV2 preset and only changing the RF value.


HB Version 5170svn x86_64 (2013101501)
RF Resolution File Size (MB) Time to Encode (min)
16.5 1080p 750.5 11
18 1080p 611.7 9
20 1080p 475 8
16.5 720p 415.2 4
18 720p 356.1 4
20 720p 303.5 4

The above test was performed using the Avatar Blu-ray and chapter 31 which is 5:19 long. The base setting was the AppleTV3 preset and only changing the RF value.

Mac Eagle said:

DYNAFLASH - I still have a couple lingering questions that hopefully you can answer.

1) From all that I have read, Blu-ray being fixed pixel count, they are not anamorphically encoded. Therefore, does using the Anamorphic setting (either strict or loose) really do anything? Is using the None option really the correct thing to do?

Click to expand...

This was my thoughts on the matter as well. I would appreciate it if someone could really explain what purpose (if any) the anamorphic settings have on a blu ray source. I've spent a lot of time searching this forum and have answered pretty much all of my questions by doing research (hence why this is my first time posting), but I can't find out why you would not want to use "none" on blu-rays, since they aren't anamorphic like dvds.

Go ahead and use none if you want on blu ray. Its a preset which means it is designed for all sources. If you want to use "none" ... then use it!

Hi Dynaflash, thanks very much for the response. Just to clarify, are there any advantages (or disadvantage) of using none versus strict or loose when it comes to blu-ray sources? I had assumed that handbrake basically ignored the anamorphic setting if the source video wasn't coded anamorphic - but I did a test encode from the same blu-ray source using anamorphic strict and none (all other settings were identical) and the files have a different checksum, so obviously handbrake did something different - just not sure what.

Sorry if this is a silly question, as I am sure the result of either setting would be impossible to distinguish from one another - I just was curious, which, if any was optimal

Personally, I use strict on everything. You should see no difference.

Thanks! I'll just stick with strict since it's what you use and it will save me from losing picture quality if I was to accidentally use my blu-ray profile with a regular dvd. I guess the only time I would ever need to use none or loose was if I wanted to convert a 1080p source to 720p, is this correct?

yep, if you use loose set the mod to 2. its as accurate as strict actually at that point.

Thanks again. Oh, you probably already are aware but on the off chance you aren't; the website listed on your profile (http://dynaflashtech.net) doesn't seem to be valid.

Infinity3 said:

Thanks again. Oh, you probably already are aware but on the off chance you aren't; the website listed on your profile (http://dynaflashtech.net) doesn't seem to be valid.

Click to expand...

I am aware. But thanks!