{"id":2191,"date":"2010-02-08T19:12:33","date_gmt":"2010-02-09T03:12:33","guid":{"rendered":"http:\/\/multimedia.cx\/eggs\/?p=2191"},"modified":"2020-07-25T23:52:30","modified_gmt":"2020-07-26T06:52:30","slug":"profiling-and-optimizing-theora","status":"publish","type":"post","link":"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/","title":{"rendered":"Profiling and Optimizing Theora"},"content":{"rendered":"<p>Because I have a short memory, I wanted to write down some of the knowledge and wisdom I&#8217;ve collected in the past few months as I have been working on optimizing <a href=\"http:\/\/ffmpeg.org\/\">FFmpeg&#8217;s<\/a> VP3\/Theora decoder.<\/p>\n<p><strong>Profiling Methods<\/strong><br \/>\nThese are some of the general tools:<br \/>\n<!--more--><\/p>\n<ul>\n<li>visual inspection via &#8216;top&#8217; and &#8216;time&#8217;: For interactive programs that are taking a lot of time to run, &#8216;top&#8217; can give a very coarse evaluation of CPU usage. For standalone programs that just need to crunch numbers to complete a task as quickly as possible, prepend the command with &#8216;time&#8217; to get a similarly coarse view of performance.<\/li>\n<li>oprofile: I like this tool a lot for profiling. Be advised that it will not work in a virtualized session.<\/li>\n<li>gprof: If you must. It gives you some decent information. The main point of the profilers is to sniff out general hotspots. For very fine-grained profiling&#8230;<\/li>\n<li>FFmpeg&#8217;s internal START_TIMER\/STOP_TIMER() macros: These tell you if your alleged optimizations are actually doing the trick. The results may surprise you.<\/li>\n<li>Remember to disable inlining for more accurate readings (-fno-inline-functions and -fno-inline-functions-called-once).<\/li>\n<\/ul>\n<p><strong>Recent Optimizations<\/strong><br \/>\nOptimizing in the multimedia domain often seems to come down to employing parallelizing SIMD instructions. However, most if not all of what can be SIMD-parallelized for VP3\/Theora has already been done for years in many cases. What I find interesting about these recent optimizations is that all have been at the algorithmic level:<\/p>\n<ul>\n<li>I <a href=\"http:\/\/multimedia.cx\/eggs\/one-gluttonous-if-statement\/\">turned a monotonous array traversal into a lean list traversal<\/a> for a giant improvement (no, I will <em>never<\/em> stop bragging about it).<\/li>\n<li>I <a href=\"http:\/\/multimedia.cx\/eggs\/optimizing-away-arrows\/\">removed a lot of structure dereference operations<\/a>. This is something that many observers state that the compiler should have managed automatically. But it didn&#8217;t and performing the same operation manually netted a performance improvement.<\/li>\n<li>Exploiting general knowledge of CPU memory caching, I moved a subsequent data operation (DC prediction reversal) to occur just after the data was decoded from the bitstream, thereby maximizing the probability that the coefficients will still be cached.<\/li>\n<li><a href=\"http:\/\/x264dev.multimedia.cx\/\">Dark Shikari<\/a> <a href=\"http:\/\/lists.mplayerhq.hu\/pipermail\/ffmpeg-devel\/attachments\/20091202\/0ca54b0a\/attachment.obj\">found some code<\/a> that looked like this:\n<pre>\r\nif (A && B)\r\n   { ... }\r\nif (A && C)\r\n   { ... }\r\nif (A && D)\r\n   { ... }\r\n<\/pre>\n<p>and factored out condition A to look like:<\/p>\n<pre>\r\n  if (A) {\r\n    if (B)\r\n      ...\r\n    if (C)\r\n      ...\r\n    if (D)\r\n      ...\r\n  }\r\n<\/pre>\n<p>thereby only evaluating condition A once per iteration. It may look minor but it makes a measurable difference.\n<\/li>\n<\/ul>\n<p><strong>Further Areas For Improvement<\/strong><br \/>\nThese are some more ideas for improving the performance of FFmpeg&#8217;s Theora decoder. It&#8217;s unlikely that I will have time or motivation to work on any of them in the near future. But I will be happy to consult (at least, for the ones I understand) if anyone else would like to take them on:<\/p>\n<ul>\n<li>Exploit the fact that many macroblocks move their 4 constituent luma blocks as a cohesive unit. This lends itself to using 16&#215;16 block functions rather than operating on 4 8&#215;8 blocks separately. This is actually tangentially related to SIMD optimization since the decoder can exploit SIMD functions to move 16 bytes at a time rather than 8. However, it will take some algorithmic rearrangement to determine when it&#8217;s safe to move a whole 16&#215;16 luma block.