[{"data":1,"prerenderedAt":813},["ShallowReactive",2],{"project-43754124":3},{"id":4,"name":5,"line":6,"line_source":7,"title":8,"url":9,"domain":10,"maker":11,"posted_at":12,"year":13,"points":14,"comments":15,"rank_in_year":16,"shown_in_year":17,"top":18,"tier":19,"category":20,"subcategory":21,"format":22,"keywords":23,"reception":28,"sentiment":29,"stance_mix":30,"scored_comments":35,"aspects":36,"strip":50,"post_text":52,"what":53,"said":79,"maker_in_thread":241,"shape":252,"grid":255,"attempts":775,"same_week":776,"maker_shows":18,"maker_hidden":805,"reader_version":806,"card":807},"43754124","Dia","An open-weights TTS model for generating realistic dialogue","title","Show HN: Dia, an open-weights TTS model for generating realistic dialogue","https:\u002F\u002Fgithub.com\u002Fnari-labs\u002Fdia","github.com","toebee",1745255227,2025,652,190,45,3837,2,"legend","ai","text-to-speech model","library",[24,25,26,27],"text to speech","open weights","dialogue generation","voice cloning","mixed",0.263,{"positive":31,"neutral":32,"mixed":33,"critical":34},36,50,8,11,105,{"execution":37,"idea":40,"performance":42,"docs":45,"design":48,"privacy":49},{"positive":38,"critical":39},7,6,{"positive":38,"critical":41},0,{"positive":43,"critical":44},10,4,{"positive":46,"critical":47},3,1,{"positive":47,"critical":41},{"positive":41,"critical":41},[47,44,51,46,51,41,46,51,41,51,46,46,46,44,46,46,18,51,46,18,46,41,44,46,46,51,51,51,46,46,46,51,46,44,44,46,46,51,51,51],5,null,{"problem":54,"audience":52,"pricing":52,"motivation":61,"time_to_build":65,"stack":69,"openness":76,"stage":77,"team_size":78},{"value":55,"source":56,"evidence":57},"Generating realistic multi-speaker dialogue audio from a transcript","maker",[58],{"id":59,"by":11,"text":60},"43754125","Hey HN! We’re Toby and Jay, creators of Dia. Dia is 1.6B parameter open-weights model that generates dialogue directly from a transcript. Unlike TTS models that generate each…",{"value":62,"source":56,"evidence":63},"Wanted to replicate podcast-style conversation after falling for NotebookLM's podcast feature",[64],{"id":59,"by":11,"text":60},{"value":66,"source":56,"evidence":67},"A bit over 3 months",[68],{"id":59,"by":11,"text":60},[70],{"value":71,"source":56,"evidence":72},"Descript Audio Codec",[73],{"id":74,"by":11,"text":75},"43757469","We use descript audio codec! I’m not sure if DAC works on iOS…","open_source","launched","team",{"praise":80,"objections":108,"requests":141,"questions":174,"alternatives":207},[81,97],{"value":82,"source":83,"evidence":84},"Audio quality and expressiveness sound realistic and human","commenter",[85,89,93],{"id":86,"by":87,"text":88},"43756741","Havoc","Sounds really good & human! Got a fair bit of unexpected artifacts though. e.g. 3 seconds hissing noise before dialogue. And music in background when I added (happy) in an…",{"id":90,"by":91,"text":92},"43755279","xbmcuser","Wow first time I have felt that this could be the end of voice acting\u002Faudio book narration etc. The speed with with the ways things are changing how soon before you can make any…",{"id":94,"by":95,"text":96},"43759096","flashblaze","I'm lost for words. This is extremely impressive!",{"value":98,"source":83,"evidence":99},"Impressive for a two-person team with no prior speech experience",[100,104],{"id":101,"by":102,"text":103},"43757244","llm_nerd","This is a pretty incredible three month creation for a couple of people who had no experience with speech models.",{"id":105,"by":106,"text":107},"43756127","codingmoh","Hey, this is really cool! Curious how good the multi-language support is. Also - pretty wild that you trained the whole thing yourselves, especially without prior experience in…",[109,124,128,135],{"value":110,"source":83,"evidence":111},"English only",[112,116,120],{"id":113,"by":114,"text":115},"43755130","isoprophlex","Incredible quality demo samples, well done. How's