Please select Into the mobile phone version | Continue to access the computer ver.
H.265 vs H.264
8828 27 2017-1-12 16:30:25
jimhare
Captain
Flight distance : 72858
Australia
Offline

Been seeing a lot of posts saying H.264 and H.265 look the same, the only difference is the H.265 file size is smaller.

This is driving me crazy because you want the file size to be the same!

H.265 at the same bitrate as H.264 will look TWICE as good in the detail.

It's very different to say they look the same but the file size is smaller, as to saying H.265 looks better at the same file size!

So someone who owns one please set me straight.  Can you not up the bitrate when shooting H.265?   

I thought the max bitrate was 100Mb/s across the board, is this not the case?
2017-1-12 16:30:25
Use props
juicedrummer
lvl.3
Flight distance : 8495
United States
Offline

i don't have super solid test to back this yet, plan on getting more this weekend, however, i have two short clips shot on different days with similar lengths and both transcoded from h.264/h.265 to ProRes 422.  These are their stats now:clip shot in h.264 at 1080p 24fps that is 30 seconds long, transcoded to prores 422 and the bit rate is
2017-1-12 17:01:59
Use props
juicedrummer
lvl.3
Flight distance : 8495
United States
Offline

Got a bit trigger happy with the enter key... Meant to post...

i don't have super solid test to back this yet, plan on getting more this weekend, however, i have two short clips shot on different days with similar lengths and both transcoded from h.264/h.265 to ProRes 422.  These are their stats now:
Clip 1 - h.264, 1080p, 24fps, 30 seconds long, transcoded to prores 422:
Size - 431.4 MB
Bit Rate - 113.81 Mbit/s

Clip 2 - h.265, 1080p, 24fps, 31 seconds long, transcoded to prores 422:
Size - 501.1 MB
Bit Rate - 130.94 Mbits/s

Again, this is just two random clips that happen to be similar length so take it for what its worth.  I hope to do a same scene same settings comparison this weekend with some sunny weather for the first time since ive had my P4P.

2017-1-12 17:06:44
Use props
Geebax
Captain
Australia
Offline

The maximum bitrate is 100 Mb/s, most likely in order to allow the system to record to commonly available SD cards. H.265 is a more recent version of MPEG encoding that allows encoding up to 8K in resolution, and it provides either superior quality encoding at the same bit rate or reduced file size at a given bit rate. It appears in the DJI implementation, they have opted for improved encoding at a given bit rate.

The only problem is, not all computer systems are yet able to handle h.265, not that it is vastly different, but they most likely reject it because they don't recognise the h.265 labelling. For this reason, DJI give you a choice of using it, or the more commonly recognised h.264.
2017-1-12 17:13:19
Use props
birdingbilly
lvl.3
Flight distance : 241542
United Kingdom
Offline

juicedrummer Posted at 2017-1-12 17:06
Got a bit trigger happy with the enter key... Meant to post...

i don't have super solid test to back this yet, plan on getting more this weekend, however, i have two short clips shot on different days with similar lengths and both transcoded from h.264/h.265 to ProRes 422.  These are their stats now:

It would be interesting to see the stats before the conversion to ProRes.  According to the DJI specs the P4P actually has a higher bitrate for 1080P on h264 than it does for h265 which I guess kind of makes sense(ish). For 4K the bitrate is the same.
2017-1-12 17:15:01
Use props
gabriel.magana
lvl.2
Flight distance : 43480
United States
Offline

@jimhare: This link explains how h265 files can be smaller given the same content (it's because of a different, better compression algorithm).
http://www.videoconverterfactory.com/tips/h265vsh264.html
2017-1-12 17:23:35
Use props
jimhare
Captain
Flight distance : 72858
Australia
Offline

Hey all, thanks but not really answering my question.  I understand what H.265 is, my question is can you max out H.265 at the same 100Mb/s as H.264, achieving the same file size but enjoy much greater quality due to the more efficient CODEC.

