Migrating Desktop Samples Into Obsidian Workflow Questions

2»

Comments

  • T4HT4H
    edited February 2019

    @Nomzai said:

    @T4H said:
    A quick update for those making the migration from desktop with big vst libraries wanting to get them into obsidian and beyond. I bought the full version of SampleRobot 6 Pro. I’m using virtual midi and the loopback application to route the midi and trigger sounds directly from Omnisphere into SampleRobot. It involved manually setting up midi learning in Omnisphere with my selected patches 1-128 however I started with 70 to see how it does. So far it’s working and sampling every 3 semitones per patch and then changing patches and repeating again. Also, you can load a text file of your patch names and it’ll automatically name and number the .wav files for u with the patch and note it’s in. For Example (001 Full Choir C3.wav). This is a great workflow once you get it up and running and I’m happy I can get my day back to myself and let that process run automatically. It’ll take around 4.5 hours to complete!

    Very interesting thread. Thanks.

    What takes 4.5 hours to complete? The initial setup?

    Thanks, glad to be able to contribute.

    No the initial setup only takes around 10-20 minutes or so depending on which vst you’re using to assign the midi program change values for each patch etc. The samples playing back (at around 15-20 seconds per single sample since they’re pads and choir patches) and then recording and it resetting and then playing the next key is what takes the hours. It’s an online ‘realtime’ bounce so each sample has to be played live automatically as it makes its rounds through the notes and patches recording everything.

  • T4HT4H
    edited February 2019

    Hey @dendy or anyone else with some understanding on NS2, the most painful process literally of all this migration now is having to type the name of every patch on a touchscreen instead of keyboard. Also, copy and paste from other programs like a text file on iPad isn’t working either. My hands are literally sore lol from creating almost 2000 obsidian patches from .wav files on the desktop and I’ve got most parts automated but this is a real bottle neck and pain in the hand 🤚. Any workflow suggestions to have the names of each patch auto generate or be copy and paste-able?

  • @T4H said:
    Hey @dendy or anyone else with some understanding on NS2, the most painful process literally of all this migration now is having to type the name of every patch on a touchscreen instead of keyboard. Also, copy and paste from other programs like a text file on iPad isn’t working either. My hands are literally sore lol from creating almost 2000 obsidian patches from .wav files on the desktop and I’ve got most parts automated but this is a real bottle neck and pain in the hand 🤚. Any workflow suggestions to have the names of each path auto generate or be copy and paste-able?

    Change them in the Files app. Copy from Notes or wherever you have your text and paste into the patch name. You still have to save the patch initially. Probably be easiest to use a letter per file (since it defaults to Untitled which is selected and with the caps key on. I’d do 26 patches (A-Z) then rename them in Files with copy/paste and go through the cycle again. Tedious but it would work. It would be nice if they keyboard allowed copy/paste or took the patch name automatically from the root of the sample file name.

  • edited February 2019

    i'm not sure i understand exactly your workflow, but isn't better rename those files from desktop ? Just activate WebDAV server in NS (Settings > WebDAV)

    Then use some WebDAV client on your desktop computer to connect to NS (i have great experience with CyberDuck (https://cyberduck.io/download/) , super stable) and rename your files from there ?

    To connect to NS WebDAV server you just need tp have both NS and desktop in same local area network... then enter NS address in WebDAV client on desktop (you see what address it is inside NS, where you activated it's WebDAV server), as login/password use whatever, it doesn't matter.

    probably not solution to your problem, just trying :-)

  • @anickt said:

    @T4H said:
    Hey @dendy or anyone else with some understanding on NS2, the most painful process literally of all this migration now is having to type the name of every patch on a touchscreen instead of keyboard. Also, copy and paste from other programs like a text file on iPad isn’t working either. My hands are literally sore lol from creating almost 2000 obsidian patches from .wav files on the desktop and I’ve got most parts automated but this is a real bottle neck and pain in the hand 🤚. Any workflow suggestions to have the names of each path auto generate or be copy and paste-able?

    Change them in the Files app. Copy from Notes or wherever you have your text and paste into the patch name. You still have to save the patch initially. Probably be easiest to use a letter per file (since it defaults to Untitled which is selected and with the caps key on. I’d do 26 patches (A-Z) then rename them in Files with copy/paste and go through the cycle again. Tedious but it would work. It would be nice if they keyboard allowed copy/paste or took the patch name automatically from the root of the sample file name.

