Add resolution and fps stats to histograms.

Add resolution and fps stats to histograms:
- "WebRTC.Video.InputWidthInPixels"
- "WebRTC.Video.InputHeightInPixels"
- "WebRTC.Video.SentWidthInPixels"
- "WebRTC.Video.SentHeightInPixels"
- "WebRTC.Video.ReceivedWidthInPixels"
- "WebRTC.Video.ReceivedHeightInPixels"
- "WebRTC.Video.RenderFramesPerSecond"

BUG=chromium:512752
Committed: https://crrev.com/d89920b74a173b7bf80c6760908a382c095a66cc
Cr-Commit-Position: refs/heads/master@{#9611}

[åsapersson, 2015-07-17 13:17:58]

Patchset #4 (id:60001) has been deleted

[åsapersson, 2015-07-17 13:18:05]

Patchset #3 (id:40001) has been deleted

[åsapersson, 2015-07-20 07:54:32]

Patchset #4 (id:100001) has been deleted

[åsapersson, 2015-07-20 08:05:42]

Patchset #4 (id:120001) has been deleted

[åsapersson, 2015-07-20 12:22:11]

asapersson@webrtc.org changed reviewers:
+ pbos@webrtc.org, stefan@webrtc.org

[pbos-webrtc, 2015-07-20 13:52:22]

overall looks good

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
File webrtc/video/receive_statistics_proxy.cc (right):

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
webrtc/video/receive_statistics_proxy.cc:46: int height =
render_height_counter_.Avg(kMinRequiredSamples);
Get both width and height at the same time, if width != -1 then you can assume
(or DCHECK, don't care) that height is not -1. But do:

int width = sample...;
int height = sample...;

if (width != -1) {
  histogram(width);
  histogram(height);
}

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
File webrtc/video/receive_statistics_proxy.h (right):

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
webrtc/video/receive_statistics_proxy.h:83: void Add(int sample) {
Move implementation-parts to .cc file.

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
webrtc/video/receive_statistics_proxy.h:91: private:
run git cl format

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
webrtc/video/receive_statistics_proxy.h:97: Clock* const clock_;
newline before, consider moving clock_ down to the rest

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/send_statis...
File webrtc/video/send_statistics_proxy.cc (right):

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/send_statis...
webrtc/video/send_statistics_proxy.cc:51: if (in_width != -1)
Same here, treat in_{width,height} as one pair, and do one check.

Then treat sent_{...} as one pair and do one check. This makes it easier to
read. :)

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/send_statis...
webrtc/video/send_statistics_proxy.cc:161: if (last_sent_frame_timestamp_ > 0 &&
Write a TODO() saying that this is incorrect if simulcast layers are encoded on
different threads and there's no guarantee (which afaik there isn't) that one
frame of all layers are encoded before the next start. Not sure if this happens
in practice, but it could make sense to keep a look out for >30fps stats
presence.

Frame 1 (320x180), Frame 1 (640x360), Frame 2 (320x180), Frame 1 (1280x720),
Frame 2 (640x480)...

In this case 720p of Frame 1 will look like a new frame, as will Frame 2 of VGA.

[åsapersson, 2015-07-20 15:06:30]

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
File webrtc/video/receive_statistics_proxy.cc (right):

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
webrtc/video/receive_statistics_proxy.cc:46: int height =
render_height_counter_.Avg(kMinRequiredSamples);
On 2015/07/20 13:52:22, pbos-webrtc wrote:
> Get both width and height at the same time, if width != -1 then you can assume
> (or DCHECK, don't care) that height is not -1. But do:
> 
> int width = sample...;
> int height = sample...;
> 
> if (width != -1) {
>   histogram(width);
>   histogram(height);
> }

Done.

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
File webrtc/video/receive_statistics_proxy.h (right):

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
webrtc/video/receive_statistics_proxy.h:83: void Add(int sample) {
On 2015/07/20 13:52:22, pbos-webrtc wrote:
> Move implementation-parts to .cc file.

Done.

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
webrtc/video/receive_statistics_proxy.h:91: private:
On 2015/07/20 13:52:22, pbos-webrtc wrote:
> run git cl format

Done.

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
webrtc/video/receive_statistics_proxy.h:97: Clock* const clock_;
On 2015/07/20 13:52:22, pbos-webrtc wrote:
> newline before, consider moving clock_ down to the rest

Done.

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/send_statis...
File webrtc/video/send_statistics_proxy.cc (right):

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/send_statis...
webrtc/video/send_statistics_proxy.cc:51: if (in_width != -1)
On 2015/07/20 13:52:22, pbos-webrtc wrote:
> Same here, treat in_{width,height} as one pair, and do one check.
> 
> Then treat sent_{...} as one pair and do one check. This makes it easier to
> read. :)

Done.

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/send_statis...
webrtc/video/send_statistics_proxy.cc:161: if (last_sent_frame_timestamp_ > 0 &&
On 2015/07/20 13:52:22, pbos-webrtc wrote:
> Write a TODO() saying that this is incorrect if simulcast layers are encoded
on
> different threads and there's no guarantee (which afaik there isn't) that one
> frame of all layers are encoded before the next start. Not sure if this
happens
> in practice, but it could make sense to keep a look out for >30fps stats
> presence.
> 
> Frame 1 (320x180), Frame 1 (640x360), Frame 2 (320x180), Frame 1 (1280x720),
> Frame 2 (640x480)...
> 
> In this case 720p of Frame 1 will look like a new frame, as will Frame 2 of
VGA.

Done, added comment.

[pbos-webrtc, 2015-07-20 15:11:07]

lgtm, add a tracking bug to BUG= if there is one

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
File webrtc/video/receive_statistics_proxy.h (right):

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
webrtc/video/receive_statistics_proxy.h:93: int num;
rename num -> num_samples

[åsapersson, 2015-07-21 06:05:47]

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
File webrtc/video/receive_statistics_proxy.h (right):

https://codereview.webrtc.org/1228393008/diff/140001/webrtc/video/receive_sta...
webrtc/video/receive_statistics_proxy.h:93: int num;
On 2015/07/20 15:11:07, pbos-webrtc wrote:
> rename num -> num_samples

Done.

[åsapersson, 2015-07-22 12:41:21]

The CQ bit was checked by asapersson@webrtc.org

[åsapersson, 2015-07-22 12:41:21]

The patchset sent to the CQ was uploaded after l-g-t-m from pbos@webrtc.org
Link to the patchset: https://codereview.webrtc.org/1228393008/#ps200001 (title:
"rebase")

[commit-bot: I haz the power, 2015-07-22 12:41:28]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1228393008/200001

[commit-bot: I haz the power, 2015-07-22 13:52:03]

Committed patchset #7 (id:200001)

[commit-bot: I haz the power, 2015-07-22 13:52:09]

Patchset 7 (id:??) landed as
https://crrev.com/d89920b74a173b7bf80c6760908a382c095a66cc
Cr-Commit-Position: refs/heads/master@{#9611}