VSCO2.sfz as .sf2

• Aug 16, 2021 - 13:42

VSCO2.sfz as sf2. nothing is changed, except the roundrobins of the ensembles play all together now. 1.6 Gb, (stereo, for the believers). woodwinds, brass, string sections, a solo violin and a handful of chromatic percussive instruments. pick it up here;

https://drive.google.com/file/d/1DOQ6yiQlJkgicOZIz0aCZtUIlrsRgfh-/view?…

when its no longer there, it is no longer there.


Comments

Thanks - there are some really nice instruments in there for sure.
It seems to be some sort of commumity orchestra with the sounds released in sfz for free use.
I wonder if they would mind re-releases in sf2 with velocity layered instruments like flute or violyns progressing from sus, expressive to vib over the velocity range? I am prepared to put a bit of work into this if allowed.
Thanks again.

In reply to by Jonky Ponky

you are completely free to do what you want with these samples. this is what they say;

"The library is under a Creative Commons 0 (i.e. public domain) license.
That's right, you can download 3 GB of samples for free with no rules, no royalties, no limits on how or when you can use it, no annoying e-mail signups, and so on."

read it yourself here:

https://vis.versilstudios.com/vsco-community.html

regards bottrop

I respect your effort.

Unfortunately, the original sfz and sampleset (VSCO2-CE), which is the basis for this, consists of unbalanced samples. First of all, these samples have low levels, secondly, have no loops, and lastly, lack integrity. So it's like every instrument was recorded in a different way.

Maybe it can be used in works at a very slow tempos. Because some instruments take a few seconds to reach normal volume from the initial stage. But they are insufficient even for a medium tempo piece. It is necessary to edit and correct one by one in an audio editor (this is how I created the ASO), but unfortunately there is no time for that.

There are no Solo Viola and Solo Cello instruments. This is also a big minus.
However, there are all kinds of unnecessary details. It is not clear why they bothered with such unnecessary details as sus-vib-quiet and staccato.

So, I think Versillian-Studios' effort on this set was almost wasted.
It's just a free starter kit for soundfont creators to pick up and play with.

I struggled with this set for a long time. I changed it to mono, adjusted the loops, increased the levels, reduced multiple layers to one, sorted by GM, etc. So I reduced the wasted space down to 266MB.

It wasn't that great after all. Although it is considered good according to the initial situation, it is not satisfactory for me. Maybe it can be used to take a few instruments from here and add them to another sound font, or to create another instrument by combining the same instruments from several different soundfonts.

In reply to by Ziya Mete Demircan

it was not too much of an effort, Ziya, i opened the sfzs in Polyphone and saved them as sf2s, then mapped all Presets in my cuckoo soundfont.
a low volume level is easily solved, just select all samples, then; Tools> Samples> Adjust Volume> Normalize at 80% (did that already) then select all the Pan and Attenuation settings in the Instruments and hit Delete).
i Unlinked all stereo samples (Tools> Global) and divided left and right in seperate instruments. this was a bit of an effort, because i still use Polyphone Version 1.9, (its better for my eyes, newer versions may look good on a telephone, but who is going to make soundfonts on a telephone?) and in 1.9 you cant Cut & Paste samples from one instrument to the other.
converting stereo samples to mono is never a good idea, since often times only the left or right channel is the bad guy and will spoil it all. pick the best and best matching samples (for your ears) from the left and right instruments and build your own mono instrument.
Quiet is the Versilian word for p (they use it as piano velocity layer).
in their Sfzs they use the staccato samples as attack (like many inexperienced people still do in sf2s) and they switch between the abundance of pizzicatoes and spiccatoes, so you dont hear the same sound from a key everytime (impossible in sf2).
after throwing away all the bad samples (Tools> Global> Remove unused elements) you can edit the samples in a wave editor (Looping, boost or attennuate attacks, remove blank sections.. etc.) and get some nice instruments.
regards bottrop

In reply to by bottrop

Adjusting low volume samples is of course not a problem. The real problem is that they were not recorded at the required volume. I've worked in studios for years, setting up microphones (for both instrument and vocals), working on mixers. I know what problems can be caused by amplifying a low recorded sound and which frequencies will decrease when viewed with a spectrometer.

Also, when I listen to a sample, I can tell what distance is between the instrument and the microphone. In the samples we mentioned, the microphone is placed very close to what is necessary. Think about it: When listening to a symphonic piece, do all the instruments play right next to your ear? They most likely used a close-up dynamic microphone or a small-diaphragm stereo condenser microphone, although they should have used a large-diaphragm condenser microphone at least 15 cm away.
Perhaps this sampleset is part of a larger project. It may consist of sample dumps that are unusable and that are too closely recorded.

When preparing stereo samples for the instrument, the recording is made either from the right and left diaphragms of a single microphone, or separately with two microphones of the same brand and model.
No matter what technique is used, there will inevitably be a difference in tone and color between the two channels. Therefore, it is fine to choose either the left or the right channel, it is problematic to choose the best ones in the right channel for one note and the left channel for the other.

Quiet samples are meaningless if they do not include a special technique such as "sul-tasto". Because these instruments already have a defined area for the v1(0-63) range.

Maybe I can understand using it as a spiccato because of the sound difference, but this staccato technique is very easy to create in a soundfont, there is no need to record separately.

It is possible to use these staccato samples as an attack part, but there is no sustain part to be used afterwards. in ASO soundfont, I created these parts one by one by cutting the samples. st=starting-part, su=sustained-part. But these were parts of the same sample, not recorded separately. For unbalanced samples this method is fine but uses two voices for one note at a time and can reduce polyphony on systems with light CPUs.

Like I said, it takes a long time to deal with each sample individually in an audio editor. This set is not of the quality to be worth doing. Maybe a few instruments will give good results. But that's all.

Although all these things I have written seem to be annoying and demotivating, in fact, it is just an explanation that shows the problems and cons of this sample set. I guess the only good thing we have left is that it's public-domain and free.

Of course, I'm not making all this up in my head or saying it as a result of a casual glance. I've worked on this sampleset, tested it with scores, and I'm talking about my experience. I attached the screenshots.

Attachment Size
zmdc-01.png 42.57 KB
zmdc-02.png 2.92 KB

Do you still have an unanswered question? Please log in first to post your question.