<\/li>\n<li>Study code around every continue statement. Is the code mindlessly iterating over another large (or even medium) array? If so, think about creating another list.<\/li>\n<li>Dark Shikari noticed that there are a lot of paranoid range checks in the code. This is no accident&#8211; I was extremely paranoid when I wrote the original code due to hard-won experience. I just wanted the code to <em>work right<\/em> at first. I think we&#8217;re well past that stage by now so it may be worth revisiting this code to see if any of those many checks are redundant or if they&#8217;re even necessary based on evaluating whether they can ever get into invalid states (accounting for the most maliciously invalid input).<\/li>\n<li>Implement partial IDCTs: The original VP3 decoder has special IDCTs to handle fragments where only the DC coefficient is non-zero or where only the first 3 or 10 coefficients are non-zero. A little printf reconnaissance to log coefficient count statistics clearly indicates that the 1-, 3- and 10-coefficient cases collectively occur far more frequently than the full 64-coefficient transform. The question is whether these transforms will necessarily translate to a notable optimization. This does involve SIMD. Thankfully, a lot of the work lives in the original VP3 source code which already sports (Win32) MMX and SSE2 variants for the 1-, 3-, 10-, and 64-coefficient IDCTs and (Mac) AltiVec versions of the 10- and 64-coefficient variations.<\/li>\n<li><a href=\"http:\/\/lists.mplayerhq.hu\/pipermail\/ffmpeg-devel\/2009-December\/079220.html\">Dark Shikari advised<\/a> that unpack_vectors() (which decodes the motion vector segment of a frame) requires an inordinate amount of CPU time and that it &#8220;can be made at least twice as fast.&#8221; No further details, but Monsieur Shikari generally has a keen eye for these kinds of hot spots. He just likes to speak in riddles, sort of.<\/li>\n<li>Decode coefficients in Hilbert order: <a href=\"http:\/\/multimedia.cx\/eggs\/theora-1-1-released\/#comment-150123\">David Conrad<\/a> has been advocating this approach for a long time and even has some partially working code to that effect. I admit that I still don&#8217;t understand what this approach entails. Dconrad reports that his prototype code demonstrates substantial speedup, even after my recent refactoring.<\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Make VP3\/Theora even faster in FFmpeg<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[33],"tags":[],"class_list":["post-2191","post","type-post","status-publish","format-standard","hentry","category-vp3theora"],"aioseo_notices":[],"aioseo_head":"\n\t\t<!-- All in One SEO 5.0.4 - aioseo.com -->\n\t<meta name=\"description\" content=\"Make VP3\/Theora even faster in FFmpeg\" \/>\n\t<meta name=\"robots\" content=\"max-image-preview:large\" \/>\n\t<meta name=\"author\" content=\"Multimedia Mike\"\/>\n\t<link rel=\"canonical\" href=\"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/\" \/>\n\t\t<meta name=\"generator\" content=\"All in One SEO (AIOSEO) 5.0.4\" \/>\n\t\t<meta property=\"og:locale\" content=\"en_US\" \/>\n\t\t<meta property=\"og:site_name\" content=\"Breaking Eggs And Making Omelettes | Topics On Multimedia Technology and Reverse Engineering\" \/>\n\t\t<meta property=\"og:type\" content=\"article\" \/>\n\t\t<meta property=\"og:title\" content=\"Profiling and Optimizing Theora | Breaking Eggs And Making Omelettes\" \/>\n\t\t<meta property=\"og:description\" content=\"Make VP3\/Theora even faster in FFmpeg\" \/>\n\t\t<meta property=\"og:url\" content=\"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/\" \/>\n\t\t<meta