Most people seem to say "it looks the same but is a smaller file size" but the point is it should look much better at the SAME file size.
2017-1-12 23:21:12
Use props
jimhare
Captain
Flight distance : 72858
Australia
Offline

juicedrummer Posted at 2017-1-12 17:06
Got a bit trigger happy with the enter key... Meant to post...

i don't have super solid test to back this yet, plan on getting more this weekend, however, i have two short clips shot on different days with similar lengths and both transcoded from h.264/h.265 to ProRes 422.  These are their stats now:

Thanks Juicedrummer but doesn't this just show what the eventual ProRes file sizes are?
2017-1-12 23:21:56
Use props
fansc423b210
lvl.2
Flight distance : 51129
Australia
Offline

You cannot change the bitrate but I have checked the bitrate of a h.265 clip and it was at touch over the 100Mbps so I am assuming the same clip captured in h.264 would be approximately the same bitrate but lower quality than h.265. I capture most of my footage at 3840 x 2160, 24p using h.265, that way you are maximising the quality of each frame.
2017-1-13 00:36:08
Use props
Skyclip
lvl.4
Flight distance : 106052
  • >>>
Germany
Offline

Product Page of P4P states :
"More powerful video processing supports H.264 4K videos at 60fps or H.265 4K at 30fps, both with a 100Mbps bitrate."
and
"For a given bitrate, H.265 doubles the amount of image processing as H.264, resulting in significantly enhanced image quality. "

Recorded in 3840x2160 / 30fps and switched between h.264 and h.265, transcoded to prores422 on Mac.
I didnt see any difference.
2017-1-13 00:54:31
Use props
Geebax
Captain
Australia
Offline

jimhare Posted at 2017-1-12 23:21
Hey all, thanks but not really answering my question.  I understand what H.265 is, my question is can you max out H.265 at the same 100Mb/s as H.264, achieving the same file size but enjoy much greater quality due to the more efficient CODEC.

Most people seem to say "it looks the same but is a smaller file size" but the point is it should look much better at the SAME file size.

I highlighted the section in my post #4 where I said it provided superior quality at the same bitrate (as h.264).  But you cannot look at the footage and see an improvement in quality, because it does not manifest itself in that way. It is mostly what you are not seeing, and that is macro-blocking on extreme fine detail, compared to h.264.

With it being winter in the US, lots of material is being shot showing trees bare of leaves, with wispy branches, that get encoded as blobby clumps, because the amount of details exceeds tha ability of the encoder to portray them correctly. But the same footage encoded in h.265 would show less of the blobs, because the encoder is able to portray them more efficiently.

So rather than looking for higher quality, you should be looking for less coding artefacts.

2017-1-13 01:57:30
Use props
birdingbilly
lvl.3
Flight distance : 241542
United Kingdom
Offline

Geebax Posted at 2017-1-13 01:57
I highlighted the section in my post #4 where I said it provided superior quality at the same bitrate (as h.264).  But you cannot look at the footage and see an improvement in quality, because it does not manifest itself in that way. It is mostly what you are not seeing, and that is macro-blocking on extreme fine detail, compared to h.264.

With it being winter in the US, lots of material is being shot showing trees bare of leaves, with wispy branches, that get encoded as blobby clumps, because the amount of details exceeds tha ability of the encoder to portray them correctly. But the same footage encoded in h.265 would show less of the blobs, because the encoder is able to portray them more efficiently.

All well and good in theory but what I suspect the OP wants and no doubt othersas well  is actual facts about file sizes (for example) and perhaps some examples of footage that actually shows that h265 yields better results than h264.  Maybe someone has does this online somewhere but I havn't seen it - maybe online footage isn't capable of showing it ?
2017-1-13 02:13:11
Use props
juicedrummer
lvl.3
Flight distance : 8495
United States
Offline

jimhare Posted at 2017-1-12 23:21
Thanks Juicedrummer but doesn't this just show what the eventual ProRes file sizes are?

Yeah exactly.  As I mentioned, I still don't have a very solid test.  I am hoping to do a couple camera tests this weekend as we may get some sunshine and little wind for the first time since I've owned my drone.  I am hoping to take a few similar clips to compare file sizes and see how each holds up in my standard workflow through color correction and final export.
2017-1-13 07:56:50
Use props
Geebax
Captain
Australia
Offline

birdingbilly Posted at 2017-1-13 02:13
All well and good in theory but what I suspect the OP wants and no doubt othersas well  is actual facts about file sizes (for example) and perhaps some examples of footage that actually shows that h265 yields better results than h264.  Maybe someone has does this online somewhere but I havn't seen it - maybe online footage isn't capable of showing it ?