    @dendy said:
    i'm not sure i understand exactly your workflow, but isn't better rename those files from desktop ? Just activate WebDAV server in NS (Settings > WebDAV)

    Then use some WebDAV client on your desktop computer to connect to NS (i have great experience with CyberDuck (https://cyberduck.io/download/) , super stable) and rename your files from there ?

    To connect to NS WebDAV server you just need tp have both NS and desktop in same local area network... then enter NS address in WebDAV client on desktop (you see what address it is inside NS, where you activated it's WebDAV server), as login/password use whatever, it doesn't matter.

    probably not solution to your problem, just trying :-)

    Thanks for this suggestions guys. I think I found a solution that’ll work until we get proper copy/paste and QWERTY keyboard support. I ended up loading all the patches and saving them in numerical order from 1-127 and then doing the WebDAV as suggested. From there I used the MacOS app ‘Better Rename 10’ and dragged my .txt file that lists all the patch names for 1-127 into it and it automatically renamed all of the .obs folders into their proper patch names! It isn’t perfect but it works! Thanks again for the ideas!

  • @T4H said:
    Thanks for this suggestions guys. I think I found a solution that’ll work until we get proper copy/paste and QWERTY keyboard support. I ended up loading all the patches and saving them in numerical order from 1-127 and then doing the WebDAV as suggested. From there I used the MacOS app ‘Better Rename 10’ and dragged my .txt file that lists all the patch names for 1-127 into it and it automatically renamed all of the .obs folders into their proper patch names! It isn’t perfect but it works! Thanks again for the ideas!

    I'm now a little confused about what you're doing/how you're doing it. Loading all the patches? Which patches?

    Until the newer posts, I thought you were using SampleRobot (or similar) to get the samples named properly on the desktop and then using Obsidian's automagical assign-to-zones-based-on-names from there. What am I missing? What are you (were you) QWERTYing?

    Part of the reason I ask: you seem pretty committed to this project and reasonably nerdy (nerdery pays my bills so...). As such, we might be able to help with a way to automate more of this process if you're open to nerd tools like the terminal.

    Possibly useful: OBS files are plain text files (well, XML-like plain text files). If you're comfortable in a text editor (not Word or Pages!), you might be able to do a bit of find and replace + rename to make this all quicker.

  • T4HT4H
    edited February 2019

    @Will said:

    @T4H said:
    Thanks for this suggestions guys. I think I found a solution that’ll work until we get proper copy/paste and QWERTY keyboard support. I ended up loading all the patches and saving them in numerical order from 1-127 and then doing the WebDAV as suggested. From there I used the MacOS app ‘Better Rename 10’ and dragged my .txt file that lists all the patch names for 1-127 into it and it automatically renamed all of the .obs folders into their proper patch names! It isn’t perfect but it works! Thanks again for the ideas!

    I'm now a little confused about what you're doing/how you're doing it. Loading all the patches? Which patches?

    Until the newer posts, I thought you were using SampleRobot (or similar) to get the samples named properly on the desktop and then using Obsidian's automagical assign-to-zones-based-on-names from there. What am I missing? What are you (were you) QWERTYing?

    Part of the reason I ask: you seem pretty committed to this project and reasonably nerdy (nerdery pays my bills so...). As such, we might be able to help with a way to automate more of this process if you're open to nerd tools like the terminal.

    Possibly useful: OBS files are plain text files (well, XML-like plain text files). If you're comfortable in a text editor (not Word or Pages!), you might be able to do a bit of find and replace + rename to make this all quicker.

    Hey @will I appreciate any suggestions you can offer to make this process easier, more automated etc. To help better understand my process thus far, and to help anyone else on the same journey now or in the future, I’ll go step by step through my workflow. This will be longwinded but please bare with me.

    Ok, so basically my objective is to capture the sounds of the presets or patches of desktop VST’s that I use. So massive, omnisphere, etc. Lets focus on Massive and my workflow using that for this example. Lets say I have a 3rd party sample set in massive format, inspired by whoever the latest trending producer is. I load all of those samples into the program section of massive. So slots 1-128 and I make the midi input the IAC driver. This will serve to make all of those massive presets able to be switched with midi program change. This makes all those patches/presets like a playlist.

