\u00c7far\u00eb \u00ebsht\u00eb kompresimi pa humbje?<\/h3>\n
Kompresimi pa humbje zvog\u00eblon madh\u00ebsin\u00eb e skedarit duke identifikuar dhe eliminuar tepric\u00ebn statistikore pa hequr asnj\u00eb informacion. Kur dekompresohet, skedari \u00ebsht\u00eb pak p\u00ebr bit identik me origjinalin, pa asnj\u00eb humbje n\u00eb cil\u00ebsi ose integritet t\u00eb t\u00eb dh\u00ebnave.<\/p>\n <\/div>\n <\/div>\n \n
Si funksionon kompresimi pa humbje<\/h3>\n
Algoritmet e kompresimit pa humbje p\u00ebrdorin teknika t\u00eb ndryshme p\u00ebr t\u00eb reduktuar madh\u00ebsin\u00eb e skedarit duke siguruar rind\u00ebrtim t\u00eb p\u00ebrsosur t\u00eb t\u00eb dh\u00ebnave origjinale. K\u00ebto metoda analizojn\u00eb modelet, frekuencat dhe strukturat brenda t\u00eb dh\u00ebnave p\u00ebr t’i koduar ato n\u00eb m\u00ebnyr\u00eb m\u00eb efikase pa humbur informacion.<\/p>\n \n
Kodimi n\u00eb gjat\u00ebsin\u00eb e ekzekutimit (RLE)<\/h4>\n
RLE z\u00ebvend\u00ebson sekuencat e elementeve identike t\u00eb t\u00eb dh\u00ebnave (drejtimet) me nj\u00eb vler\u00eb dhe num\u00ebrim t\u00eb vet\u00ebm. P\u00ebr shembull, “AAAAAABBBCCCCC” b\u00ebhet “6A3B5C”, duke reduktuar ndjesh\u00ebm madh\u00ebsin\u00eb p\u00ebr t\u00eb dh\u00ebnat me shum\u00eb sekuenca t\u00eb p\u00ebrs\u00ebritura.<\/p>\n
Original: WWWWWWWWWWBBBWWWWWWWWWWWWBBBWWWWWWWWWW\nCompressed: 10W3B12W3B10W<\/pre>\n <\/div>\n <\/div>\n \n\n\n <\/i>\n <\/div>\nKodimi Huffman<\/h4>\n
Kjo teknik\u00eb cakton kode me gjat\u00ebsi t\u00eb ndryshueshme p\u00ebr karakteret hyr\u00ebse, me kode m\u00eb t\u00eb shkurtra p\u00ebr karaktere m\u00eb t\u00eb shpeshta. Kjo qasje statistikore optimizon kodimin bazuar n\u00eb shp\u00ebrndarjen e frekuenc\u00ebs s\u00eb karaktereve.<\/p>\n
\n Shembull:<\/strong>\nFrequent character 'e': 101\nLess frequent 'z': 1010101011<\/pre>\n <\/div>\n <\/div>\n \n\n\n <\/i>\n <\/div>\nAlgoritmet LZ77 & LZ78<\/h4>\n
K\u00ebto metoda t\u00eb bazuara n\u00eb fjalor z\u00ebvend\u00ebsojn\u00eb shfaqjet e p\u00ebrs\u00ebritura t\u00eb t\u00eb dh\u00ebnave me referenca n\u00eb nj\u00eb kopje t\u00eb vetme tashm\u00eb t\u00eb pranishme n\u00eb rrjedh\u00ebn e pakompresuar. Ato formojn\u00eb baz\u00ebn p\u00ebr formatet e njohura si ZIP dhe GIF.<\/p>\n
\n Shembull:<\/strong>\nInstead of storing \"compression compression\"\nStore \"compression [pointer to earlier instance]\"<\/pre>\n <\/div>\n <\/div>\n \n\n\n <\/i>\n <\/div>\nAlgoritmi i shfryrjes<\/h4>\n
Duke kombinuar kodimin LZ77 dhe Huffman, Deflate siguron kompresim t\u00eb shk\u00eblqyer me shpejt\u00ebsi t\u00eb mir\u00eb. P\u00ebrdoret n\u00eb kompresimin ZIP, PNG dhe HTTP (gzip), duke e b\u00ebr\u00eb at\u00eb nj\u00eb nga algoritmet m\u00eb t\u00eb p\u00ebrhapur.<\/p>\n
\n Aplikimet:<\/strong>\n\n
- Arkivat ZIP<\/li>\n
- Imazhet PNG<\/li>\n
