IRC log of images on 2019-09-18

Timestamps are in UTC.

01:45:10 [RRSAgent]
RRSAgent has joined #images
01:45:10 [RRSAgent]
logging to https://www.w3.org/2019/09/18-images-irc
01:45:12 [koalie]
RRSAgent, make logs public
01:45:15 [koalie]
koalie has changed the topic to: https://w3c.github.io/tpac-breakouts/sessions.html
01:45:18 [koalie]
koalie has left #images
01:59:48 [toshiakikoike]
toshiakikoike has joined #images
04:28:10 [toshiakikoike]
toshiakikoike has joined #images
05:31:08 [toshiakikoike]
toshiakikoike has joined #images
06:05:22 [shimazu]
shimazu has joined #images
06:32:05 [shimazu]
shimazu has joined #images
07:32:14 [smfr]
smfr has joined #images
07:34:13 [toshiakikoike]
toshiakikoike has joined #images
07:34:53 [shimazu]
shimazu has joined #images
07:36:44 [smfr]
smfr has joined #images
08:05:41 [kzms2]
kzms2 has joined #images
08:17:31 [hyojin]
hyojin has joined #images
08:27:34 [shimazu]
shimazu has joined #images
08:28:29 [smfr]
smfr has joined #images
08:30:44 [ChrisLilley]
ChrisLilley has joined #images
08:31:14 [ChrisLilley]
rrsagent, here
08:31:14 [RRSAgent]
See https://www.w3.org/2019/09/18-images-irc#T08-31-14
08:31:23 [ChrisLilley]
present+
08:31:58 [gkatsev]
gkatsev has joined #images
08:32:16 [zcorpan]
zcorpan has joined #images
08:32:39 [JakeA]
JakeA has joined #images
08:32:41 [zcorpan]
Meeting: Images
08:32:44 [zcorpan]
present+
08:32:45 [MasakazuKitahara]
MasakazuKitahara has joined #images
08:32:45 [smfr]
present+
08:32:46 [gkatsev]
present+
08:32:47 [zcorpan]
RRSAgent, make minutes
08:32:47 [RRSAgent]
I have made the request to generate https://www.w3.org/2019/09/18-images-minutes.html zcorpan
08:32:47 [mstange]
mstange has joined #images
08:32:48 [lizheming]
lizheming has joined #images
08:32:49 [JakeA]
present+
08:32:50 [mstange]
present+
08:32:54 [HeWenli]
HeWenli has joined #images
08:32:56 [domfarolino]
domfarolino has joined #images
08:32:57 [dlibby]
dlibby has joined #images
08:33:00 [zcorpan]
RRSAgent, make logs world-visible
08:33:04 [MasakazuKitahara]
present+
08:33:05 [zcorpan]
RRSAgent, make minutes
08:33:05 [RRSAgent]
I have made the request to generate https://www.w3.org/2019/09/18-images-minutes.html zcorpan
08:33:09 [toshiakikoike]
toshiakikoike has joined #images
08:33:12 [markw]
markw has joined #images
08:33:17 [domfarolino]
present+
08:33:17 [zcorpan]
scribenick: zcorpan
08:33:18 [markw]
present+
08:33:29 [bdekoz]
bdekoz has joined #images
08:33:39 [zcorpan]
[presentation]
08:34:13 [yoav]
yoav has joined #images
08:34:13 [annevk]
annevk has joined #images
08:34:13 [jib]
jib has joined #images
08:34:23 [yoav]
present+
08:34:30 [cyril]
cyril has joined #images
08:34:37 [bdekoz]
present+
08:34:56 [zcorpan]
cyril: container format used heavily for video, now proposed for images
08:35:00 [zcorpan]
cyril: HEIF
08:35:11 [bkardell_]
bkardell_ has joined #images
08:35:14 [zcorpan]
cyril: lots of images on a web page, on average 30 per page
08:35:18 [bkardell_]
present+
08:35:19 [zcorpan]
cyril: small in resolution
08:35:21 [zcorpan]
cyril: for icons
08:35:29 [jya]
jya has joined #images
08:35:30 [zcorpan]
cyril: constraints cna be very different compared to video
08:35:35 [zcorpan]