    Next I use Bbedit and drag my folder of presets/patches that I’m going to record into it from the massive folder located in the documents folder. This will generate a text file of all the Massive presets/patches that I intend to record in order. This is crucial for the name assignments of all the sounds I intend to record not having to be typed in manually. Now, I go into sample robot and import that text file and it generates its own batch recording playlist, that matches that of Massives 1-128 Program list. After the initial set up I hit record and it will batch record all of those massive presets/patches in order 1-128. This takes a couple hours and so i’ll either start the same process on another computer or find something constructive to do lol.

    Now, after SampleRobot is finished, I have to export each sound one by one which takes 20-40 minutes total for 127 sounds/patches/presets depending on how quickly I’m moving, since I haven’t found a batch export option. I’m guessing they want you to test the sounds and/or loops before you bounce them. When I hit export the sampled instruments are automatically saved with their name from the earlier mentioned text file and the key that was recorded, for example one instrument being exported would be these 3 files: ‘001 Super Piano C2.wav, 001 Super Piano C3.wav, 001 Super Piano C4.wav’ Next, once i’ve exported all 127 patches/presets I batch folder create for all the instruments using ‘Big Mean Folder Machine 2’. This just helps with organization and it automatically names the folders based on the first however many designated letters of the .wav file and places all 3 .wavs of each instrument into them automatically.

    Now I have 127 Folders that each have 3.wavs in them ready to be imported into NS2. I airdrop them over to NS2 Library/Audio/Inbox. Next I go into obsidian and hit the automap feature and it automatically maps 001 Super Piano C2.wav, 001 Super Piano C3.wav, 001 Super Piano C4.wav into their corresponding keys.

    Now its time to save it as an obsidian patch/preset and where we get to the dilemma, NS2 doesn’t automatically name the patch/preset based on the .wav it imported, so you need to type in the name of the patch/preset manually. Problem is, NS2 doesn’t support a QWERTY keyboard, nor can you copy and paste from a text file on iPad into Ns2 to make it easier. You physically have to type in each patch, on the touch screen. Sounds like a small deal until you’ve repeated this a few dozen or 100’s of times. Eventually the hand fatigue sets in lol, especially when organization and labeling is a big deal to you, which up until this point has been perfectly preserved.

    I appreciate your suggestion @will and experimented with the .obs file and saw the line where i could swap the .wav sound for the new one but replicating a text file 127 times and doing find and replace seemed more tedious than I cared to do, however I would of done it if not for coming up with this solution. Saving each file in numerical order 1,2,3,4,5,6,7,8 etc (or a-z several times as @anickt suggested) in obsidian manually but making sure Im tagging the category as I’m saving. Hitting the Keys tag for Keys, etc. This is to make life easier in the long run when sourcing sounds for songs.

    Then doing the WebDav as @dendy suggested and from there using ‘Better Rename 10’ to batch rename all of those 1-128 .obs folder names into the correct preset/patch name using the same .txt file that I made at the beginning of this whole process. This names each .obs folder preset properly and in the proper order. Once complete, I go back into obsidian and hit refresh and now i’ve got a perfect list of presets/patches that mirror those I had in Massive.

    Any suggestions or alteration to this workflow to improve it are always appreciated. Eliminating just one step can save hours when batching on this scale. so please let me know your thoughts. I have no coding, scripting know how what so ever, but if there is a easy way to implement one to improve the speed of this process I’m game.

  • edited February 2019

    Next I use Bbedit and drag my folder of presets/patches that I’m going to record into it from the massive folder located in the documents folder.

    That is brilliant. Had no idea bbedit did that.

    No, I don't think I have any suggestions to make this faster—sounds like you have a really well oiled process in place! I'd email sample robot about the lack of batch export through—that seems an odd omission.

    If each OBS file is basically the same—all use three layers of samples, all samples have the same note suffixes—there might be a way to script it all but would probably take as long to set it up as it takes to just do what you're doing!

    I'm thinking it would look something this (again, prolly not worth the time but thinking out loud anyway...)

    1. Create a 'master template' OBS file with the conventions used by all. Use simple-to-find-and-replace sample names for this one.
    2. Then, using something like Python or AppleScript or SH or whatever...
    3. Scan your Bbedit generated file listing and for each line
    4. Create a copy of the 'master' OBS file named after that line
    5. Within the new OBS file find the old known sample names and replace them with the new sample names (which it sounds like you ca get from the text file too)
    6. Move the OBS file to where ever it's supposed to be, relative to the samples (or vice-versa)


    If it works, you should be able to skip the obsidian automap, the numerical saving, the webdav renaming...