- Kompresimi HTTP (gzip)<\/li>\n <\/ul>\n <\/div>\n <\/div>\n \n
\n\n <\/i>\n <\/div>\nKodimi aritmetik<\/h4>\n
Kjo teknik\u00eb p\u00ebrfaq\u00ebson nj\u00eb mesazh si nj\u00eb varg numrash midis 0 dhe 1. Ajo mund t\u00eb arrij\u00eb raporte kompresimi af\u00ebr kufirit teorik t\u00eb entropis\u00eb, duke e b\u00ebr\u00eb at\u00eb shum\u00eb efikase p\u00ebr lloje t\u00eb caktuara t\u00eb dh\u00ebnash.<\/p>\n
\n Avantazhi:<\/strong>\nMund t\u00eb kodoj\u00eb pjes\u00eb t\u00eb pjesshme p\u00ebr simbol, duke ofruar kompresim m\u00eb t\u00eb mir\u00eb se Huffman p\u00ebr shum\u00eb burime.<\/p>\n <\/div>\n <\/div>\n \n
\n\n <\/i>\n <\/div>\nKodimi Delta<\/h4>\n
N\u00eb vend t\u00eb ruajtjes s\u00eb vlerave absolute, kodimi delta ruan dallimet midis vlerave t\u00eb nj\u00ebpasnj\u00ebshme. Kjo \u00ebsht\u00eb ve\u00e7an\u00ebrisht efektive p\u00ebr t\u00eb dh\u00ebnat ku vlerat ngjitur jan\u00eb t\u00eb ngjashme, si mostrat audio ose leximet e sensor\u00ebve.<\/p>\n
\n Shembull:<\/strong>\nOriginal: 105, 107, 106, 110, 108\nDelta: 105, +2, -1, +4, -2<\/pre>\n <\/div>\n <\/div>\n <\/div>\n \nFormatet e zakonshme t\u00eb skedar\u00ebve pa humbje<\/h3>\n
\n\n\n <\/i>\nArkivat<\/h4>\n <\/div>\n
\nZIP<\/div>\nRAR<\/div>\n7Z<\/div>\nGZIP<\/div>\nBZIP2<\/div>\nTAR<\/div>\n <\/div>\n <\/div>\n \n\n\n <\/i>\nImazhet<\/h4>\n <\/div>\n
\nPNG<\/div>\nTIFF<\/div>\nBMP<\/div>\nGIF<\/div>\nWebP (pa humbje)<\/div>\n <\/div>\n <\/div>\n \n\n\n <\/i>\nAudio<\/h4>\n <\/div>\n
\nFLAC<\/div>\nALAC<\/div>\nWAV<\/div>\nAPE<\/div>\nWavPack<\/div>\n <\/div>\n <\/div>\n <\/div>\n <\/div>\n <\/section>\n \n \n\n Shpjegohet kompresimi me humbje<\/h2>\n \n
\n\n\n <\/i>\n <\/div>\n\n\u00c7far\u00eb \u00ebsht\u00eb kompresimi me humbje?<\/h3>\n
Kompresimi me humbje zvog\u00eblon madh\u00ebsin\u00eb e skedarit duke eliminuar p\u00ebrgjithmon\u00eb informacione t\u00eb caktuara, ve\u00e7an\u00ebrisht t\u00eb dh\u00ebna t\u00eb tep\u00ebrta ose m\u00eb pak t\u00eb r\u00ebnd\u00ebsishme n\u00eb perceptim. Skedari i dekompresuar \u00ebsht\u00eb i ndrysh\u00ebm nga origjinali, por ndryshimet jan\u00eb krijuar q\u00eb t\u00eb jen\u00eb t\u00eb v\u00ebshtira ose t\u00eb pamundura p\u00ebr t’u perceptuar nga njer\u00ebzit n\u00eb kushte normale.<\/p>\n <\/div>\n <\/div>\n \n
Si funksionon kompresimi me humbje<\/h3>\n
Kompresimi me humbje arrin raporte kompresimi duksh\u00ebm m\u00eb t\u00eb larta duke marr\u00eb vendime strategjike se cilat t\u00eb dh\u00ebna duhet t\u00eb hidhen. K\u00ebto algoritme shfryt\u00ebzojn\u00eb njohurit\u00eb rreth perceptimit njer\u00ebzor \u2013 at\u00eb q\u00eb syt\u00eb dhe vesh\u00ebt tan\u00eb mund dhe nuk mund t\u00eb zbulojn\u00eb \u2013 p\u00ebr t\u00eb hequr informacionin n\u00eb m\u00ebnyra q\u00eb minimizojn\u00eb ndikimin e duksh\u00ebm n\u00eb cil\u00ebsi.<\/p>\n \n