property=\"article:published_time\" content=\"2010-02-09T03:12:33+00:00\" \/>\n\t\t<meta property=\"article:modified_time\" content=\"2020-07-26T06:52:30+00:00\" \/>\n\t\t<meta name=\"twitter:card\" content=\"summary\" \/>\n\t\t<meta name=\"twitter:title\" content=\"Profiling and Optimizing Theora | Breaking Eggs And Making Omelettes\" \/>\n\t\t<meta name=\"twitter:description\" content=\"Make VP3\/Theora even faster in FFmpeg\" \/>\n\t\t<script type=\"application\/ld+json\" class=\"aioseo-schema\">\n\t\t\t{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/profiling-and-optimizing-theora\\\/#article\",\"name\":\"Profiling and Optimizing Theora | Breaking Eggs And Making Omelettes\",\"headline\":\"Profiling and Optimizing Theora\",\"author\":{\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/author\\\/multimedia-mike\\\/#author\"},\"publisher\":{\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/#organization\"},\"datePublished\":\"2010-02-08T19:12:33-08:00\",\"dateModified\":\"2020-07-25T23:52:30-07:00\",\"inLanguage\":\"en-US\",\"commentCount\":3,\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/profiling-and-optimizing-theora\\\/#webpage\"},\"isPartOf\":{\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/profiling-and-optimizing-theora\\\/#webpage\"},\"articleSection\":\"VP3\\\/Theora\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/profiling-and-optimizing-theora\\\/#breadcrumblist\",\"itemListElement\":[{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/#listItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/\",\"nextItem\":{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/category\\\/vp3theora\\\/#listItem\",\"name\":\"VP3\\\/Theora\"}},{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/category\\\/vp3theora\\\/#listItem\",\"position\":2,\"name\":\"VP3\\\/Theora\",\"item\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/category\\\/vp3theora\\\/\",\"nextItem\":{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/profiling-and-optimizing-theora\\\/#listItem\",\"name\":\"Profiling and Optimizing Theora\"},\"previousItem\":{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/#listItem\",\"name\":\"Home\"}},{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/profiling-and-optimizing-theora\\\/#listItem\",\"position\":3,\"name\":\"Profiling and Optimizing Theora\",\"previousItem\":{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/category\\\/vp3theora\\\/#listItem\",\"name\":\"VP3\\\/Theora\"}}]},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/#organization\",\"name\":\"Breaking Eggs And Making Omelettes\",\"description\":\"Topics On Multimedia Technology and Reverse Engineering\",\"url\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/\"},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/author\\\/multimedia-mike\\\/#author\",\"url\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/author\\\/multimedia-mike\\\/\",\"name\":\"Multimedia Mike\",\"image\":{\"@type\":\"ImageObject\",\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/profiling-and-optimizing-theora\\\/#authorImage\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/2eb93f7cada658f7986b5092062849fa255779eada3342c8f6b9a40764b95fdd?s=96&d=identicon&r=g\",\"width\":96,\"height\":96,\"caption\":\"Multimedia Mike\"}},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/profiling-and-optimizing-theora\\\/#webpage\",\"url\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/profiling-and-optimizing-theora\\\/\",\"name\":\"Profiling and Optimizing Theora | Breaking Eggs And Making Omelettes\",\"description\":\"Make VP3\\\/Theora even faster in FFmpeg\",\"inLanguage\":\"en-US\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/#website\"},\"breadcrumb\":{\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/profiling-and-optimizing-theora\\\/#breadcrumblist\"},\"author\":{\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/author\\\/multimedia-mike\\\/#author\"},\"creator\":{\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/author\\\/multimedia-mike\\\/#author\"},\"datePublished\":\"2010-02-08T19:12:33-08:00\",\"dateModified\":\"2020-07-25T23:52:30-07:00\"},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/#website\",\"url\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/\",\"name\":\"Breaking Eggs And Making Omelettes\",\"description\":\"Topics On Multimedia Technology and Reverse Engineering\",\"inLanguage\":\"en-US\",\"publisher\":{\"@id\":\"https:\\\/\\\/multimedia.cx\\\/eggs\\\/#organization\"}}]}\n\t\t<\/script>\n\t\t<!