Forums

    15 posts

  • avatar
    73 sounds
    88 posts
    RE: Newbie - Sample Pack (questions?)


    Hi folks,

    If I was to go about creating my own SAMPLE PACKS (say, of 'MOOG' synth tones), then what would be the best way to go about it please?

    Do I simply play (eg.) '3 x notes' per octave across the range of the synthesizer and EXPORT them as 48kHz / 24-bit Wav files?

    I have never done this before but it is something I have been thinking seriously about for a couple of years now.

    Many thanks in advance for any kind advice offered here . . .

    Best,

    Paul David Seaman (UK)

    Fender Rhodes Suitcase / Fender Rhodes Stage / Hohner E7 Clavinet / Yamaha DX7 Mk1 & DX21 / Mini Moog / Multi Moog / Korg Poly 61 / Roland Juno 6 / Roland XV5050 (module). Cubase Pro 8.5 / Sibelius 7. Dell Inspiron 3737, Windows 10 Professional, 1.70GHz, 16GB RAM, 64-bit, Intel Core i3.
  • avatar
    2064 sounds
    2241 posts


    That seems perfectly reasonable. And put all related sounds into a pack (e.g 3 notes per octave, 5 octaves = 15 sounds).

    I want to believe.
  • avatar
    73 sounds
    88 posts


    Hi Alien,

    Thank you. I got it on that level now cheers.

    But what I am really stuck on is the actual PROCESS of how to go about the SAMPLING procedure. For example, do I need to set 'Cubase' to '16-bit' or '24' if I am recording a vintage MOOG synth? And do I need to EXPORT the audio Wav file at the exact same bit-depth as I RECORDED in?

    Is there an ideal 'dB' (say '-4') which I should adhere to when RECORDING in the single notes? Also, because the volume of the SAMPLES (eg. Moog) will differ depending on which notes I am playing (ie. low notes will probably sound louder than higher ones), should I modify the INPUT GAIN (in 'Cubase') accordingly to try to maintain a consistent (say) '-4dB' throughout the entire RANGE of the synth (ie. so that low to high notes are all the same volume)?

    As I say, I am totally new to this . . .

    Ta,

    Paul

    Fender Rhodes Suitcase / Fender Rhodes Stage / Hohner E7 Clavinet / Yamaha DX7 Mk1 & DX21 / Mini Moog / Multi Moog / Korg Poly 61 / Roland Juno 6 / Roland XV5050 (module). Cubase Pro 8.5 / Sibelius 7. Dell Inspiron 3737, Windows 10 Professional, 1.70GHz, 16GB RAM, 64-bit, Intel Core i3.
  • avatar
    2064 sounds
    2241 posts


    The Moog is analogue so, in theory, the higher the bit-rate and bit-depth, the better your recording will be.
    It is usually said that the weakest element in a signal chain will determine the quality of the overal chain. - So, imagine that you have a faulty cable taht crackles. Even if you have a high quality microphone, a quiet room, good amplifier and digitize at a high bit-rate and bit-depth, the recording will still be noisy. And recording at 16 or 24 bit will make no difference to this.

    I imagine you are plugging your synth to the audio interface directly (no microphone). The above was just an example.

    If you have a good quality audio interface, yeah go for it and record at 24-bit.

    As for converting the file when exporting. This is a tricky one. I have many times recorded at a higher bit-rate than I needed and then converted down. I found this to result in better sound quality - but this was many years ago, with worse recording equipment than we have today.

    You make a very important point about recording volume. This is indeed the critical point.
    Before recording setup your synth to the sound you want to record. Set the volume to what you think will be a good volume (say -4dB). Then play the notes that you want to record one at a time and watch the meter - make sure you are not going into the red.
    You should try to record at a volume that is high enough but making sure you never go into the red.
    If you can trigger your synth from your DAW do it this way so you know the exact velocity and all notes have the same velocity.

    If you are very patient, you can also repeat this process and record the notes at different velocities. In that case, you should do the volume setting using the highest velocity that you will play.

    I want to believe.
  • avatar
    73 sounds
    88 posts


    Hi Alien,

    I am wondering if I might be better starting off with sampling the MOOG. It's in MONO and also I don't really have to stress about the 'DAC' and all that as much. I can just MAX OUT the soundcard/Cubase audio settings and get going.

