Hello giga-users. Your assistance is kindly requested to help with the final stage demo-testing of my new piano gigainstrument.
Because of the limitation I have encountered, being unable to exceed a file size of 2 Gigabytes for a Gigasampler piano instrument file on my PC, I am changing my approach slightly. I have decided to create two different piano gigainstrument versions from the new samples.
One will be a smaller version, containing fewer actually recorded strike layers. At the moment, it looks like I can fit 8 strike layers (4 pedal-up and 4 pedal down) comfortably under the 2 gigabyte limit, with up to 20 second decays for the bass octaves. In order to make up for the missing layers, I am creating pseudo-layers within the instrument file using the gigasampler filtering capabilities. This helps greatly in smoothing the transitions between the actual recorded layers.
This also has the advantage of being able to run (but just barely) on a PC with only 128MB of RAM .
The second, larger version, containing all the 16 layers, and whatever other effects I can model, I will defer till some time later when I have had a chance to see if the upcoming Gigastudio can indeed eliminate the problem with the 2 gigabyte file size for that new giga-instrument format.
The giga-compression, which Nemesys has kindly offer to let me use, may be implemented for either or both of these versions as necessary to keep within the target file sizes or other limits.
So, with this change in strategy, I am close to finalizing the new 2 gigabyte version. In order to get some early, pre-release feedback, I would welcome the opportunity to create some demo MP3 audio files from MIDI files sent in by readers of this BBS.
I have been using a variety of test case MIDI files, and have already been posting a changing set of MP3 demos of those, but I realize that my own taste in musical styles may not coincide with yours.
If you are willing to send a MIDI file of your own for me to render to MP3 audio, and can offer constructive feedback on the results, please let me know (e-mail to: firstname.lastname@example.org). I expect to be ready to work with externally contributed demo MIDI files possibly by the end of the week (April 20/21). The current 2 gigabyte version has both pedal-down and pedal-up samples merged into the instrument.
I just posted 3 new in-process demo MP3 files, live-played with pedaling included.
These are on my page at: http://www.wstco.com/gigastatus/
Extra bonus \"brownie-points\" (sorry, no real prizes) will be given to whoever first correctly identifies how the 1906 song in the third demo was later used.
Thanks in advance for any assistance and feedback.
How about a version with 8-velocity splits with pedal up and none down? I personal don\'t use the pedal that much anyway, and blv the more velocity splits the better the expressiveness. (though, maybe the pedal down samples, even if only triggered occasionally, make more of a difference than I realize).
<BLOCKQUOTE><font size=\"1\" face=\"Verdana, Arial\">quote:</font><HR>I personal don\'t use the pedal that much anyway<HR></BLOCKQUOTE>
I personally ALWAYS use the hold pedal. I guess it relates to style of playing and type of music. Hopefully GigaStudio will allow for larger instruments, but I don\'t understand why we would have to wait for the release of GigaStudio to find out about this. Surely this is not a secret? They\'ve released a lot of details about the upgrade!
The winner of the \"trivia contest\" was Michael McInnis (http://www.mm-pro.com) who was the first one to correctly identify the 1906 song \"When The Mocking Birds Are Singing In The Wildwood\" as later being transformed into the theme song for \"The Three Stooges\". Contrary to my posting above, I will send Michael a free upgrade copy of the new 8 layer version, when it is ready, as a prize.
How about a version with 8-velocity splits with
pedal up and none down?
I too use the sustain pedal only sparingly for ragtime/stride piano, which is my personal favorite musical style. That is the reason why my first set of samples did not bother to include the pedal-down samples. I was fairly surprised myself by the difference that the added pedal-down layers make. The word that I think best describes it, is that it gives the piano a more \"powerful\" overall sound.
I also find that the additional layers make a large difference in improving the expressiveness, and smoothness of dynamic timbre response, for the piano gigainstrument. This is the reason I wanted to capture all the extra layers.
I don\'t really want to offer too many different versions of the piano sets, as this becomes confusing to users. I already find that I spend far too much time answering questions about my current variety of soundfont/bankset/wavefile/gigasampler format versions, even though I have tried to cover everything in detail on my website.
If there is an overwhelming groundswell of requests for an 8-layer pedal-up-only version for the regular gigasampler, I\'m willing to consider adding this as an option. For the moment, the 8-layer-mixed (4 up, 4 down) for the gigasampler and the 16 layer-mixed (8 up, 8 down) strike layer versions for the upcoming gigastudio, seem like the most widely useful options to me.
I don\'t understand why we would have to wait for
the release of GigaStudio to find out about this
I\'m \"pushing the envelope\" for all the soft-synth programs that I use.
I believe that my current 16 layer piano sample set is the largest ever produced to date (from over 32 gigabytes of raw note sample data). I was evidently the first one to encounter the 2 gigabyte file size limit. The Nemesys folks did not think this limit should have occurred until a 4 gigabyte file size. So, the moral of the story is that you don\'t really know for sure until you have tried it.
There are still many features that I would like to see implemented in the soft-synths to improve the modelling of various aspects of real piano behavior and sound. I believe that the extra capability of the gigastudio may allow me to program some of these characteristics better than I am able to do within the limitations of the gigasampler. However, as I said above, I won\'t know for sure until I give it a try with the actual program.