I think Jim made it clear in his post #7 that he wanted to know if it looked better. He did not seem concerned about file sizes.

The whole aim of MPEG encoding is really a magician's trick, sleight of hand. It aims to only encode as little as possible in full detail, and takes advantage of the fact that our eyes are quite poor at seeing fleeting details. So areas in the picture we are less likely to notice are not given the same detailed encoding. But all that goes out the window when you convert the pictures to another form where it can be easily examined frame by frame during an editing process. Other forms such as ProRes or AVI are much easier to edit with, but when MPEG is converted to those frorms, you can see all the tricks that MPEG was getting away with. With h.265, there are less of those tricks visible and therefore the subjective quality is higher.



2017-1-13 13:36:47
Use props
fans73888802
lvl.1
United States
Offline

Geebax Posted at 2017-1-13 13:36
I think Jim made it clear in his post #7 that he wanted to know if it looked better. He did not seem concerned about file sizes.

The whole aim of MPEG encoding is really a magician's trick, sleight of hand. It aims to only encode as little as possible in full detail, and takes advantage of the fact that our eyes are quite poor at seeing fleeting details. So areas in the picture we are less likely to notice are not given the same detailed encoding. But all that goes out the window when you convert the pictures to another form where it can be easily examined frame by frame during an editing process. Other forms such as ProRes or AVI are much easier to edit with, but when MPEG is converted to those frorms, you can see all the tricks that MPEG was getting away with. With h.265, there are less of those tricks visible and therefore the subjective quality is higher.

I wouldn't say that there are fewer tricks...just smarter ones.
2017-1-13 13:50:24
Use props
jimhare
Captain
Flight distance : 72858
Australia
Offline

Geebax Posted at 2017-1-13 01:57
I highlighted the section in my post #4 where I said it provided superior quality at the same bitrate (as h.264).  But you cannot look at the footage and see an improvement in quality, because it does not manifest itself in that way. It is mostly what you are not seeing, and that is macro-blocking on extreme fine detail, compared to h.264.

With it being winter in the US, lots of material is being shot showing trees bare of leaves, with wispy branches, that get encoded as blobby clumps, because the amount of details exceeds tha ability of the encoder to portray them correctly. But the same footage encoded in h.265 would show less of the blobs, because the encoder is able to portray them more efficiently.

Macroblocking and detail are the ONLY places I'm interested in.   I was pretty happy with the X3 other than the image falling apart when shooting detail.  Also was disappointed in the low light noise but I think the larger and improved sensor will help with that.
2017-1-13 17:39:08
Use props
jimhare
Captain
Flight distance : 72858
Australia
Offline

Geebax Posted at 2017-1-13 13:36
I think Jim made it clear in his post #7 that he wanted to know if it looked better. He did not seem concerned about file sizes.

The whole aim of MPEG encoding is really a magician's trick, sleight of hand. It aims to only encode as little as possible in full detail, and takes advantage of the fact that our eyes are quite poor at seeing fleeting details. So areas in the picture we are less likely to notice are not given the same detailed encoding. But all that goes out the window when you convert the pictures to another form where it can be easily examined frame by frame during an editing process. Other forms such as ProRes or AVI are much easier to edit with, but when MPEG is converted to those frorms, you can see all the tricks that MPEG was getting away with. With h.265, there are less of those tricks visible and therefore the subjective quality is higher.

Correct.  My post was related to all the reports of people saying the image didn't look better, was just a greatly reduced file size in H.265.  This didn't make sense because the goal would be to use the same bitrate and enjoy the additional quality, hence the file size should be the same.
2017-1-13 17:41:19
Use props
jimhare
Captain
Flight distance : 72858
Australia
Offline

juicedrummer Posted at 2017-1-13 07:56
Yeah exactly.  As I mentioned, I still don't have a very solid test.  I am hoping to do a couple camera tests this weekend as we may get some sunshine and little wind for the first time since I've owned my drone.  I am hoping to take a few similar clips to compare file sizes and see how each holds up in my standard workflow through color correction and final export.