the performance for multilingual generation?",{"id":117,"by":118,"text":119},"43754602","brumar","Impressive! Is it english only at the moment?",{"id":121,"by":122,"text":123},"43757107","hiAndrewQuinn","Is this English-only? I'm looking for a local model for Finnish dialogue to run.",{"value":125,"source":83,"evidence":126},"Outputs are unstable with artifacts such as hissing or background music",[127],{"id":86,"by":87,"text":88},{"value":129,"source":83,"evidence":130},"Requires about 10GB VRAM, too heavy for many GPUs",[131],{"id":132,"by":133,"text":134},"43754400","ivape","Darn, don't have the appropriate hardware. The full version of Dia requires around 10GB of VRAM to run. If you have a 16gb of VRAM, I guess you could pair this with a 3B param…",{"value":136,"source":56,"evidence":137},"Limited to two speakers",[138],{"id":139,"by":11,"text":140},"43758094","Only two voices at the moment... We will need to upgrade the dataset to make that happen, and are considering that as one of the next steps.",[142,154,160,167],{"value":143,"source":83,"evidence":144},"Multi-language support",[145,149,150],{"id":146,"by":147,"text":148},"43755174","999900000999","Does this only support English? I would absolutely love something like this for practicing Chinese, or even just adding Chinese dialogue to a project.",{"id":121,"by":122,"text":123},{"id":151,"by":152,"text":153},"43756302","bzuker","hey, this looks (or rather, sounds) amazing! Does it work with different languages or is it English only?",{"value":155,"source":56,"evidence":156},"Quantized version for lower VRAM",[157],{"id":158,"by":11,"text":159},"43754481","We will work on a quantized version of the model, so hopefully you will be able to run it soon! We've seen Bark from Suno go from 16GB requirement -> 4GB requirement + running on…",{"value":161,"source":83,"evidence":162},"Streaming output for lower latency",[163],{"id":164,"by":165,"text":166},"43754416","sarangzambare","Impressive demo! We'd love to use this at https:\u002F\u002Fuseponder.ai time to first audio is something that is crucial for us to reduce the latency - wondering if dia works with output…",{"value":168,"source":83,"evidence":169},"AMD GPU support",[170],{"id":171,"by":172,"text":173},"43757584","oehtXRwMkIs","Any plans for AMD GPU support? Maybe I'm missing something, but it's not working out of the box on a 7900xtx.",[175,189,200],{"value":176,"source":83,"evidence":177},"Where the training data came from",[178,182,186],{"id":179,"by":180,"text":181},"43757653","dangoodmanUT","I know it’s taboo to ask, but I must: where’s the dataset from? Very eager to play around with audio models myself, but I find existing datasets limiting",{"id":183,"by":184,"text":185},"43754758","gfaure","Amazing that you developed this over the course of three months! Can you drop any insight into how you pulled together the audio data?",{"id":187,"by":114,"text":188},"43755082","+1 to this, amazing how you managed to deliver this, and iff you're willing to share i'd be most interested in learning what you did in terms of train data..!",{"value":190,"source":83,"evidence":191},"How to keep a consistent voice across turns",[192,196],{"id":193,"by":194,"text":195},"43756773","xhkkffbf","Are there different voices? Or only [s1] and [s2] in the examples?",{"id":197,"by":198,"text":199},"43754550","xienze","How do you declare which voice should be used for a particular speaker? And can it created a cloned speaker voice from a sample?",{"value":201,"source":83,"evidence":202},"Licensing terms for research and commercial use",[203],{"id":204,"by":205,"text":206},"43755952","rustc","Is this Apache licensed or a custom one? The README contains this: > This project is licensed under the Apache License 2.0 - see the LICENSE file for details. > This project…",[208,219,230,237],{"value":209,"source":83,"evidence":210},"ElevenLabs",[211,215],{"id":212,"by":213,"text":214},"43754285","stuartjohnson12","Impressive