cyril: images need to be decoded very fast
08:35:38 [zcorpan]
cyril: footprint is a problem
08:35:45 [zcorpan]
cyril: small footprint for video decoder
08:35:51 [zcorpan]
cyril: gpu for video
08:36:06 [zcorpan]
cyril: goal is to try to understand the consequences for browsers, content providers, users
08:36:10 [zcorpan]
cyril: overlap and differences
08:36:20 [zcorpan]
cyril: not everyone is familiar, want to give more details
08:36:28 [hhan1]
hhan1 has joined #images
08:36:31 [zcorpan]
cyril: ISOBMFF
08:36:33 [ericc]
ericc has joined #images
08:36:37 [zcorpan]
cyril: "boxes"
08:36:47 [zcorpan]
cyril: MSE-type video "moov" "moof"
08:36:52 [zcorpan]
cyril: for images "meta" box
08:36:58 [zcorpan]
cyril: freely available
08:37:01 [zcorpan]
cyril: link to spec
08:37:27 [zcorpan]
cyril: more recent spec defines hwo to use the "meta" box to store props for th eimage
08:37:39 [zcorpan]
cyril: new type of track, "pict" for animations
08:37:43 [zcorpan]
cyril: how to store thumbnails
08:37:48 [zcorpan]
cyril: something channels
08:37:54 [zcorpan]
cyril: codec agnostic
08:38:11 [zcorpan]
cyril: annexes how to store HEVC, AVC, JPEG, but can use any codec
08:38:35 [markw]
ISO BMFF: https://standards.iso.org/ittf/PubliclyAvailableStandards/c068960_ISO_IEC_14496-12_2015.zip
08:38:41 [zcorpan]
pal: freely available? Rf?
08:38:48 [zcorpan]
cyril: the spec is freely available.
08:39:05 [zcorpan]
cyril: in terms of royalties, the spec has been available for 20 years...
08:39:13 [zcorpan]
pal: ISO patent database...
08:39:17 [zcorpan]
cyril: doesn't mean naything
08:39:20 [hhan2]
hhan2 has joined #images
08:39:34 [markw]
MPEG Image File Format: https://standards.iso.org/ittf/PubliclyAvailableStandards/c066067_ISO_IEC_23008-12_2017.zip
08:39:40 [zcorpan]
cyril: if someone impl it, and has been sued, that's relevant
08:40:04 [zcorpan]
cyril: large spec with lots of tech, not all of it is useful
08:40:15 [zcorpan]
cyril: MIAF
08:40:18 [pal]
pal has joined #images
08:40:25 [zcorpan]
cyril: meta box is required
08:40:29 [zcorpan]
cyril: for specifying base image
08:40:35 [zcorpan]
cyril: even if you have a sequence of images
08:40:44 [zcorpan]
cyril: e.g. middle image is representative, used for thumbnail
08:40:49 [zcorpan]
cyril: media profiles
08:40:56 [zcorpan]
cyril: brands, special profiles
08:41:01 [zcorpan]
cyril: progressively loaded
08:41:03 [zcorpan]
cyril: burst of images
08:41:10 [zcorpan]
cyril: spec is not freely downloadable
08:41:14 [zcorpan]
cyril: can pay to get it
08:41:23 [zcorpan]
cyril: AVIF
08:41:26 [zcorpan]
cyril: hosted on github
08:41:29 [zcorpan]
cyril: freely available
08:41:35 [zcorpan]
cyril: is based on MIAF
08:41:42 [zcorpan]
cyril: uses AV1 codec
08:41:49 [zcorpan]
cyril: transparency
08:41:53 [zcorpan]
cyril: two media profiles
08:41:57 [zcorpan]
cyril: baseline, advanced
08:42:02 [zcorpan]
cyril: main and high
08:42:11 [zcorpan]
cyril: reference software
08:42:16 [zcorpan]
cyril: C impl, portable
08:42:20 [zcorpan]
cyril: encode and decode
08:42:23 [zcorpan]
cyril: including alpha
08:42:39 [zcorpan]
cyril: the support, HEIF-based images, in browsers is ... currently limited
08:42:57 [zcorpan]
cyril: does apple support HEIF?