All good.  Just want to make sure we're on the same page that the original H.264 and H.265 files should be roughly the same size before transcoding.   If they look the same then perhaps find more complex subject matter like running water and grassy areas.   Detail is where compression gets challenged.     
2017-1-13 17:43:02
Use props
Geebax
Captain
Australia
Offline

jimhare Posted at 2017-1-13 17:39
Macroblocking and detail are the ONLY places I'm interested in.   I was pretty happy with the X3 other than the image falling apart when shooting detail.  Also was disappointed in the low light noise but I think the larger and improved sensor will help with that.

I see the complaints on here from people who don't like the P4P+ images and I can't understand where they are coming from. My whole goal in obtaining a Phantom was to get good pictures, I have a P3, but would kill to have the higher quality camera the P4P+ offers. And the fact that DJI offered h.265 encoding was a big plus in my view. It is clear however, that quite a lot of people who do not like the images are not very familiar with camera systems in general.
2017-1-13 18:01:15
Use props
Airwolf13
lvl.3
Flight distance : 7234
United States
Offline

I tried shooting 265 today. My files would not open in any program tried. Can someone tell me why?
2017-1-13 21:47:12
Use props
Geebax
Captain
Australia
Offline

Airwolf13 Posted at 2017-1-13 21:47
I tried shooting 265 today. My files would not open in any program tried. Can someone tell me why?

I have already answered that question in post #4

'The only problem is, not all computer systems are yet able to handle h.265, not that it is vastly different, but they most likely reject it because they don't recognise the h.265 labelling. For this reason, DJI give you a choice of using it, or the more commonly recognised h.264.'


2017-1-13 22:20:16
Use props
Airwolf13
lvl.3
Flight distance : 7234
United States
Offline

Geebax Posted at 2017-1-13 22:20
I have already answered that question in post #4

'The only problem is, not all computer systems are yet able to handle h.265, not that it is vastly different, but they most likely reject it because they don't recognise the h.265 labelling. For this reason, DJI give you a choice of using it, or the more commonly recognised h.264.'

Has nothing to do with what computer I use (Mac)  NONE of my editing or viewing programs can open it.  Do you have a suggestion?
2017-1-14 03:35:15
Use props
Finnborg-Braga
lvl.2
Flight distance : 74539
United Kingdom
Offline

Airwolf13 Posted at 2017-1-14 03:35
Has nothing to do with what computer I use (Mac)  NONE of my editing or viewing programs can open it.  Do you have a suggestion?

Adobe Premiere CC 2017 with latest updates (=bug fixes) works fine on a PC, and I would expect the Mac version to be the same.

Before the latest update Premiere had a problem with some (but not all) P4P H.265 files, complaining about "no valid video streams".
2017-1-14 04:51:02
Use props
Jkeller84
First Officer
Flight distance : 120992
  • >>>
United States
Offline

Finnborg-Braga Posted at 2017-1-14 04:51
Adobe Premiere CC 2017 with latest updates (=bug fixes) works fine on a PC, and I would expect the Mac version to be the same.

Before the latest update Premiere had a problem with some (but not all) P4P H.265 files, complaining about "no valid video streams".

It works on a Mac also, but you have to use the .mp4 container.
2017-1-14 05:07:33
Use props
mxffiles
lvl.1
China
Offline

gabriel.magana Posted at 2017-1-12 17:23
@jimhare: This link explains how h265 files can be smaller given the same content (it's because of a different, better compression algorithm).
http://www.videoconverterfactory.com/tips/h265vsh264.html

H.265 has better compression performance and lower bandwidth utilization ratio.  at the same quality of coding, H.265 video can save around 50% of the bit rate, and the coding efficiency can be doubled than H. 264, as H.265 provides more different tools to reduce the bit rate. For example, in H. 264, each macroblock is fixed to the size of 16 x16 pixels, but the coding unit of H.265 can be chosen from a minimum of 8 x 8 to the maximum of 64 x 64.ts videos
2017-4-7 22:09:27
Use props
Bullflyer
lvl.4
Spain
Offline

Take a look at these  links:



2017-4-7 22:25:37
Use props
Dara Clark
Banned
United States
Offline

Tip: the author has been banned or deleted automatically shield
2017-4-21 01:39:21
Use props
Sebb
lvl.4
Flight distance : 49107
Germany
Offline