\n\n\n <\/i>\n <\/div>\nTransformoni kodimin<\/h4>\n
Kjo teknik\u00eb transformon t\u00eb dh\u00ebnat nga nj\u00eb domen (si hap\u00ebsinor) n\u00eb tjetrin (si frekuenca) ku kompresimi mund t\u00eb zbatohet n\u00eb m\u00ebnyr\u00eb m\u00eb efektive. Transformimi i kosinusit diskret (DCT) i p\u00ebrdorur n\u00eb JPEG \u00ebsht\u00eb nj\u00eb shembull kryesor.<\/p>\n
\n Procesi:<\/strong>\n\n
- Konvertoni blloqet e imazhit n\u00eb komponent\u00eb t\u00eb frekuenc\u00ebs<\/li>\n
- Kuantizoni komponent\u00ebt me frekuenc\u00eb t\u00eb lart\u00eb n\u00eb m\u00ebnyr\u00eb m\u00eb agresive<\/li>\n
- Syt\u00eb e njeriut jan\u00eb m\u00eb pak t\u00eb ndjesh\u00ebm ndaj k\u00ebtyre frekuencave<\/li>\n <\/ul>\n <\/div>\n <\/div>\n \n
\n\n <\/i>\n <\/div>\nKuantizimi<\/h4>\n
Kuantizimi zvog\u00eblon sakt\u00ebsin\u00eb e vlerave t\u00eb t\u00eb dh\u00ebnave. Ai harton nj\u00eb s\u00ebr\u00eb vlerash hyr\u00ebse n\u00eb nj\u00eb grup m\u00eb t\u00eb vog\u00ebl vlerash dal\u00ebse, duke reduktuar n\u00eb m\u00ebnyr\u00eb efektive numrin e biteve t\u00eb nevojsh\u00ebm p\u00ebr t\u00eb p\u00ebrfaq\u00ebsuar t\u00eb dh\u00ebnat.<\/p>\n
\n Shembull:<\/strong>\nOriginal values: 4.13, 4.28, 4.97, 4.02\nQuantized to: 4, 4, 5, 4<\/pre>\n <\/div>\n <\/div>\n \n\n\n <\/i>\n <\/div>\nModelimi Psikoakustik<\/h4>\n
E p\u00ebrdorur n\u00eb kompresimin audio, kjo teknik\u00eb shfryt\u00ebzon kufizimet e d\u00ebgjimit njer\u00ebzor. Ai identifikon se cil\u00ebt komponent\u00eb audio mund t\u00eb hiqen pa ndikuar n\u00eb cil\u00ebsin\u00eb e perceptuar t\u00eb z\u00ebrit.<\/p>\n
\n Konceptet kryesore:<\/strong>\n\n
- Maskimi d\u00ebgjimor: Tingujt m\u00eb t\u00eb fort\u00eb maskojn\u00eb tingujt m\u00eb t\u00eb qet\u00eb<\/li>\n
- Ndjeshm\u00ebria ndaj frekuenc\u00ebs: Njer\u00ebzit d\u00ebgjojn\u00eb m\u00eb s\u00eb miri frekuencat e intervalit t\u00eb mes\u00ebm<\/li>\n
- Maskimi i p\u00ebrkohsh\u00ebm: Tingujt mund t\u00eb maskojn\u00eb t\u00eb tjer\u00ebt q\u00eb ndodhin pak para\/pas<\/li>\n <\/ul>\n <\/div>\n <\/div>\n \n
\n\n <\/i>\n <\/div>\nKodimi perceptues<\/h4>\n
Ngjash\u00ebm me modelimin psikoakustik, por p\u00ebr t\u00eb dh\u00ebnat vizuale, kjo qasje heq informacionin q\u00eb syt\u00eb e njeriut kan\u00eb m\u00eb pak gjasa t\u00eb v\u00ebrejn\u00eb, ve\u00e7an\u00ebrisht n\u00eb detajet me frekuenc\u00eb t\u00eb lart\u00eb dhe variacionet e ngjyrave.<\/p>\n
\n Aplikimet:<\/strong>\nP\u00ebrdoret n\u00eb JPEG, MPEG dhe standarde t\u00eb tjera t\u00eb kompresimit vizual p\u00ebr t’i dh\u00ebn\u00eb p\u00ebrpar\u00ebsi t\u00eb dh\u00ebnave t\u00eb r\u00ebnd\u00ebsishme perceptuale.<\/p>\n <\/div>\n <\/div>\n \n
\n\n <\/i>\n <\/div>\nKompensimi i l\u00ebvizjes<\/h4>\n