08:43:27 [zcorpan]
OS supports it but not web-exposed
08:43:57 [zcorpan]
JakeA: you can encode but not decode in Safari macOS
08:44:04 [zcorpan]
cyril: other browsers, dunno?
08:44:20 [zcorpan]
mstange: firefox doesn't support
08:44:36 [zcorpan]
cyril: chrome bug...
08:44:40 [zcorpan]
annevk: we're looking into it
08:44:53 [zcorpan]
annevk: open issues with AVIF
08:45:27 [zcorpan]
cyril: open questions... is it reasonbale to assume that if a video codec is supported, will the image codec be too?
08:45:31 [zcorpan]
annevk: once we support both, sure
08:45:33 [zcorpan]
LOL
08:45:45 [zcorpan]
cyril: don't want to use multiple decoders
08:46:01 [zcorpan]
??: path for gecko, between image and video totally different
08:46:08 [zcorpan]
??: but can change
08:46:22 [zcorpan]
mstange: in the past the answer has been no. for webp and webm
08:46:42 [zcorpan]
mstange: we don't want the web platform to rely on (something) image formats
08:46:52 [zcorpan]
yoav: it may be technically possible but there are other considerations
08:47:19 [zcorpan]
cyril: would media capabilities ?
08:47:24 [zcorpan]
mstange: i don't see why not
08:47:35 [zcorpan]
yoav: accept headers...
08:47:50 [zcorpan]
ChrisLilley: accept were expected to have that role, but failed
08:47:57 [zcorpan]
yoav: ok but sounds like a reasonable use case
08:48:14 [zcorpan]
YY: possible that the device can support both
08:48:21 [zcorpan]
YY: which should you choose
08:49:00 [zcorpan]
s/YY/markw/
08:49:01 [zcorpan]
s/YY/markw/
08:49:12 [zcorpan]
yoav: nowehere near that spec but sounds reasonable use case
08:49:31 [zcorpan]
cyril: AVIF format, supported in video
08:49:37 [shimazu]
shimazu has joined #images
08:49:39 [Yves]
Yves has joined #images
08:49:42 [zcorpan]
cyril: how likely is it to get uniform support
08:49:48 [zcorpan]
cyril: delta to achieve parsers
08:50:07 [zcorpan]
??: common demuxing and decoding video and images
08:50:09 [Yves]
https://tools.ietf.org/id/draft-ietf-httpbis-variants-00.html is relevant to the 'accept' header discussion
08:50:10 [Rochw]
Rochw has joined #images
08:50:16 [zcorpan]
annevk: for image loading there's some byte sniffing
08:50:32 [zcorpan]
annevk: some challenge
08:50:39 [zcorpan]
annevk: don't want to add new sniffing
08:50:57 [zcorpan]
annevk: strict mime types for new formats
08:51:09 [zcorpan]
yoav: strict mime types seems reasonable
08:51:13 [zcorpan]
annevk: more than reasoanble
08:51:26 [zcorpan]
ChrisLilley: do you say ??????