Dara Clark Posted at 2017-4-21 01:39
HEVC vs H.264: Compression Ratio     According to wiki and official test, what differs H.265 from H.264 is the compression efficiency. H.265 (aka HEVC) doubles the coding efficiency compared with its predecessor H.264. This means H.265 video saves around 50% of the bit rate at the same quality of coding. Specifically, the average bit reduction for H.265 is 64% at 4K UHD, 62% at 1080p, 56% at 720p and 52% at 480p. For instance, the H.265 replaces the 16x16 pixel macroblocks used with H.264 and uses larger block structures of up to 64x64 samples. Therefore, by reducing the design of flow rate, H.265 further lowers the cost of storage and transmission and delivers better visual quality, compared with H.264.     HEVC vs H.264: Bandwidth Use     H.265 is superior to H.264 regarding to the bandwidth usage. Because H.265's algorithm uses the efficient coding, H.265 promises approximately 40-50% of reduction in transmission bandwidth needed to compress the video (e.g. in 720p) over H.264 at the same video quality. So you can enjoy 4k video smoothly even at 1-2 Mbs bandwidth constrained network connection. On the contrary, the UHD videos encoded with h.264 will be watched at a stuttering mode.     HEVC vs H.264: Quality     The big difference between H.265 and H.264 lies in the video quality at the same bit rate. In H.264, the border areas of the block are likely to be distorted, because each macroblock is fixed and the data is independent to each other. While H.265 offers sharper detailed on faces and fabrics and smooth gradient areas with less blocking and fewer artifacts, since the new standard determines the size of code unit based on regional information. So H.265 is better than H.264 when it comes to compress a video with better image quality.     HEVC vs H.264: File Size     The great compression ratio also has a great relationship with the digital storage requirement of video streams and transmission. The reduced bandwidth leads to smaller file size. Test shows a video encoded with H.264 is almost 1-3 times larger than that with H.265. This is favorable for hard drive storage or the device with limited storage space required to house the video data.  HEVC vs H.264: Performance Comparison     H.265 vs H.264, which one is more suitable for you for videos playback? Of course, H.264 codec is applicable for almost all common devices. But things are not feasible to H.265. For devices that can decode HEVC video, performance is a concern. Here we tested the performance of various devices on average CPU usage, median clock speed and battery life for a reference of final decision of H.265 and H.264.[view_image]

Hi Dara,

that is great explanation, but what does it mean for P4P users? In practice, here is what I have observed:

like all codecs, h.264 visually breaks down at lower bitrates. Take 60Mbit (phantom 3/4/Mavic!) when filming certain scenes, e.g. flying very close to tree branches where full frame has fast moving detail, it looks like crap; also there is pulsing noticable on i-frames; also cloud covers get stuttery; list goes on, 60Mbit clearly not enough for 4K h.264 aerial footage.
h.265 is a codec that is specialized to counter issues resulting from overwhelming codecs, so at 60Mbit it would perform much better, most likely not have those issues and show all the performance gains you describe.

However, h.264 P4pro is 100Mbit. It does not have those issues nearly as much, it performs much better, much cleaner image, to a point where i am really happy with it.

Therefore, h.265 at 100Mbit on P4pro shows little to no difference in practice. To me, there really is no use case, it cannot materialize its strengths. Cannot fix whats not broken. and no, it will not magically show 2x the detail - h.264 already shows 100% of whats doable for the lens more or less.

On the flip side, h.265 complicates workflow significantly, since you need very modern PC to transcode it. Working directly on a timeline is very cumbersome, proxy footage takes long time etc.etc.

Also, h.265 compresses colors much more than h.264. That is actually not good for 8bit footage at all, which is already vulnerable to color noise, banding in sky in DLOG etc. etc. such issues are theoretically even worse in h.265.

But mostly: I have yet to see footage where h.265 visually outperforms h.264 P4pro footage. So from my experience and testing, h.265 is not worth it. Maybe someone can point me to such footage?

Cheers,
Sebb




2017-4-21 06:50:22
Use props
You need to log in before you can reply Login | Register now

Credit Rules

Close

DJI RecommendsPrevious /1 Next