Monday, December 5, 2011

Woeing Wails of Whales

Baru lps present demo dkt pelawat td. Fuhhh legaaaa. Mmg bbrp hari sblm ni dah stop kerja2 lab and fokus adalah tertumpu di situ. Sbb nya, pelawat bukan calang2 pelawat, tp x lain x bukan drpd director company contemporary dance terkemuka, ‘Rambert’ yg based in London, tp kini touring di Plymouth utk show di The*atre Ro*yal. Siap SV aku pesan dalam emel nya lg, “There’s a chance for collaboration. We need to impress”. Mana x panic. Walaupon aku x rasa depa nak collaborate dgn keja aku yg agak dhaif ni, aku x nak la jd penyebab diorg elok2 nak wat collaboration, sekali tgk keja aku terus kansel niat suci murni tu plak.

Apapon, Alhamdulillah, I think the demo went well. I guess this time I invested more time and energy for this so-called presentation sbb nya I wanted to make up for a sucky one I did in October. Nyata lah, sebenar nya kalo kita buat sesuatu tu dgn lebih teliti dan budget masa yg secukupnya (well-prepared la kira nya ni), hasil nya jauh lebih baik drpd buat kerja dalam keadaan terkocoh2. Walaupon still ada jg moments ala2 last minute td, tp it was extra2 stuff that came into my head as the day went, but the main part of the presentation dah siap awal2 lg, so was not as panicky as last time.

Walau rasa ralat sgt hari tu sbb present mcm hampeh, but I have to admit the wake up call did me good. I NEEDED it. Because I didn’t get to deliver the punch line of my PhD during last seminar (my own fault as I was too frazzled trying to get latest results in without trying to digest and analyze them first – mmg mengundang utk kena fry aje lah kan), it got me thinking of what did I really wanted to say about my research. Although I was down and depressed for I think a few days after that, I spent a whole 3 weeks afterwards revising my objectives (write, re-write, re-re-write) and refining my tests, and I feel that I have found a ‘chord’ in my work, which was all worth it (I hope!).

Smlm ada seminar group kami. Kali ni Hilmi tu yg present. Oleh sbb bidang aku dan Hilmi lebih kurg sama, byk jgk aku dpt observe dan baca drpd potential internal examiners nya soalan2 (internal examiner ni ahli lab jg – maybe postdoc or lecturing staff). Rata2 nak tahu use and application, x mau sgt theory2, kena bg seketoi contoh kegunaan yg mendatangkan manfaat dan memudahkan cara.

Oleh yg demikian, aku pon buat la composition sket2. I never really got serious with composition because I kept thinking that I am doing this for the technology side, so x perlu spend lama2 on the artistic side, but thoughts like that might just grill me alive during viva. So like it or not, I have to show and demonstrate that there IS a space and need for my work in the music community.

The timing for pelawat td dtg couldn’t have come at a better time, because everyone was asked to prepare a showcase of their work. So utk julung2 kali nya, I feel like sharing a bit of what my work is really about.

I’m working on concatenative sound synthesis (CSS). Simplified, it’s where you use snippets of audios from various sources to make a completely new sound, based on a target sound that you ask the ‘concatenative system’ to generate. My work involves finding out what affects the concatenation, and of course, developing the system that addresses some of the shortcomings of existing systems (it is a relatively new field). Unlike Music Information Retrieval, CSS is not really focused on returning an EXACT match, but more like, cari sound yg serupa tp x sama. Perhaps the pics below would help clarify further (it’s from a slide, but I took only part of it. Could have used SlideShare or something like that, but I have materials I’d rather not made public yet (bahan x published lg), and am too lazy to edit. So snapshot aje lah solution nya.

Okay, so what sort of sounds would you expect from my CSS system? Something like below, so I have to say that this might not be everyone’s cup of tea. Some may even hate this piece, but then again, it’s like that with art, especially avant-garde art, very risqué gitu. I have to say this piece is quite raw. I didn’t have much time to clean up the audio, there are clicks in the audio that I need to get rid of, and I couldn’t be bothered to do fancy transformation on the piece, so basic fade out would have to do at the moment. I incorporate beats that I took of the internet (for free, thank you!), and combined it with the sound of the whales ‘singing’, that was generated automatically using my CSS system.

The title of this piece is ‘Woeing Wails of Whales’. It’s meant to portray the woeing and wailing sounds that whales make prior to and during their act of mating. The original sound file of the whales before the synthesis took place was actually sounds of mating whales, and I added the beat to sort of show the ‘speed’ of the process, i.e. courting (slow, cautious – first 15 seconds, and then mating (where the wailings start and beats got faster), and an abrupt end at the peak (as in, dare I write this, orga*sm (???), as you’d normally expect in the animal kingdom.

Ok, mmg ada unsur2 erotika plak, but what’s contemporary art without one kan? You can’t go past a gallery without a nude scene, so themes like these are heavily played in this field. Looks like all that ‘learning and studying’ at the galleries have paid off. Heh.

Also, pemilihan imej tersebut bukan lah suka2 sbb dia gmbr whales. You might say imej ni resolution teruk lah, berkotak2. It's not a smooth image because the picture is actually made of composite of other images, in a process known as Photo Mosaic. If you look closely, the pic of the whale is made up of smaller pics of ramdom pics in my pics folder. Similarly, CSS uses smaller, random sound files to compose a bigger, coherent sound.

Even if it sounds godawful, please bear in mind that my work is on the Artificial Intelligence side of it, lots of calculation and coding, and trial and error, to actually get a COMPUTER to COMPOSE this. So we can't really expect much (yet). Apapon, I'm already happy that SOMETHING is coming out of the system. Just need to justtify lah in the thesis and argue the empirical findings after this (which is not yet finished, oh dear me) :P

So what do you think if the piece?


Unknown said...

edgy gitu.. hebat.. i wouldn't dare to comment further sebab tak reti pun.. haha..

Mama Zharfan said...

hehe me zero bab2 ur research area tu :) apa2 pun me doakan u excel in ur viva :) bila vivanya?

isabelle said...

i ni buah cempedak di luar pagar.
xdapek nak komen lebih2.

but i bet, if it is ur product, it should be great.

Oyis said...

mast: hah i like tht word, 'edgy'. makes it sounds cooler than what it actually is :P

mamazharfan: thanks kak suzie. moga2 makbul doa akak. saya ni kalo bole nak la merasa kejayaan mcm akak jg. hish, lambat lg viva. kena finish writing dulu ni baru leh jawab soalan tu huhu T_T

isabelle: mau i jolok itu cempedak? hehe... oh, that last sentence is really too kind. malu i...

MRM said...

Information retrieval in general, is not to look for EXACT match actually, more towards those that are 'most similar' to the intended material. Kalau exact match, I think would be recognition (I think)?