project! We'd love to use something like this over at Delfa (https:\u002F\u002Fdelfa.ai). How does this hold up from the perspective of stability? I've spoken to various folks…",{"id":216,"by":217,"text":218},"43759361","LarsDu88","Fantastic model. I'm going to write a Unity plugin for this. Have been using ElevenLabs for my VR game side project, but this appears to be better",{"value":220,"source":83,"evidence":221},"Sesame CSM-1B",[222,226],{"id":223,"by":224,"text":225},"43754670","pzo","Sounds great. Hope more language support in the future. In comparison Sesame CSM-1B sounds like trained on stoned people.",{"id":227,"by":228,"text":229},"43756758","jokethrowaway","Looking forward to try. My current go-to solution is E5-F2 (great cloning, decent delivery, ok audio quality, a lot of incoherence here and there forcing you to do multiple…",{"value":231,"source":83,"evidence":232},"Spark TTS",[233],{"id":234,"by":235,"text":236},"43755207","verghese","How does this compare with Spark TTS? https:\u002F\u002Fgithub.com\u002FSparkAudio\u002FSpark-TTS",{"value":238,"source":56,"evidence":239},"Bark",[240],{"id":158,"by":11,"text":159},{"comments":242,"answered":243,"key":244},38,21,[245,248,251],{"id":246,"text":247},"43757397","Thank you for the kind words \u003C3",{"id":249,"text":250},"43757403","Thank you!! Indeed the script was inspired from a scene in the Office.",{"id":59,"text":60},{"commenters":253,"max_depth":38,"maker_replies":242,"first_comment_mins":41,"lifespan_hours":254},121,106,[256,260,263,267,269,273,276,280,283,287,290,293,297,299,302,305,309,312,315,318,320,323,326,329,332,335,338,340,342,345,347,350,353,355,358,362,366,369,373,376,379,382,384,387,390,393,396,399,403,404,408,410,413,416,418,419,423,425,427,430,432,434,435,437,439,441,445,447,448,452,454,457,460,463,465,467,469,472,475,478,481,484,487,490,492,495,496,499,502,505,507,511,515,517,520,522,525,527,530,532,535,538,541,543,546,549,554,555,557,558,562,564,567,569,572,574,577,580,584,587,589,592,594,596,599,601,603,605,609,611,615,618,620,622,625,627,629,632,634,636,640,641,643,645,647,648,650,652,654,658,662,664,668,672,674,676,678,682,684,686,688,690,692,695,696,699,702,706,708,710,714,716,718,720,722,726,730,733,737,739,743,746,749,751,753,755,759,763,767,771],[257,41,47,258,259,41],"43760276","project\u002Fcritical\u002Fexecution","sebstefan",[261,41,47,258,262,41],"43760727","throwaway-alpha",[264,41,46,265,266,41],"43763994","project\u002Fmixed\u002Fexecution","degosuke",[268,41,47,258,259,41],"43769514",[270,41,46,271,272,41],"43776952","project\u002Fneutral\u002Fexecution","yencabulator",[274,41,46,271,275,41],"43764526","xdfgh1112",[277,41,51,278,279,41],"43770558","project\u002Fpositive\u002Fexecution","yahoozoo",[281,41,46,271,282,41],"43761202","abrookewood",[284,41,46,285,286,41],"43761684","project\u002Fneutral\u002Fnone","bt1a",[288,41,51,278,289,41],"43761857","antiraza",[291,41,46,285,292,41],"43800464","cchance",[294,47,51,295,296,41],"43755481","project\u002Fpositive\u002Fidea","hemloc_io",[246,47,41,298,11,47],"maker\u002Fpositive\u002Fnone",[300,47,46,285,301,41],"43758495","kreelman",[303,47,46,285,304,41],"43759095","JonathanFly",[306,47,41,307,308,41],"43760210","other\u002Fneutral\u002Fpricing","miki123211",[310,47,41,311,224,41],"43760781","other\u002Fneutral\u002Fnone",[313,47,41,311,314,41],"43761084","bavell",[316,47,41,317,308,41],"43770320","other\u002Fneutral\u002Fidea",[319,47,41,311,292,41],"43800468",[321,18,51,295,322,41],"43754941","Versipelle",[324,18,41,317,325,41],"43755849","mclau157",[327,18,41,311,328,41],"43759666","hleszek",[330,18,41,311,331,41],"43755991","azinman2",[333,18,41,311,334,41],"43756515","senordevnyc",[336,18,41,311,337,41],"43756762","DrSiemer",[339,18,41,311,331,41],"43757590",[341,18,41,311,322,41],"43756799",[343,18,41,311,344,41