TAL Sampler - discussion

[chibchan]

Well then
+1 for TAL Discord

[EmptyVessel]

I'll give it some thought, it might be fun. It just needs to be more than, like, 6 of us to be worthwhile and it's not actually that easy to let the majority of TAL users know about stuff.
I'd love to build an active community like the Kilohearts Discord, that sh t is crazy but I don't know if it's possible and I don't have a huge amount of time to put into tirelessly building it.
I think it's fair to say Patrick doesn't really have the time or inclination to be there much, or at all, and a lot of what's keeping the KHs server engaged and active is the pretty much constant presence of the KHs team.
Not a criticism of Patrick in the slightest, I can see why people are a bit fatigued with social media of any sort and just would rather go outside. I feel a bit that way myself for all the ability to connect with like-minded people from all over the world is deeply cool.
Sorry. What a rambly comment. thx u for coming to my TED talk about internets.

[chibchan]

Maybe it's best to keep things simple and just report bugs to Patrick and post them here as a new topic. I didn't realize the complexity of it all and we certainly don't want to add any more of it on to others.

+1 tal-software @kvraudio

[pummel]

I really liked version 3.6.2. Yet relative to that, version 4.5.2 includes such a huge feature set improvement I was blown away. I really appreciate all the effort that went into the added features. The only thing I really miss from 3.6.2 is the ability to auto-load / audition presets quickly with the up/down arrow keys (versus mouse double click to load).

In 4.5.2, what does the "L" button do (to the right of the Service control button)?

[pummel]

Here is some info on TAL sampler v452 color schemes that is not in covered the manual
If you are inclined to create your own color scheme, I hope this saves you some time!

On windows PC, the TAL color scheme files are saved in "documents"

The eight color boxes change the following aspects:

frame_background, frame_text, frame_outline, sample_background
frame_labels, knob_rings, sample_foreground, knob_centers

The color box 6-digit codes can be entered via keyboard (LMB double-click to enter)

Or, the eight color box codes can be entered manually in text file
0xff<color1>;0xff<color2>;0xff<color3>; ... ;0xff<color7>;0xff<color8>

Here is an example file for you to try
filename: warmGoldv452.talsamplercolor
file contents:
0xff272525;0xffffd6a6;0xffa6e184;0xffd84040;0xffbcaf72;0xffffc47e;0xff252731;0xff7b6c4f

[DuX]

I have just found out there is a reasonably big subsonic (~0-40Hz) DC offset in TAL-Sampler 4, which is also present in TAL-Sampler 3. It's unbelievable that it remained unnoticed for all these years. It can be removed by DC offset removal plugin, or a high pass filter set to filter out very low frequencies. I am using Airwindows' "Infrasonic" and it does it perfectly without affecting low frequencies.

While at it, I checked all other TAL plugins, and as it turned out there are other TAL plugins which have DC offset on their outputs and need a DC offset filter on the output. TAL J8 has the biggest DC offset of them all, and TAL-Noisemaker and TAL-Dub-X have a bit lower DC offset same or similar to TAL-Samplers. All other TALs are fine.

This is good to keep in mind when mixing with these plugins as DC offset is something to avoid.

I am using Linux VST2 versions of TALs, but I don't think that matters as code is code. I will check VST3 versions of Sampler, J8, Noisemaker, and Dub-X to see if its present there. If I don't edit this post, it means they behave the same.

Cheers!

[imrae]

Is that DC in the input sample, or appearing elsewhere?

I imagine the psytrance crowd etc. probably prefer their sampler to not mess with the original phase if possible.

[DuX]

It appears on the output as soon as you load TAL-Sampler, either 3 or 4. This DC offset is not so bad at -136dBFS as I made it to look, but if you tend to process its output further with compressors, limiters etc. it could make its way up into the more audible territory, so it's best to put some DC offset remover on Sampler's output. Airwindows' "Infrasonic" is perfect! If you're using more audio outputs from Sampler it should be on each output.

Now, I've spent a couple of hours testing all TAL plugins, and it's not as bad when it's so low at -136dB like it is with TAL-Sampler, but TAL-Drum doesn't have it at all. Maybe if I used more of its FX... these are usually the culprit.