08:51:36 [zcorpan]
??: only to determine the container, not the codec
08:52:10 [zcorpan]
zcorpan: for top level navigation, sniffing to figure out type of document
08:52:18 [zcorpan]
cyril: the way HEIF define mime types
08:52:22 [zcorpan]
cyril: one for static images
08:52:26 [zcorpan]
cyril: one for image sequences
08:52:47 [zcorpan]
??: discussion a while back with w3c on media whether you listen to the server mime type
08:52:57 [zcorpan]
??: can only determine by sniffing, not by what the server says
08:53:01 [zcorpan]
yoav: servers lie
08:53:11 [zcorpan]
yoav: but if we add new stuff, we can break it
08:53:46 [zcorpan]
??: for firefox/safari (some format) , need sniffing
08:53:59 [zcorpan]
??: issue of trust
08:54:08 [zcorpan]
??: already don't trust mime type
08:54:13 [zcorpan]
yoav: we should trust and verify
08:54:20 [zcorpan]
yoav: if they lie, nothing runs
08:54:30 [zcorpan]
yoav: people should figure it out and fix it
08:55:02 [zcorpan]
annevk: we require the mime type to be correct, for webvtt, javascript modules...
08:55:12 [zcorpan]
??: video in particular
08:55:22 [zcorpan]
??: what makes major difference is the container
08:55:29 [zcorpan]
annevk: video element does a request
08:55:39 [zcorpan]
??: media source, you specify what you're going to be fed
08:55:45 [zcorpan]
annevk: you pass into an API
08:55:55 [zcorpan]
??: when you create your SourceBuffer
08:56:01 [zcorpan]
??: (missed)
08:56:07 [zcorpan]
yoav: the server mime type is irrelevant
08:56:29 [zcorpan]
JakeA: sniffing is fine if you have full access to (something?)
08:56:36 [jib]
s/??/jya/
08:56:40 [zcorpan]
annevk: sniffing i'm talking about if you navigate to a random url
08:56:52 [zcorpan]
annevk: or if we load cross-origin resources
08:56:58 [zcorpan]
annevk: safe-listing media resources
08:57:02 [zcorpan]
annevk: others we'd block
08:57:13 [zcorpan]
annevk: i'd like to restrict the sniffing. not keep adding
08:57:26 [zcorpan]
cyril: the first bytes of the file contains the type of file
08:57:34 [zcorpan]
cyril: the mime type tells you that
08:57:43 [zcorpan]
jya: video src = bla
08:57:48 [zcorpan]
yoav: server sends a mime type
08:57:55 [zcorpan]
yoav: it needs to correspond to the content
08:58:03 [zcorpan]
yoav: otherwise we need to add to the sniffing
08:58:05 [shimazu]
shimazu has joined #images
08:58:34 [zcorpan]
JakeA: if ... CORS, would just work
08:58:37 [zcorpan]
annevk: maybe
08:58:40 [igarashi_]
igarashi_ has joined #images
08:58:52 [zcorpan]
cyril: specific mime type
08:59:00 [zcorpan]
cyril: issue is "why 2 mime types"
08:59:16 [zcorpan]
cyril: either you change the spec, if the static mime type is given but the content is actually a sequence
08:59:21 [zcorpan]
cyril: rendered as the static
08:59:28 [zcorpan]
cyril: or get rid of the second mime type
08:59:35 [zcorpan]
ChrisLilley: given it has htumbnail
08:59:39 [zcorpan]
ChrisLilley: show static image
08:59:48 [zcorpan]
ChrisLilley: may want to do that deliberately
09:00:04 [zcorpan]
yoav: are there scenarios where static will be supported but sequence not?
09:00:13 [zcorpan]
yoav: if so we need 2 mime types
09:00:28 [zcorpan]
cyril: when i talk to media folks at google, they don't know
09:00:35 [zcorpan]
mstange: animations
09:00:44 [zcorpan]
jya: does the user want to be annoyed with sound
09:00:57 [zcorpan]
yoav: can be scenarios where people will have multiple ... dunno
09:01:24 [zcorpan]
cyril: no requirement in the spec to treat the mime types differently
09:01:41 [zcorpan]
annevk: problem is the mime type has no impact on processing model, will end up the same
09:01:48 [zcorpan]
yoav: <picture>, preload
09:01:57 [zcorpan]
yoav: if i specify a mime type that is not supported
09:02:02 [zcorpan]
yoav: it will be skipped
09:02:23 [zcorpan]
JakeA: if you specify jpeg but return png, it works
09:02:39 [zcorpan]
yoav: if you support only one, you want Accept to reflect that (hopefully), and type="" should work
09:02:56 [zcorpan]
annevk: why would you switch between those two mime types?