-- All in One SEO -->\n\n","aioseo_head_json":{"title":"Profiling and Optimizing Theora | Breaking Eggs And Making Omelettes","description":"Make VP3\/Theora even faster in FFmpeg","canonical_url":"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/#article","name":"Profiling and Optimizing Theora | Breaking Eggs And Making Omelettes","headline":"Profiling and Optimizing Theora","author":{"@id":"https:\/\/multimedia.cx\/eggs\/author\/multimedia-mike\/#author"},"publisher":{"@id":"https:\/\/multimedia.cx\/eggs\/#organization"},"datePublished":"2010-02-08T19:12:33-08:00","dateModified":"2020-07-25T23:52:30-07:00","inLanguage":"en-US","commentCount":3,"mainEntityOfPage":{"@id":"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/#webpage"},"isPartOf":{"@id":"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/#webpage"},"articleSection":"VP3\/Theora"},{"@type":"BreadcrumbList","@id":"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/#breadcrumblist","itemListElement":[{"@type":"ListItem","@id":"https:\/\/multimedia.cx\/eggs\/#listItem","position":1,"name":"Home","item":"https:\/\/multimedia.cx\/eggs\/","nextItem":{"@type":"ListItem","@id":"https:\/\/multimedia.cx\/eggs\/category\/vp3theora\/#listItem","name":"VP3\/Theora"}},{"@type":"ListItem","@id":"https:\/\/multimedia.cx\/eggs\/category\/vp3theora\/#listItem","position":2,"name":"VP3\/Theora","item":"https:\/\/multimedia.cx\/eggs\/category\/vp3theora\/","nextItem":{"@type":"ListItem","@id":"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/#listItem","name":"Profiling and Optimizing Theora"},"previousItem":{"@type":"ListItem","@id":"https:\/\/multimedia.cx\/eggs\/#listItem","name":"Home"}},{"@type":"ListItem","@id":"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/#listItem","position":3,"name":"Profiling and Optimizing Theora","previousItem":{"@type":"ListItem","@id":"https:\/\/multimedia.cx\/eggs\/category\/vp3theora\/#listItem","name":"VP3\/Theora"}}]},{"@type":"Organization","@id":"https:\/\/multimedia.cx\/eggs\/#organization","name":"Breaking Eggs And Making Omelettes","description":"Topics On Multimedia Technology and Reverse Engineering","url":"https:\/\/multimedia.cx\/eggs\/"},{"@type":"Person","@id":"https:\/\/multimedia.cx\/eggs\/author\/multimedia-mike\/#author","url":"https:\/\/multimedia.cx\/eggs\/author\/multimedia-mike\/","name":"Multimedia Mike","image":{"@type":"ImageObject","@id":"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/#authorImage","url":"https:\/\/secure.gravatar.com\/avatar\/2eb93f7cada658f7986b5092062849fa255779eada3342c8f6b9a40764b95fdd?s=96&d=identicon&r=g","width":96,"height":96,"caption":"Multimedia Mike"}},{"@type":"WebPage","@id":"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/#webpage","url":"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/","name":"Profiling and Optimizing Theora | Breaking Eggs And Making Omelettes","description":"Make VP3\/Theora even faster in FFmpeg","inLanguage":"en-US","isPartOf":{"@id":"https:\/\/multimedia.cx\/eggs\/#website"},"breadcrumb":{"@id":"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/#breadcrumblist"},"author":{"@id":"https:\/\/multimedia.cx\/eggs\/author\/multimedia-mike\/#author"},"creator":{"@id":"https:\/\/multimedia.cx\/eggs\/author\/multimedia-mike\/#author"},"datePublished":"2010-02-08T19:12:33-08:00","dateModified":"2020-07-25T23:52:30-07:00"},{"@type":"WebSite","@id":"https:\/\/multimedia.cx\/eggs\/#website","url":"https:\/\/multimedia.cx\/eggs\/","name":"Breaking Eggs And Making Omelettes","description":"Topics On Multimedia Technology and Reverse Engineering","inLanguage":"en-US","publisher":{"@id":"https:\/\/multimedia.cx\/eggs\/#organization"}}]},"og:locale":"en_US","og:site_name":"Breaking