What I've also found out is that both in J-8 and U-NO-LX-V2 the DC offset is caused by Juno Chorus, and it's far more significant than in TAL-Sampler at ~ -90dBFS. Turn off the Chorus and DC offset disappears. Interestingly, TAL-Chorus-LX doesn't produce any DC offset, so you can use it instead. I tested both VST2 and VST3 versions and they both exhibit same DC offset. The DC offset in Dub-X is a story on it's own. They are different in mode X, I and II. X is the worst at around -100dBFS, followed by mode II at -114dBFS, and I doesn't produce any DC offset.

What is important to keep in mind is that Sampler's DC offset is at -140dBFS, that's like dither level offset. It's nothing significant. -90dBFS is. However, as I said if you boost the signal by significant amount of dBs it could creep in into the audible territory and... well, imagine what 10hz sounds like if boosted enough. You would need to boost it by 60-70dB to break something, though. It is rare to find yourself needing to boost something by more than 10-20dB, at least I don't have to.

Nevertheless, I'll be putting Infrasonic by default after Sampler, J-8, Noisemaker and Dub-X, just so I don't have to worry about it later when mixing.

Cheers!

[benedettodue]

In the TAL Sampler software there is a maximum polyphony limit of 12 voices: considering that most "sample players" have polyphonies of 64-128 voices, is it possible to increase this value or are there any technical reasons?

[mmccarthy]

The only thing I really miss from 3.6.2 is the ability to auto-load / audition presets quickly with the up/down arrow keys (versus mouse double click to load).

Big +1 on this

Updated from a 2.x version to the most recent 4.x version there and it took a while to figure out how to load presets. It was also nice that once you got to the last preset in a folder, clicking next would go to the first preset in the next folder. It feels like a bit of a step back IMO.

Though I've just started using it for the first time today, so it may well be just that I'm not used to it yet.

[rens]

A question for the group-

I am really liking TAL sampler (new user as of last month) but I'd like to dedicate a bit of hardware to it so that it can sit with my hardware synths, etc.

Does anyone have advice on a smallish rig specifically to recreate some of the 'vibe' of a rack sampler with TAL?

I was thinking maybe getting an old surface pro to run linux on with a 2x2 interface or the like.

anyone doing this?

[AtomOfScent]

The concept of TAL Sampler is great.
I wouldn't want (or expect) it to turn into something like Kontakt, though after using it for a while I've made a list of some usability enhancements and additional functionality I think could expand what it's capable of without sacrificing immediacy and use of use.

• Presets:
• Favorite presets, not just folders. Users will have far more presets than folders so favorite presets is much more useful.
• Locking for each layer when changing presets. Modifier-click on presets to load only audio sources, only keep audio sources (load everything else.)
• Copy audio checkbox in preset save dialog to save samples in subfolder.
  (Embed in project can cause a lot of bloat with project backup/autosave files.)
• Option to save preset samples compressed as FLAC.
• Indicate preset has been edited with * in name.
• Don't show extension in save dialog. It's more cumbersome when clicking on existing filenames and tweaking the name.
• Lock/presets for granular.
  
  Effects/Filters:
• Filter 2..
• Wet/dry and stereo width for chorus (like LX.) I'm a fan of Chorus-LX, but never fully wet. Many times I've wanted to add some chorus, but I never want all chorus.
• Output gain for FX. They can change the overall volume and cause clipping or other gain-staging issues with the other effects.
• Convolution effect, for loading reverb, speaker emulations, sound design etc.
• Send FX (from ALL layer tab.)
• Reorder FX
• Global EQ: Add LP, selectable slope for HP/LP
• Yamaha VSS-30 and/or Casio SK DAC emulation. The VSS-30 has a cool U-Turn playback mode, a great "echo", and useful AM too, (FM not so much.)
• Sine shape for LFOs (something imperfect and slightly drifty of course.)
  
  Layers:
• User ROM folder to add custom simple waveforms.
• Solo layer in individual tabs. Modifier-click on/off to solo in ALL tab.
• Layers to individual outs.
• Snap-X option in layer view.
• Zoom/finer control over velocity scaling, limit highest/lowest vel (like Omni2 or Pianoteq)
• Automap root key from file name(s). Drag and drop multisamples, map by filenames.
  
  Bugs:
• The 6dB filter sometimes defaults to LP at 0.0Hz or the BP defaults to 22kHz, so no audible output.