    I currently have a Steinberg 'UR22' soundcard but I am getting the 'UR44' over the summer:

    http://www.ebay.co.uk/itm/272265535223?_trksid=p2055119.m1438.l2649&ssPageName;=STRK%3AMEBIDX%3AIT

    It does not have 32-bit float though.

    But yes, the MOOG would go into the soundcard in MONO (jack-to-jack), and then into the PC (USB).

    I think I would run the entire sessions @ 48kHz/24-bit. That would probably enable me to whack a few self-creations up on Google Drive for people. If I go up to 96kHz then files will really start eating up GB real estate.

    Please note, the MOOG does not have MIDI nor is it touch sensitive. The latter is a positive because it means all notes will not vary in application of velocity. Although I am sure that there will still be a variation across the range of the synth with respect to volume. In which case, would you:

    1 - adjust the MOOG volume dial,

    2 - adjust the soundcard input gain,

    3 - adjust the channel fader (in Cubase),

    4 - adjust the master fader (in Cubase)?

    Nothing is ever straight forward! lol

    Ta,

    Paul

    Fender Rhodes Suitcase / Fender Rhodes Stage / Hohner E7 Clavinet / Yamaha DX7 Mk1 & DX21 / Mini Moog / Multi Moog / Korg Poly 61 / Roland Juno 6 / Roland XV5050 (module). Cubase Pro 8.5 / Sibelius 7. Dell Inspiron 3737, Windows 10 Professional, 1.70GHz, 16GB RAM, 64-bit, Intel Core i3.
  • avatar
    73 sounds
    88 posts


    Hi again,

    Secondly, would you apply (to the recorded audio samples):

    1- Compression,

    2- Limiter,

    3- Maximizer?

    Obviously these would be applied post-recording.

    Here is a track I made last year which has NOTHING added. So it is totally the MOOG genuine article. Everyone said it sounded brilliant and that I should leave it untouched (because I was totally paranoid it need many of the infinite VST 'FX' applied to it). In the end, I actually agreed with everyone!

    It comprises a Bass line and then a Melody line overdubbed.

    https://www.freesound.org/people/monsterjazzlicks/sounds/347511/

    Paul

    Fender Rhodes Suitcase / Fender Rhodes Stage / Hohner E7 Clavinet / Yamaha DX7 Mk1 & DX21 / Mini Moog / Multi Moog / Korg Poly 61 / Roland Juno 6 / Roland XV5050 (module). Cubase Pro 8.5 / Sibelius 7. Dell Inspiron 3737, Windows 10 Professional, 1.70GHz, 16GB RAM, 64-bit, Intel Core i3.
  • avatar
    2064 sounds
    2241 posts


    Hi Paul

    First, let me begin by saying that there is no right or wrong answer to some of the questions you asked. Feels like you are over-agonizing over some of this (too many options, too many knobs to twidle, too much worry...)
    Let me make it as easy as it can be.

    The only volume that really matters in the signal chain you have described is the input volume into your soundcard. - If that is too high the sounds will be clipped and irreversibly distorted.
    Volume adjustments in the DAW itself are already in the 'digital domain'. If the sound clipped when entering the soundcard, it is too late to do anything about it with those faders.
    The key thing is not clipping the sound on the input.
    The fact the Moog is not velocity sensitive makes everything much easier - so we are in luck.

    I would go about it like this:
    1 - Set input gain on the sound card to middle value and not touch it anymore unless you have to.
    2 - You probably have playied the Moog through your soundcard before and have an idea of what volume would work well. If you have no idea, start low (~10% of max volume on the Moog) and work your way up.
    Press a key and watch the meters. At this point you probably just want to go to ~50% of the meter range. Adjust only the Moog output vol and keep all faders on your DAW at 0dB, is the easiest.
    3 - Now, play all the notes that you will be recording. If any pushses into the red (or close), lower the vol on the Moog. If you are still far from the red, push it up a bit.
    I would say if you are hitting -10dB or even -5dB on the strongest notes, that should be fine and you can start your recording session.
    4 - If you can't get to -10dB with the Moog at max vol, it is time to go back to the soundcard input knob and try increasing that.