Eggs And Making Omelettes | Topics On Multimedia Technology and Reverse Engineering","og:type":"article","og:title":"Profiling and Optimizing Theora | Breaking Eggs And Making Omelettes","og:description":"Make VP3\/Theora even faster in FFmpeg","og:url":"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/","article:published_time":"2010-02-09T03:12:33+00:00","article:modified_time":"2020-07-26T06:52:30+00:00","twitter:card":"summary","twitter:title":"Profiling and Optimizing Theora | Breaking Eggs And Making Omelettes","twitter:description":"Make VP3\/Theora even faster in FFmpeg"},"aioseo_meta_data":{"post_id":"2191","title":null,"description":null,"keywords":null,"keyphrases":null,"primary_term":null,"canonical_url":null,"og_title":null,"og_description":null,"og_object_type":"default","og_image_type":"default","og_image_url":null,"og_image_width":null,"og_image_height":null,"og_image_custom_url":null,"og_image_custom_fields":null,"og_video":null,"og_custom_url":null,"og_article_section":null,"og_article_tags":null,"twitter_use_og":false,"twitter_card":"default","twitter_image_type":"default","twitter_image_url":null,"twitter_image_custom_url":null,"twitter_image_custom_fields":null,"twitter_title":null,"twitter_description":null,"schema":{"blockGraphs":[],"customGraphs":[],"default":{"data":{"Article":[],"Course":[],"Dataset":[],"FAQPage":[],"Movie":[],"Person":[],"Product":[],"ProductReview":[],"Car":[],"Recipe":[],"Service":[],"SoftwareApplication":[],"WebPage":[]},"graphName":"","isEnabled":true},"graphs":[]},"schema_type":null,"schema_type_options":null,"pillar_content":false,"robots_default":true,"robots_noindex":false,"robots_noarchive":false,"robots_nosnippet":false,"robots_nofollow":false,"robots_noimageindex":false,"robots_noodp":false,"robots_notranslate":false,"robots_max_snippet":null,"robots_max_videopreview":null,"robots_max_imagepreview":"large","priority":null,"frequency":null,"location":null,"local_seo":null,"breadcrumb_settings":null,"limit_modified_date":false,"ai":null,"created":"2020-12-31 21:59:32","updated":"2025-12-31 19:21:07","seo_analyzer_scan_date":null,"focus_keyword":null,"additional_keywords":null,"truseo_locale":null},"aioseo_breadcrumb":"<div class=\"aioseo-breadcrumbs\"><span class=\"aioseo-breadcrumb\">\n\t\t\t<a href=\"https:\/\/multimedia.cx\/eggs\/\" title=\"Home\">Home<\/a>\n\t\t<\/span><span class=\"aioseo-breadcrumb-separator\">\u00bb<\/span><span class=\"aioseo-breadcrumb\">\n\t\t\t<a href=\"https:\/\/multimedia.cx\/eggs\/category\/vp3theora\/\" title=\"VP3\/Theora\">VP3\/Theora<\/a>\n\t\t<\/span><span class=\"aioseo-breadcrumb-separator\">\u00bb<\/span><span class=\"aioseo-breadcrumb\">\n\t\t\tProfiling and Optimizing Theora\n\t\t<\/span><\/div>","aioseo_breadcrumb_json":[{"label":"Home","link":"https:\/\/multimedia.cx\/eggs\/"},{"label":"VP3\/Theora","link":"https:\/\/multimedia.cx\/eggs\/category\/vp3theora\/"},{"label":"Profiling and Optimizing Theora","link":"https:\/\/multimedia.cx\/eggs\/profiling-and-optimizing-theora\/"}],"_links":{"self":[{"href":"https:\/\/multimedia.cx\/eggs\/wp-json\/wp\/v2\/posts\/2191","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/multimedia.cx\/eggs\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/multimedia.cx\/eggs\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/multimedia.cx\/eggs\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/multimedia.cx\/eggs\/wp-json\/wp\/v2\/comments?post=2191"}],"version-history":[{"count":13,"href":"https:\/\/multimedia.cx\/eggs\/wp-json\/wp\/v2\/posts\/2191\/revisions"}],"predecessor-version":[{"id":4645,"href":"https:\/\/multimedia.cx\/eggs\/wp-json\/wp\/v2\/posts\/2191\/revisions\/4645"}],"wp:attachment":[{"href":"https:\/\/multimedia.cx\/eggs\/wp-json\/wp\/v2\/media?parent=2191"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/multimedia.cx\/eggs\/wp-json\/wp\/v2\/categories?post=2191"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/multimedia.cx\/eggs\/wp-json\/wp\/v2\/tags?post=2191"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}