    Seems to me that the "three folders for three layers" is a really handy convention but presuming there was some other naming convention that you could find and replace, you might be able skip that step with a script.

  • edited February 2019

    Wow! Hardcore! Muchos respect!

    It would certainly be nice to have a big old library like that. For me I am just using Renoise as a sort of VST sound module (midi out of NS2) and when I have something going that I like I just use Renoise’s 'Render plugin to Instrument' option. It is nice being able to hear the vst, then just switch the midi channel and then hear the rendered out version to compare.

  • T4HT4H
    edited February 2019

    @Will said:

    Next I use Bbedit and drag my folder of presets/patches that I’m going to record into it from the massive folder located in the documents folder.

    That is brilliant. Had no idea bbedit did that.

    No, I don't think I have any suggestions to make this faster—sounds like you have a really well oiled process in place! I'd email sample robot about the lack of batch export through—that seems an odd omission.

    Yes, Bbedit is coming to the rescue in a major way for me on this, as is SampleRobot being able to import the text file and run with it. It's a beautiful thing indeed. With the exception of this lack of batch export which I agree, makes absolutely no sense...especially when you can batch record, batch save, etc, why can't I Batch Export!? And before spending the $300 on the software you'll never find this omission since you really can't export meaningfully in the demo. I would've taken the stripped down $100 version for my needs, but they showed a near future update with direct ITB VST recording at NAMM on the Pro version, which required no midi routing and didn't require a standalone Plug in. So I went for it since I'm going for such a high volume of work, with high value sample libraries I've spent years building, and fully committing to NS2, which I think is THE tool for future producers, as long as we get a couple of the promised updates.

    I'm thinking it would look something this (again, prolly not worth the time but thinking out loud anyway...)

    1. Create a 'master template' OBS file with the conventions used by all. Use simple-to-find-and-replace sample names for this one.
    2. Then, using something like Python or AppleScript or SH or whatever...
    3. Scan your Bbedit generated file listing and for each line
    4. Create a copy of the 'master' OBS file named after that line
    5. Within the new OBS file find the old known sample names and replace them with the new sample names (which it sounds like you ca get from the text file too)
    6. Move the OBS file to where ever it's supposed to be, relative to the samples (or vice-versa)


    If it works, you should be able to skip the obsidian automap, the numerical saving, the webdav renaming...

    Nice, I just did a variation of this using 'Araxis' which lets you batch import text files to do find and replace. First i simply duplicated the 'master .obs' file 127 times which obviously had the same 'find' line in it for the find and replace, and went through it one by one in araxis replacing the same text with each different file name for the .wav. It took as long if not longer and the items weren't tagged so thats another added process now for later within obsidian, but overall no biggie since i've got them named by category as a redundancy, for example BA Bassline 1, PA Pad 1, LD Lead 1, etc. However, it was easier on my hands due to being able to use track pads and shortcuts on physical keys so thats a plus. If only I could take the process in Araxis and have it replace automatically based on a text file list of patches 1-128, I'd be golden!

  • @Audiogus said:
    Wow! Hardcore! Muchos respect!

    It would certainly be nice to have a big old library like that. For me I am just using Renoise as a sort of VST sound module (midi out of NS2) and when I have something going that I like I just use Renoise’s 'Render plugin to Instrument' option. It is nice being able to hear the vst, then just switch the midi channel and then hear the rendered out version to compare.

    Thanks! That sounds pretty cool and I've been hearing some good things about Renoise so I'm definitely gonna have to check it out as soon as I find the time

  • T4HT4H
    edited February 2019

    post deleted.

  • T4HT4H
    edited February 2019

    @Will I figured out a solution for the goal from my previous post, which was to change the names of the .wav files inside the .obs text files in a batch on the desktop instead of manually making obsidian patches one by one from the .wav automapping on the iPad. To do this I used WebDAV, and made a copy of the 127 .obs text files from one of the soundkits (with 127 patches) I’ve imported from earlier and used those as a master to create another sound kit (meaning a kit that is multiple obsidian patches/presets) from it collectively, and transferred that to the desktop. Let’s call that ‘kit A’. I retained the original text file I mentioned in earlier posts with the list of sounds I used for SampleRobot of ‘Kit A’ and drag and drop that text file into ‘Numbers’ to make a it spreadsheet list. This will from now on be our master file from which all other kits are made in conjunction with those 127 .obs files of ‘Kit A’ that I made a copy of.