    Special considerations, when yo apply compression etc:
    a) Do you want the pure Moog sound or an improved version? IF you want pure, stay pure. If you are being creative, use whatever effects you want smile
    b) Remember, you can always do both - pure and processed versions...
    c) When do you need it?
    Limiting is a waste of time, I think. Since any clipping at the recording stage cannot be fixed anyway, there is not much point limiting afterwards. So don't bother.
    If the particular sound you are recording sounds very faint for some notes and very high for others you can consider compression/maximizing. Since this can also boost any line noise present in the original recording, if the difference is really big, maybe just re-record the faint notes with a higher output vol on the Moog.

    After you have completed your recording session you can:
    A) Noramlize the complete recording then cut each note and save. This makes the loudes point on the whole recording = 0dB.
    B) Cut each note, maximize and then save. This makes the loudest point on each note 0dB.
    Option A gives you the true sound of the Moog, including the volume balance between the different notes. Version B makes all notes max volume.
    I would go with option A, since people downloading the samples can easily normalize them if they so wish.

    I want to believe.
  • avatar
    73 sounds
    88 posts


    Thanks Alien,

    Wow, that reads almost as good as a MANUAL! lol btw - Do you use Cubase yourself please?

    You obviously picked up on the fact that I am being perhaps over-analytical here. This is for a few reasons:

    I have never done this before and so I would like to try and get at least the general working PROCEDURE correct (even if the aural results are not 100% tip-top). So it is more a matter of gaining an understanding of the process (ie. the plugging in of leads, which dials to tweak, etc) which you have just kindly explained. I am much more clearer on these points now thank you wink

    In actual fact, a lot of what you are suggesting is more or less the same as if I was recording band or vocalist! But for some reason, I have always felt that a completely different set of rules come into play with reference to the world of SAMPLING. I think that because of the definition of the word itself, the implication is that a very defined SCIENTIFIC approach is called upon; meaning I have viewed it as being highly more intricate and something of a 'dark art' when compared with (say) programming a 'DX7' or recording a Guitar/Vocalist.

    Yes, I agree that it would be preferable to deliver everything as 'dry'. However, there may be the odd occasion where, to the contrary, I would add 'FX' if it were to enhance the identity of the timber, and be with keeping of the spirit of the genre. For example, I might want to create a 'Sci-fi' soundtrack using (say) 'Sample & Hold' and the MOOG tone would be too stark and somewhat meaningless without having (say) an exaggerated 'Delay' applied to it.

    Much appreciated,

    Paul

    Fender Rhodes Suitcase / Fender Rhodes Stage / Hohner E7 Clavinet / Yamaha DX7 Mk1 & DX21 / Mini Moog / Multi Moog / Korg Poly 61 / Roland Juno 6 / Roland XV5050 (module). Cubase Pro 8.5 / Sibelius 7. Dell Inspiron 3737, Windows 10 Professional, 1.70GHz, 16GB RAM, 64-bit, Intel Core i3.
  • avatar
    73 sounds
    88 posts


    Hi Alien,

    Ok, so I have made a decision now thanks. What I am going to do is this:

    RECORD @ 24-bit.

    Then, RENDER it out in TWO versions:

    44.1kHz / 16-bit,

    and

    48kHz / 24-bit.

    Rather than reconfigure settings for each project, just leave them set to RECORD as aforementioned. So a one-size-fits-all of sorts.

    Do you think that is a viable method?

    Paul

    Fender Rhodes Suitcase / Fender Rhodes Stage / Hohner E7 Clavinet / Yamaha DX7 Mk1 & DX21 / Mini Moog / Multi Moog / Korg Poly 61 / Roland Juno 6 / Roland XV5050 (module). Cubase Pro 8.5 / Sibelius 7. Dell Inspiron 3737, Windows 10 Professional, 1.70GHz, 16GB RAM, 64-bit, Intel Core i3.
  • avatar
    2064 sounds
    2241 posts


    Go for it.

    I want to believe.
  • avatar
    73 sounds
    88 posts


    Alien,

    Would you simply hold down a single-note for a few seconds and stop/start Cubase's record function (repeating this process 20 x times until all notes are documented)? If so, it will be quite tricky to perform this to exactly the same duration for each note because the Moogs and Poly_61 are not MIDI. With the DX21, it is MIDI so I can draw in the MIDI data on the Cubase Piano-Roll (with the pencil tool) and have a cup of tea while it is triggering the whole recording!

    Also, the Moog and Poly_61 (and I dare say possibly the DX21) will have to have their ends trimmed (I think this is called 'truncated'). And so the degree of 'SNAP' (as it is called in Cubase [or 'Quantize' resolution more commonly elsewhere]) will need to be firmly established before any such editing takes place.

