Meeting yg Tertangguh On Friday Itu..

Baru habis meeting dgn SV. Yg patutnya hari Jumaat lps tu, tp dia tiba2 ada kecemasan plak katanya, so dibawa ke hari ini. Fuhhh lega... Ada 3 part aku presented kat dia td, test result utk concatenation without any algo, then with Random Algorithm, and lastly based on the features derived from the perceptual model. The first two parts tu sbnarnya dah buat test skala kecil2an dah before balik Msia hari tu, tp this time buat properly; I plotted graphs to show the relationship between each pitch classes, illustrated the ratios of original versus random generation, and did some statistical analysis jg, just to see the results more quantitatively.

Yg part 3 of the test ni something completely new la, x pernah tunjuk kat SV lg, but have discussed before. Part 3 ni was quite a challenge, even more so than the other two parts, bcoz it doesn't only deal with pitch information tetapi other audio information jg. I actually had to go back to findings and tools from my MSc research, including guna MARSYAS and WEKA tools to aid me. In the end, I think he was quite happy to see what the results have been, although x de anything entirely bravado can be said about it yet. He did mention that throughout this summer kena bekerja dgn lebih keras lg, bcoz time really is catching up on us, especially after a minor setback sebab balik Msia hari tu. Personally I think dia quite gutted sebab x dpt publish apa2 at this stage, and masih tiada breakthrough results lg, tu yg kena push and push and push! Oh, and he also said that he needed my 'system' ni by the end of summer, sbb nak guna to aid his composition (or recomposition, rather) of Mozart's classical pieces using concatenative synthesis *jatuh terlepuk haku di lantai!* <--- from a mixture of pressure and jg elemt of flattery, sket2 ;P

Good news is that we are working towards the memory mechanism now, which really is the HEART of this PhD. Bunyi mcm senang, tp sebenarnya aku rasa adalah susah sekali sbb nya masih x de idea nak proceed mcm mana dgn cik memory nih. Also I have no idea how to handle multi-dimensional data, and spt yg telah dipersetujui ramai semasa aku present dpn work colleagues aku semasa di Msia (ye, sempat lg present masa balik 3 minggu itu hari ye), I really should avoid getting myself sucked into using Principal Component Analysis if I can help it. Mmg ternyata aku kena on kan balik reading, carik solution through other people's past work. Tu la masalah dgn aku ni, kalo dah coding, lupa reading, kalo reading, lupa coding plak. Another thing, we have agreed that now I have to make sure everything is done on-the-fly, dan bkn lg ada delay spt my prototype skang yg hanya mengupdate matrix table after every 20 cycles. Takut mmg takut sbb coding has never been my strongest point, so I might take longer than I wanted to spend on this, tp nak buat mcm mana kan, redah je la.

Things to do:
1. Continue with features testing
2. Re-read Marcelo's thesis, and understand how he handles different features
3. Build the memory on-the-fly
4. Add a sub-part to Part 2 test, repeat using normalized frequency instead of pitch classes

I also told him about UM*IES 2010 paper being accepted, and we are working towards finding a means to get there. X nak menaruh harapan, sebab x tahu apa kesudahan sbb x de funding lg as of yet. Will reveal later lah kalo ada rezeki. Kalo x de rezeki mmg konfem la x dak publication ini tahun. Lain tahun la ye, Insya Allah. Finally we discussed a bit on my last 3 months progress (Feb, March, April 2010) and overall he thinks I am on the right track, which is good to hear.

Because my UM*IES paper was on Gamelan (a little side project, ada buat preliminary test je early last year and ada sedikit je kena mengena dgn research aku skang, but worth publishing - me thinks!), it kind of jogged my SV's memory and he mentioned that the Faculty of Arts dkt UoP ni telah membuat purchase satu set Gamelan baru2 ini. In fact, Mike M (lecturer Arts di sini yg aku pernah attended seminar dia), telah pon pergi ke Msia baru2 utk membuat purchase tersebut. So imagine betapa berbunga2 nya hati aku sbg ex-Gamelan player cum Gamelan instructor x bertauliah ini apabila mendgr akan berita tersebut.

And when my SV asked if I play the Gamelan, mmg ayat dia x abis lg aku dah jerit "Yes!", nasib baik x berbunyi ala2 orgasmik di situ tgn aku x terhangguk2 menepuk imaginary gendang gamelan aje terus (I had mainly been the percussionist for our Gamelan team, altho I still can play the other instruments on the ensemble). And when aku tanya bole ke org bkn Arts spt aku nak join, and my SV replied, 'Of course', mmg aku di awang2an kjp. And when he continued by saying, 'Maybe you could even give lessons, I will let Mike know', nyaris2 aku terpengsan di situ jua. Tp aku telah cpt2 tersedar balik ke alam realiti, when he quickly added, 'All the more reasons to speed up work, then!' *Duh! aduh pedih*

Mmg bunyinya over tp hanya org yg pernah terlibat dgn seni ni aje tahu betapa gediks nya segala anggota utk kembali perform. Hopefully, if all things work well, that's my summer sorted then :)