],"43757678","ks2048",[346,18,41,311,292,41],"43758680",[348,18,51,295,349,41],"43759650","satvikpendem",[351,18,47,258,352,41],"43759961","iamsaitam",[354,18,41,311,349,41],"43762071",[356,18,41,311,357,41],"43773492","fennecfoxy",[359,46,51,360,361,41],"43755003","project\u002Fpositive\u002Fperformance","tyrauber",[363,46,46,364,365,41],"43755963","project\u002Fneutral\u002Fdocs","jinay",[367,46,51,278,368,41],"43756236","mrandish",[370,46,46,371,372,41],"43756357","project\u002Fmixed\u002Fperformance","DoctorOW",[374,46,41,311,375,41],"43757460","selimthegrim",[377,46,41,311,378,41],"43757575","wisemang",[380,46,41,311,381,41],"43760252","Cthulhu_",[383,46,41,311,372,41],"43777892",[385,46,41,311,386,41],"43760682","3by7",[388,46,51,360,389,41],"43761303","dostick",[391,46,51,360,392,41],"43763439","intalentive",[394,46,51,360,395,41],"43756347","nojs",[397,46,46,285,398,41],"43756563","fwip",[400,46,46,401,402,41],"43756931","project\u002Fneutral\u002Fidea","3abiton",[249,46,41,298,11,47],[405,46,46,406,407,41],"43763123","project\u002Fneutral\u002Fdesign","hombre_fatal",[59,44,41,409,11,47],"maker\u002Fneutral\u002Fidea",[411,44,46,371,412,41],"43754718","new_user_final",[414,44,41,415,11,47],"43758134","maker\u002Fneutral\u002Fperformance",[183,44,46,417,184,41],"maker\u002Fneutral\u002Fnone",[187,44,46,417,114,41],[420,44,46,421,422,41],"43755567","project\u002Fneutral\u002Fperformance","nickthegreek",[424,44,41,417,11,47],"43757964",[426,44,46,401,292,41],"43758674",[428,44,46,401,429,41],"43756264","heystefan",[431,44,41,409,11,47],"43758120",[433,44,41,311,349,41],"43759641",[151,44,46,285,152,41],[436,44,41,285,11,47],"43758132",[101,44,51,438,102,41],"maker\u002Fpositive\u002Fexecution",[440,44,41,298,11,47],"43758130",[442,44,46,443,444,41],"43757317","project\u002Fmixed\u002Fdesign","smusamashah",[446,44,41,285,11,47],"43758127",[179,44,46,285,180,41],[449,44,41,450,451,41],"43758141","commenter\u002Fneutral\u002Fnone","zelphirkalt",[453,44,41,450,180,41],"43758318",[455,44,41,450,456,41],"43758511","tough",[458,44,41,450,459,41],"43760208","deng",[461,44,41,450,462,41],"43784025","isaacfung",[464,44,41,450,459,41],"43796728",[466,44,41,450,357,41],"43773535",[468,44,46,285,275,41],"43765198",[470,44,46,285,471,41],"43758343","karimf",[473,44,47,474,292,41],"43758672","project\u002Fcritical\u002Fperformance",[476,44,46,421,477,41],"43760433","claiir",[479,44,46,401,480,41],"43768981","amp-lifier",[482,51,51,278,483,41],"43755044","notdian",[485,51,51,278,486,41],"43756713","noiv",[488,51,51,360,489,41],"43757852","rahimnathwani",[491,51,41,438,11,47],"43757958",[493,51,46,371,494,41],"43758272","emmelaich",[204,39,46,364,205,41],[497,39,46,364,498,41],"43757420","montroser",[500,39,41,501,11,47],"43757797","maker\u002Fneutral\u002Fdocs",[503,39,46,364,504,41],"43758300","crooked-v",[506,39,41,501,11,47],"43758694",[508,38,47,509,510,41],"43758604","project\u002Fcritical\u002Fdocs","moritonal",[512,38,46,513,514,41],"43758686","project\u002Fmixed\u002Fidea","kevmo314",[516,38,41,409,11,47],"43758745",[518,38,46,401,519,41],"43759382","kamranjon",[521,38,41,450,272,41],"43777132",[523,33,46,285,524,41],"43754236","strobe",[526,33,41,417,11,47],"43754343",[528,33,41,450,529,41],"43754781","aclark",[531,33,41,311,368,41],"43756248",[533,33,41,311,534,41],"43754991","Magma7404",[536,33,41,450,537,41],"43755121","teddyh",[539,33,41,450,540,41],"43756075","gapan",[542,33,41,311,537,41],"43764078",[544,33,41,311,545,41],"43755194","SoKamil",[547,33,41,311,548,41],"43755245","freedomben",[550,551,51,552,553,41],"43766084",9,"project\u002Fpositive\u002Fnone","devnen",[212,43,46,421,213,41],[556,43,41,417,11,47],"43754365",[86,34,46,371,87,41],[559,34,41,560,561,41],"43756851","commenter\u002Fneutral\u002Fdocs","yjftsjthsd-h",[563,34,41,560,87,41],"43756944",[565,34,41,560,566,41],"43764861","genewitch",[568,34,41,417,11,47],"43757435