Jun 3 Podcast #481: 720p vs 1080p or More Compression vs Less Compression
Podcast 
Last September Apple released the AppleTV 2 and one of the things people complained about was that it was only 720p. We have talked on this program many times that resolution is not the most important specification of a good display or projector. Especially when you talk screen sizes less than 50 inches. There is study after study that says MOST people need to sit six feet away from a 50 inch TV to see the difference between 720p and 1080p.
Amazon Prime: Stream movies and save on shipping; Mmmmm, streaming...
Listen to the show
Today's Show:
Blu-ray Reviews:
News:
- NPD: Samsung Tops In Plasma TVs
- 3D LCD TV Displays Increase In Q1
- Mitsubishi 2011 Line-Up
- Over-The-Air Broadcasts Losing Viewers
Other:
720p vs 1080p or More Compression vs Less Compression
Last September Apple released the AppleTV 2 and one of the things people complained about was that it was only 720p. We have talked on this program many times that resolution is not the most important specification of a good display or projector. Especially when you talk screen sizes less than 50 inches. There is study after study that says MOST people need to sit six feet away from a 50 inch TV to see the difference between 720p and 1080p.
So we set out to see if people could see the difference in resolution sitting 14 feet away from a professionally calibrated 100 inch screen. For this test we used an Apple TV and an Oppo Blu-ray player. The movies we chose were Harry Potter Deathly Hallows and Beverly Hills Cop.
It only took us about ten seconds to see the flaw in our test. That flaw was that our experiment was more a test in compression than on resolution. Regardless we proceeded because everyone was already on their way over and there might be some decent conclusions anyway.
Process
The experiment was done by starting each movie at random scenes and switching between the AppleTV and Oppo player. This was done so that each individual was looking at each scene against itself and not in comparison to the other. A total of five scenes were selected for each movie. The subjects were given a score card with the name of the movie and two columns, 720p and 1080p. Ara kept the master. There were five people who participated in the test and only one of those was a self professed videophile who claimed he would be able to see the difference.
Results
The values below reflect the percentage of time the subject could correctly identify the difference between 720p and 1080p.
Harry Potter
- Subject 1 (Videophile) - 100%
- Subject 2 - 40%
- Subject 3 - 60%
- Subject 4 - 60%
- Subject 5 - 60%
Beverly Hills Cop
- Subject 1 (Videophile) - 100%
- Subject 2 - 80%
- Subject 3 - 80%
- Subject 4 - 100%
- Subject 5 - 60%
The Video expert nailed both tests. When we asked him what he keyed on he said, “The Macro Blocking”. In many of the scenes the Macro Blocking was a dead giveaway that the subjects weren’t watching Blu-ray. This ended up being the biggest tell to which source material was on screen. So much so that watching Beverly Hills Cop most subjects could easily tell the difference in resolution.
But even with the compression issues the non experts were not able to see the difference between the two sources all the time. It did matter what scene was being displayed. If there was a lot of motion then there is typically less compression. In Beverly Hills Cop the office scenes were always a give away since there are large mono colored walls that compress easily and show macro blocking just as easily. Bright outdoor scenes with a lot of motion were much harder to see a difference.
We figured that because Harry Potter is darker it would be harder to see differences between the two sources. The results backed our initial thoughts with all but the videophile scoring lower.
What does this mean?
It essentially confirms what we have been saying all along. If you want the highest quality picture you will want to watch Blu-ray. At this point in time most streaming solutions will be a compromise in video quality. The one exception is Vudu HQX. We will try to do a similar test with HQX video in the near future.
We conclude that for many people , streamed 720p will be an acceptable viewing experience. Even with the compression artifacts present some content will be difficult to see the difference between Blu-ray. If streaming providers were to reduce the compression (increase the bit-rate) the difference between 1080p and 720p material would be harder to distinguish and the resolution issue would go away.
The main reason the providers use highly compressed 720p content is because of bandwidth. It costs them more to stream all those bits to you. You may not want so much data streamed to you either. Some of you may be subject to data caps limited by your ISP. That would make watching high data rate 1080p content an expensive proposition.
Finally, just for grins we watched the same content on a 58 inch plasma. While the artifacts were still present they were much harder to see siting 12 feet away. The take away is that if you have a screen less than 50 inches, streaming video from Apple, Amazon, Netflix and Vudu will not only be convenient but will be of good enough quality for most people.
Next steps
The original intent of this experiment was to compare 720p vs 1080p at reasonable watching distances to see if most people could see the difference. What we didn’t consider was that our choice to use overly compressed streamed 720p content would taint the results. We will retry the experiment but only after we have secured high quality 720p content. Once we eliminate any compression artifacts from the equation we feel that it will be a much tougher test and only the true videophile will be able to see the difference.
Reader Comments (4)
If you are using an OPPO player that has quick resolution select options on the remote, wouldn't the test be more accurate to switch between resolutions from the same source material on the same device? I understand that there are then going to be other factors introduced into the test with scaling and such but it seems to be more relevant that Blu vs. Streams.
I guess that's what happens when you compare Apples to Oppos...sorry I couldn't resist! :-)
Regarding taking down OTA broadcasting to free spectrum for more economically lucrative and useful applications...as a consumer of OTA content I would miss it, but I would not try to lobby politically for keeping it. I would be perfectly content if the networks and their broadcast affiliates went with online streaming. The one lost element would be, of course HDTV signal quality. At today's internet bandwidths, I don't see how they could keep the same near-Blu-ray quality.
If you're looking purely to test 1080p vs 720p then your best bet is Handbrake and playing the resulting files on an HTPC.
In Handbrake do two separate conversions using the quality indicator rather than file size indicator. Then, with all other settings remaining exactly the same, run one conversion at 1080p and one at 720p. That way any h264/conversion artifacting will affect both equally (except where resolution assists).
The end result should be the 1080p file being approximately 2.25 times as big as the 720p file, which means your per-pixel bitrate is equal between the two files. That will give a truly unbiased comparison of resolution differences.
I've done these tests on several movies (randomizing the file names so I wouldn't be subject to the placebo effect) and my wife and I could only tell the difference in rare instances - tests being performed on our (personally, amateurly calibrated) 58" Plasma at about 9-10 feet. In the end we decided to go with 720p files for our HTPC running XBMC since the minuscule difference when we were looking for it couldn't justify the 2.25x increase in file size.
I do use highly customized Handbrake settings, but the biggest thing is that I've set quality to '15'. The Handbrake forums typically talk about 18-20 (lower number = higher quality) claiming that they can't see the difference past that. Our own tests found the threshold to be around 15-16, but even at that 'overkill' level we can get movies shrunk down pretty small. Some have been as small as 2.5 GB, Avatar ended up being 10 GB; the great thing about the quality setting is that it uses higher bitrates when the content needs it but saves space when the content won't really be impacted.
For your tests, since file size isn't an issue at all, I'd use a setting of 10. The resulting 1080p file will probably be pretty close to original size, but it will have been subjected to the exact same processing as the 720p file.
In regards to OTA TV-
Lot's of work left to be done before we can ditch OTA tv.
I saw a study today that shows OTA tv viewing is UP to 15% of households.
http://thehill.com/blogs/hillicon-valley/technology/164911-over-the-air-tv-viewers-are-up
Which one of these studies is full of it?
I would have no problem getting my programming via broadband. A few problems-
-Bandwidth. The fastest connection available to me is 6 mbps. That's not going to cut it for HD live tv. TV quality would have to go down.
-Caps. I am only allowed to use 150 gbs per month by the fine folks at at&t. How many hours of HD TV per month is that (+ all the other things I need to use the internet for)?
Love the show. Thanks.