    Next I’ve done all my sampling and .wav exporting for our new kit, lets call it ‘Kit B’ in Sample Robot. During the batch import I used a text file listing all the sounds specific to ‘Kit B’ to auto name it’s patches, just like i had done for ‘Kit A’. Now I take ‘Kit B’s’ text file and also import it into ‘Numbers’ into the same file I have kit A’s sounds listed in. The patches names (which are the root name of every .wav file) from Kit A and Kit B are now side by side in 2 columns. I export this as a .csv

    Next I use ‘String Replacer’ which I downloaded from the Mac App Store. In the ‘sources’ tab I Drag and Drop all 127 .obs files that are my “master” files (making sure I’ve made a copy of them for future use!). Next In the ‘Strings’ tab I import the .csv file. It will ask which column is for ‘find’ (Kit A) and which column is for ‘replace’ (Kit B ). Once everything is set up I hit Replace all.

    Now all the .wav files inside the .obs text files that were set up for ‘Kit A’ have been renamed to the patch names of the Kit B! I then place all the newly changed .obs files in the 127 folders for ‘Kit B’ I made earlier in the previous posts. Now I use a batch renamer to add .obs at the end of all 127 folders that contain the .wav and the new .obs text file that corresponds to it.

    Then I airdrop or WebDAV those folders into obsidians user folder folder and it’s done. 127 patches made without any manual labor on the iPad. This can be done endlessly as long as you keep the original master files for Kit A (.obs and the original patch list that the .wav files were named from that you import into numbers) unchanged and only edit copies of them.

  • i really admire your commitment and effort you did, and also that you're sharing all those valuable informations .. looks like at the end you reached almost perfect workflow, which is great !

    Btw. looking forward to some of you production with all those prepared patches ! Sounds like tons of inspiration under you hands after you finish this patch collecting story.

  • @T4H said:
    Then I airdrop or WebDAV those folders into obsidians user folder folder and it’s done. 127 patches made without any manual labor on the iPad. This can be done endlessly as long as you keep the original master files for Kit A (.obs and the original patch list that the .wav files were named from that you import into numbers) unchanged and only edit copies of them.

    Epic. Way to go!

  • T4HT4H
    edited March 2019

    Here's a video outlining the process of migrating NI Massive Patches over to Obsidian. Im thinking of making a video series out of it with different VST's and workflows. It's currently unlisted as I want to get feedback from the forum beforehand, and make the necessary edits or changes before posting it officially. Please let me know your thoughts. Thanks!

  • Really great video ! Very informative and very professionally made ! Just publish it !

    I'm totally amazed by this workflow you developed.. complex, deep, pro with very pro level result ..
    :+1::+1::+1:

  • Thanks 😃 @dendy ! Quick question, does the information in the video go by to quickly? Should I slow down the pace in other videos going forward or does that speed/pace work? Thanks

  • edited March 2019

    hm.. yes, especially for beginers there is a lot information so if you can slow it down a bit - maybe it would be good idea..

    But it is not critical, so if it's too much work then do not worry about that - people will use stop and rewind :-))

    Good idea would be to add also links to used apps in Video description..

  • @T4H said:
    Thanks 😃 @dendy ! Quick question, does the information in the video go by to quickly? Should I slow down the pace in other videos going forward or does that speed/pace work? Thanks

    The voiceover is well done but it is very fast. She needs to take a breath once in a while! Good job overall though. That's a lot of information. Links to apps is a good idea per @dendy

  • T4HT4H
    edited March 2019

    @dendy Links have been added!

    @anickt said:

    @T4H said:
    Thanks 😃 @dendy ! Quick question, does the information in the video go by to quickly? Should I slow down the pace in other videos going forward or does that speed/pace work? Thanks

    The voiceover is well done but it is very fast. She needs to take a breath once in a while! Good job overall though. That's a lot of information. Links to apps is a good idea per @dendy

    LOL, noted. Thanks!

Sign In or Register to comment.