",[570,34,46,285,571,41],"43757925","dragonwriter",[90,573,51,295,91,41],12,[575,576,41,285,11,47],"43758012",13,[578,576,47,474,579,41],"43758641","daemonologist",[581,582,46,285,583,41],"43757134",14,"dindindin",[585,582,41,311,586,41],"43757307","jelling",[588,582,41,311,566,41],"43764866",[105,590,51,591,106,41],15,"maker\u002Fpositive\u002Fdocs",[593,590,41,285,11,47],"43758232",[595,590,51,591,314,41],"43761415",[597,598,41,417,11,47],"43754649",16,[164,600,51,360,165,41],17,[602,600,41,421,11,47],"43754471",[604,600,46,417,165,41],"43754500",[606,607,51,295,608,41],"43756200",18,"eob",[610,607,41,298,11,47],"43758082",[612,613,46,364,614,41],"43754347",19,"IshKebab",[616,613,41,560,617,41],"43754354","flakiness",[619,613,41,285,11,47],"43754404",[216,621,51,295,217,41],20,[623,243,46,401,624,41],"43755296","a2128",[626,243,41,364,11,47],"43758080",[146,628,46,285,147,41],22,[630,628,46,285,631,41],"43757745","buttercrab",[171,633,47,258,172,41],23,[635,633,41,421,11,47],"43758226",[637,638,46,401,639,41],"43755495",24,"lostmsu",[139,638,41,401,11,47],[113,642,51,360,114,41],25,[644,642,41,285,11,47],"43758258",[132,646,47,474,133,41],26,[158,646,41,421,11,47],[649,646,51,295,133,41],"43754581",[651,646,41,285,11,47],"43758083",[653,646,51,438,133,41],"43766019",[655,656,51,591,657,41],"43759155",27,"howon92",[659,660,46,371,661,41],"43757481",28,"enodios",[663,660,41,285,11,47],"43758076",[665,666,46,401,667,41],"43770004",29,"nicman23",[669,670,46,401,671,41],"43755275",30,"youssefabdelm",[673,670,41,364,11,47],"43758100",[675,670,46,401,671,41],"43762131",[234,677,46,285,235,41],31,[679,680,51,552,681,41],"43755366",32,"popalchemist",[197,683,46,364,198,41],33,[685,683,41,364,11,47],"43754643",[117,687,46,285,118,41],34,[689,687,41,285,11,47],"43754638",[223,691,51,552,224,41],35,[693,31,46,285,694,41],"43757314","instagary",[74,31,41,285,11,47],[697,698,47,474,180,41],"43757517",37,[700,242,51,701,486,41],"43756380","project\u002Fpositive\u002Fdesign",[703,704,47,258,705,41],"43756332",39,"vagabund",[707,704,41,271,11,47],"43757710",[709,704,41,271,11,47],"43758087",[711,712,51,552,713,41],"43758883",40,"user_4028b09",[193,715,46,401,194,41],41,[717,715,41,401,11,47],"43758096",[227,719,46,421,228,41],42,[721,719,41,285,11,47],"43758239",[723,724,41,311,725,41],"43762688",43,"basilgohar",[727,728,46,271,729,41],"43758412",44,"sroussey",[731,16,41,311,732,41],"43758257","qwertytyyuu",[734,735,51,552,736,41],"43754701",46,"film42",[94,738,51,552,95,41],47,[740,741,51,552,742,41],"43761576",48,"elia_42",[744,745,46,421,681,41],"43776875",49,[747,32,51,360,748,41],"43756118","zhyder",[750,32,41,417,11,47],"43758250",[752,32,51,360,304,41],"43758498",[121,754,46,285,122,41],51,[756,757,41,311,758,41],"43760254",52,"bazlan",[760,761,51,552,762,41],"43760015",53,"benterix",[764,765,51,360,766,41],"43761726",54,"jackchina",[768,769,46,401,770,41],"43760488",55,"gitroom",[772,773,46,774,325,41],"43755881",56,"project\u002Fneutral\u002Fprivacy",[],[777,782,787,792,796,801],{"id":778,"name":779,"points":780,"category":781},"43739456","Tutorial-Codebase-Knowledge",923,"dev_tools",{"id":783,"name":784,"points":785,"category":786},"43729683","The Backdooms",531,"games",{"id":788,"name":789,"points":790,"category":791},"43734910","Undercutf1",301,"consumer_social",{"id":793,"name":794,"points":795,"category":791},"43745554","JuryNow",263,{"id":797,"name":798,"points":799,"category":800},"43771365","Scrimba",258,"learning_science",{"id":802,"name":803,"points":804,"category":800},"43734583","Gridbach",244,false,"v2",{"id":808,"name":5,"line":6,"ai":805,"cat":20,"year":13,"points":14,"comments":15,"rec":28,"s":809,"top":18,"kw":810,"maker":11,"domain":10,"tier":19,"attempts":41,"strip":811,"url":9,"mix":812,"praise":82,"objection":110,"sub":21},43754124,0.26,[24,25,26,27],"1453503505333433253230433555333534433555",[31,32,33,34],1794036673231]