09:03:22 [zcorpan]
yoav: may want to fall back to other format
09:03:29 [zcorpan]
cyril: let's say you intend to support an api
09:03:35 [zcorpan]
cyril: would you support only static?
09:03:39 [zcorpan]
annevk: moz would support both
09:03:53 [zcorpan]
yoav: problem with webp
09:04:07 [zcorpan]
s/webp/apng/
09:04:36 [zcorpan]
ChrisLilley: PNG group said "you should register video/png" but they didn't want to do that
09:04:48 [zcorpan]
yoav: question: does anyone want to ship one but not the other?
09:05:10 [zcorpan]
ChrisLilley: use case for static or animated?
09:05:27 [zcorpan]
yoav: for goth type attr and accept header
09:05:32 [zcorpan]
yoav: if UA supports the format for both
09:05:35 [zcorpan]
yoav: same signal
09:05:46 [zcorpan]
yoav: no way for the web dev to distinguish based on mime type
09:05:53 [zcorpan]
cyril: third bullet point
09:06:07 [zcorpan]
cyril: consistent support between <img> and <video>
09:06:35 [zcorpan]
cyril: will we see ... will the overlap , unified model, convergance between the two elements?
09:06:40 [zcorpan]
yoav: wouldn't expect it
09:06:50 [zcorpan]
cyril: difference is subtle
09:07:18 [zcorpan]
yoav: there are multiple image formats that have animation, gif, webp
09:07:27 [zcorpan]
yoav: conv with our media team
09:07:35 [zcorpan]
yoav: needs convincing that the complexity is worth it
09:07:42 [zcorpan]
yoav: for video codec in <img>
09:07:47 [zcorpan]
yoav: if you have data
09:07:52 [zcorpan]
yoav: byte-savings e.g.
09:08:02 [zcorpan]
yoav: i can cummunicate it
09:08:30 [zcorpan]
yoav: ability to sandbox
09:08:43 [zcorpan]
cyril: safari team?
09:08:57 [zcorpan]
smfr: if you think about... yes
09:09:05 [zcorpan]
smfr: when we display an mp4 file
09:09:08 [zcorpan]
smfr: use hardward decoder
09:09:14 [zcorpan]
smfr: decode one frame at a time
09:09:19 [zcorpan]
smfr: we don't need to keep them all around
09:09:22 [zcorpan]
smfr: with animated gif
09:09:31 [zcorpan]
smfr: may not want to decode each frame every loop
09:09:38 [zcorpan]
smfr: savings
09:09:50 [zcorpan]
JakeA: comparison is <img src=video> vs <video muted>
09:10:14 [zcorpan]
smfr: can't use video as background image
09:10:26 [zcorpan]
yoav: problem of not everyone can control and modify their markup
09:10:32 [zcorpan]
yoav: i argued for that internally
09:10:41 [zcorpan]
ChrisLilley: ...lost?
09:10:46 [zcorpan]
yoav: well yes
09:10:52 [zcorpan]
yoav: animated webp is assumed to be enough
09:10:57 [zcorpan]
yoav: for background-image
09:11:02 [zcorpan]
yoav: lack of control for markup...
09:11:04 [JunGamo]
JunGamo has joined #images
09:11:08 [zcorpan]
yoav: otherwise <video loop muted> would work
09:11:15 [zcorpan]
smfr: arguing power and memory...
09:11:24 [zcorpan]
yoav: so... based on gpu decoding?