Teknika e kompresimit t\u00eb videos q\u00eb shfryt\u00ebzon tepric\u00ebn e p\u00ebrkohshme duke koduar dallimet midis kornizave dhe jo \u00e7do kornize t\u00eb plot\u00eb. Vet\u00ebm ndryshimet nga nj\u00eb korniz\u00eb n\u00eb tjetr\u00ebn jan\u00eb plot\u00ebsisht t\u00eb koduara.<\/p>\n
\n Procesi:<\/strong>\n\n
- Ruani “kornizat kryesore” t\u00eb plota (I-korniza) n\u00eb m\u00ebnyr\u00eb periodike<\/li>\n
- P\u00ebr kornizat e tjera, ruani vet\u00ebm dallimet (kornizat P) ose dallimet dydrejtim\u00ebshe (kornizat B)<\/li>\n
- Rezulton n\u00eb zvog\u00eblimin dramatik t\u00eb madh\u00ebsis\u00eb s\u00eb skedarit p\u00ebr video<\/li>\n <\/ul>\n <\/div>\n <\/div>\n \n
\n\n <\/i>\n <\/div>\nN\u00ebnmostrimi i Chroma<\/h4>\n
Kjo teknik\u00eb zvog\u00eblon informacionin e ngjyrave m\u00eb shum\u00eb sesa informacionin e shk\u00eblqimit, duke p\u00ebrfituar nga ndjeshm\u00ebria m\u00eb e madhe e syrit t\u00eb njeriut ndaj shk\u00eblqimit sesa ndaj dallimeve t\u00eb ngjyrave.<\/p>\n
\n Format e zakonshme:<\/strong>\n\n
- 4:4:4 – Nuk ka n\u00ebn-mostrim (me ngjyr\u00eb t\u00eb plot\u00eb)<\/li>\n
- 4:2:2 – P\u00ebrgjysmon rezolucionin horizontal t\u00eb ngjyrave<\/li>\n
- 4:2:0 – P\u00ebrgjysmon rezolucionin horizontal dhe vertikal t\u00eb ngjyrave<\/li>\n <\/ul>\n <\/div>\n <\/div>\n <\/div>\n \n
Formatet e zakonshme t\u00eb skedar\u00ebve me humbje<\/h3>\n
\n\n\n <\/i>\nImazhet<\/h4>\n <\/div>\n
\nJPEG<\/div>\nWebP (me humbje)<\/div>\nJPEG 2000<\/div>\nHEIF<\/div>\nAVIF<\/div>\n <\/div>\n <\/div>\n \n\n\n <\/i>\nAudio<\/h4>\n <\/div>\n
\nMP3<\/div>\nAAC<\/div>\nVorbis<\/div>\nOpus<\/div>\nWMA<\/div>\n <\/div>\n <\/div>\n \n\n\n <\/i>\nVideo<\/h4>\n <\/div>\n
\nH.264<\/div>\nH.265<\/div>\nVP9<\/div>\nAV1<\/div>\nWebM<\/div>\n <\/div>\n <\/div>\n <\/div>\n <\/div>\n <\/section>\n \n \n\n Aplikime praktike dhe raste p\u00ebrdorimi<\/h2>\n \n
\n\n\n\n <\/i>\n <\/div>\nFotografi Dixhitale<\/h3>\n <\/div>\n
\nKompresim pa humbje<\/h4>\n
\n
- Ruajtja e formatit RAW p\u00ebr fotograf\u00ebt profesionist\u00eb<\/li>\n
- Ruajtja me cil\u00ebsi t\u00eb arkivit t\u00eb fotografive t\u00eb r\u00ebnd\u00ebsishme<\/li>\n
- Imazhet q\u00eb k\u00ebrkojn\u00eb p\u00ebrpunim ose redaktim t\u00eb gjer\u00eb<\/li>\n
- Format PNG p\u00ebr grafika me tekst ose skaje t\u00eb mprehta<\/li>\n <\/ul>\n \n
Kompresim me humbje<\/h4>\n
\n
- JPEG p\u00ebr fotografit\u00eb e p\u00ebrditshme dhe shp\u00ebrndarjen n\u00eb ueb<\/li>\n
- Gjenerimi i fotografive p\u00ebr galerit\u00eb dhe pamjet paraprake<\/li>\n
- Ngarkimet e mediave sociale ku zbatohen kufijt\u00eb e madh\u00ebsis\u00eb<\/li>\n
- Bashk\u00ebngjitjet e post\u00ebs elektronike dhe aplikacionet e mesazheve<\/li>\n <\/ul>\n \n
\n Praktika m\u00eb e mir\u00eb:<\/strong> Regjistroni n\u00eb formatin RAW ose pa humbje, krijoni versione me humbje p\u00ebr ndarje, mbani masterat pa humbje p\u00ebr arkivim.\n <\/div>\n <\/div>\n <\/div>\n \n\n\n\n <\/i>\n <\/div>\n