    To illustrate this problem, if you care to have a quick listen to a 'DX7' piece I made last month consisting entirely of 2 x Bar & 4 x Bar samples (or 'loops' [call them what you may]), you can hear that the Orchestral type tones (which enter half way thru) do not line-up correctly because (I believe to be the case) the front-ends were not sliced at identical points! This was my first attempt though and so there were bound to be numerous glitches:

    https://www.youtube.com/watch?v=tJgIocjrZoY

    It will probably be easier made single note or one-shot samples, but I would thing that the same rules remain, albeit to a less pedantic degree.

    Ta,

    Paul

    Fender Rhodes Suitcase / Fender Rhodes Stage / Hohner E7 Clavinet / Yamaha DX7 Mk1 & DX21 / Mini Moog / Multi Moog / Korg Poly 61 / Roland Juno 6 / Roland XV5050 (module). Cubase Pro 8.5 / Sibelius 7. Dell Inspiron 3737, Windows 10 Professional, 1.70GHz, 16GB RAM, 64-bit, Intel Core i3.
  • avatar
    2064 sounds
    2241 posts


    Best to do one single recording session. That is why I suggested checking the volume setting would work for all notes before starting to record (see my previous posts).

    Press record and then play the notes you want to record, one after the other. Take care to make sure the tail of each sound has finished completely before triggering the next sound.
    Does not matter if they are not all exactly the same duration. As long as they are not massively different.

    You end up with one big recording, which you can then later cut into individual samples.
    I am not familiar with the sample editing capabilities in Cubase. But if you find these are insufficient to edit and cut out your individual sounds, then I recommend you try Audacity. A free sample editor which covers simple functions well and even some advanced ones too.

    Cutting out single notes is easier than cutting out perfect loops.
    I sugges you start with ouds with a clear attack, since then it is easy to find the starting point of each sound.

    As you practice this, you will become better at cutting out 'seamless loops'. There are some techniques for that as well, but won't cover them now. Start with simple one-shot samples.

    I want to believe.
  • avatar
    73 sounds
    88 posts


    Thanks Alien,

    Oh ok, so you mean just LEAVE Cubase recording and make the whole recoding in a single pass. It's only single notes and so it will be unlikely that there will be any mistakes.

    I was over concerned about making each sample (say) exactly 2 x seconds precisely. But are the professional sample packs not absolutely spot-on? Or do you think near-as-damn-it is suffice for most peoples wants and needs.

    Cubase is very much capable of performing the editing tasks required. However, my question was/is more to do with:

    in the "YAMMY" You-Tube piece I linked above, I was not sure how far to zoom-in on the waveform before executing the cut. From a distance, you can easily see the wavy-line where the tone begins, but as you zoom-in, the line is so magnified that I find it indistinguishable as to where the actual sound starts? Hence, I think that was the problem with "YAMMY" in that I made the cuts slight to early/late and so the audio files come in out-of-time!

    Cheers,

    Paul

    Fender Rhodes Suitcase / Fender Rhodes Stage / Hohner E7 Clavinet / Yamaha DX7 Mk1 & DX21 / Mini Moog / Multi Moog / Korg Poly 61 / Roland Juno 6 / Roland XV5050 (module). Cubase Pro 8.5 / Sibelius 7. Dell Inspiron 3737, Windows 10 Professional, 1.70GHz, 16GB RAM, 64-bit, Intel Core i3.
  • avatar
    2064 sounds
    2241 posts


    Now is time to stop talking and start recording!

    Cutting single notes is much easier than cutting phrases to loop seamlessly.

    Even if you make some mistakes, that is part of the process and you will learn from them.

    I want to believe.
  • avatar
    73 sounds
    88 posts


    Hi,

    Yes, I agree. Slicing and lining up loops is more problematic than with single-note files.

    Paul

    Fender Rhodes Suitcase / Fender Rhodes Stage / Hohner E7 Clavinet / Yamaha DX7 Mk1 & DX21 / Mini Moog / Multi Moog / Korg Poly 61 / Roland Juno 6 / Roland XV5050 (module). Cubase Pro 8.5 / Sibelius 7. Dell Inspiron 3737, Windows 10 Professional, 1.70GHz, 16GB RAM, 64-bit, Intel Core i3.

    15 posts