09:11:25 [zcorpan]
smfr: yes
09:11:41 [zcorpan]
yoav: argument was that images and animated gifs tend to come in lots vs video is only one
09:11:52 [zcorpan]
yoav: you could have multiple animated gifs
09:12:09 [zcorpan]
yoav: i will convey your skeptisism smfr
09:12:37 [smfr]
s/smfr/ericc/ ^
09:12:38 [zcorpan]
Richard: gpu or cpu, does it matter?
09:12:49 [zcorpan]
Richard: <img> tag, or if they merge
09:13:00 [zcorpan]
mstange: video where img is supported...
09:13:08 [zcorpan]
mstange: img is supported in many places
09:13:23 [zcorpan]
smfr: we support video everywhere
09:14:03 [zcorpan]
cyril: (missed)
09:14:05 [zcorpan]
yoav: bring data
09:14:11 [zcorpan]
annevk: why do you want that?
09:14:21 [zcorpan]
cyril: when only some browser support
09:14:30 [zcorpan]
cyril: same format, same codec
09:14:48 [zcorpan]
jya: video format was just so much better at compressing compared to gif
09:15:00 [zcorpan]
jya: you see on reddit, 100MB gif
09:15:06 [zcorpan]
jya: is it relevant
09:15:15 [zcorpan]
jya: reason is it compresses better
09:15:22 [zcorpan]
jya: now images compress well, is it relevant?
09:15:28 [zcorpan]
annevk: also special cases
09:15:44 [zcorpan]
jya: would you like your images to be (missed)
09:15:48 [zcorpan]
cyril: depends on format?
09:15:55 [zcorpan]
cyril: can follow image pipeline
09:16:02 [zcorpan]
jya: webp, use image decoder
09:16:06 [zcorpan]
jya: different objective
09:16:14 [zcorpan]
jya: want to be able to pause
09:16:19 [zcorpan]
jya: can't do that with video
09:16:29 [zcorpan]
jya: interrupts page loading
09:16:39 [zcorpan]
jya: purely technical
09:16:49 [zcorpan]
cyril: can still provide same decoding features
09:17:18 [zcorpan]
jya: ok if you have video that is gpu decoded, may not have same features in cpu
09:17:28 [zcorpan]
smfr: fundamental differences between img and video
09:17:32 [zcorpan]
smfr: color profiles different
09:17:38 [zcorpan]
smfr: not sure those would converge
09:17:52 [zcorpan]
smfr: might be a while before those converge
09:18:01 [zcorpan]
smfr: netflix is interested in mixing the two
09:18:02 [zcorpan]
cyril: yes
09:18:12 [zcorpan]
jya: if it were easy, everyone would have done it
09:18:15 [zcorpan]
cyril: some have done it
09:18:20 [zcorpan]
jya: one format
09:18:28 [zcorpan]
cyril: adjurn
09:18:33 [zcorpan]
rrsagent, make minutes
09:18:33 [RRSAgent]
I have made the request to generate https://www.w3.org/2019/09/18-images-minutes.html zcorpan
09:18:35 [JunGamo]
JunGamo has left #images
09:22:15 [ChrisLi18]
ChrisLi18 has joined #images
09:29:06 [ChrisLilley]
ChrisLilley has joined #images
09:50:41 [ChrisLilley]
ChrisLilley has joined #images
11:30:34 [pal]
pal has joined #images
12:13:05 [ChrisLilley]
ChrisLilley has joined #images
12:13:43 [ChrisLi63]
ChrisLi63 has joined #images
12:15:43 [ChrisLi32]
ChrisLi32 has joined #images
12:22:11 [ChrisLilley]
ChrisLilley has joined #images
12:35:10 [ChrisLi59]
ChrisLi59 has joined #images
15:16:26 [shimazu]
shimazu has joined #images
15:54:54 [shimazu]
shimazu has joined #images
16:28:50 [shimazu]
shimazu has joined #images
16:34:40 [ChrisLilley]
ChrisLilley has joined #images
17:37:05 [shimazu]
shimazu has joined #images