[mcbpete]

[*]Yamaha VSS-30 and/or Casio SK DAC emulation

This is definitely something I'd personally love to see developed over everything else. Not just these but any sampler emulation - we've got the Emu II, AM6070, S1000 but imagine if it had all these too (from Chipcrusher documentation):

Simple Encodings
-PCM (Linear Pulse-Code Modulation): Examples: Compact Disc (if 16-bit/44.1kHz), E-Mu SP12 (12-bit/27.5kHz) E-Mu SP1200 (12-bit/26.04kHz) <- Already present in TAL-Sampler
-FPCM (Floating-point pulse-code modulation): Example: The YM301X series DAC used for Yamaha’s 4OP FM synthesizers encoded audio on 10.3-bits.
-DPCM (Differential pulse-code modulation): Example: Delta modulation channel provided by the RP2A03/RP2A07 CPUs used in the Big N’s 8-bit console uses Bits=6 and Parameter=1

Adaptive Encodings
-BRR (Bit Rate Reduction) Example: Specific ADPCM Encoding used in the S-DSP of the SNES.
-YADPCM : Examples: N’s ‘wii-mote’ and various Yamaha FM chips for sample playback.
-HAAR Transform: A unique blend of triangle waveforms.
-A-Law and μ-Law: Examples: Early US samplers and drum machines which used AM6070/AM6072 ICs like the LINN Drum machines, and the E-MU Emulator I/II/Drumulator. <- Already present in TAL-Sampler
-DIALOGIC: Example: MSM(OKI)6295 Used in LOTS of late 80′s and early 90′s arcade titles (All non Q-sound CPS1 titles)
-CVSD (continuously variable slope delta modulation): Example: Suzanne Ciani's voice in the Pinball classic 'Xenon'

1-bit Density Encodings
-PWM (Pulse-width modulation): Examples: Some legacy software on the Apple II SoftDAC (3 and 4-bit versions) and DAC522. You will find that they are NASTY sounding. You have been warned.
-PDM (Pulse-density modulation): This is usually run at VERY high sample rates, SACD/DSD uses 2.8224 MHz, but since here we only have straight PCM for input, whose rate depends on the host rate (typically 44.1/48/96kHz), it is quite useless to use PDM at higher frequency, since we don’t have extra information to encode in time.

Vocal Encodings
-LPC10: Example: Speak & Spell
-MOZER: Example: Berzerk, Impossible Mission
-SP0256: Example: Intellivoice, Various speech modules for vintage computers.

Imagine a big drop down menu split into something like:
Professional Samplers
Sampling Keyboards
Computers / Consoles
Toys

and each of those with sub sections of brands (e.g. Akai, Ensoniq, E-Mu Systems, etc.).

TAL-Sampler wouldn't just be a sampler, it'd be ALL the samplers (Hey, I can dream !)

[EmptyVessel]

The concept of TAL Sampler is great.
I wouldn't want (or expect) it to turn into something like Kontakt, though after using it for a while I've made a list of some usability enhancements and additional functionality I think could expand what it's capable of without sacrificing immediacy and use of use.

• Presets:
• Favorite presets, not just folders. Users will have far more presets than folders so favorite presets is much more useful.
• Locking for each layer when changing presets. Modifier-click on presets to load only audio sources, only keep audio sources (load everything else.)
• Copy audio checkbox in preset save dialog to save samples in subfolder.
  (Embed in project can cause a lot of bloat with project backup/autosave files.)
• Option to save preset samples compressed as FLAC.
• Indicate preset has been edited with * in name.
• Don't show extension in save dialog. It's more cumbersome when clicking on existing filenames and tweaking the name.
• Lock/presets for granular.
  
  Effects/Filters:
• Filter 2..
• Wet/dry and stereo width for chorus (like LX.) I'm a fan of Chorus-LX, but never fully wet. Many times I've wanted to add some chorus, but I never want all chorus.
• Output gain for FX. They can change the overall volume and cause clipping or other gain-staging issues with the other effects.
• Convolution effect, for loading reverb, speaker emulations, sound design etc.
• Send FX (from ALL layer tab.)
• Reorder FX
• Global EQ: Add LP, selectable slope for HP/LP
• Yamaha VSS-30 and/or Casio SK DAC emulation. The VSS-30 has a cool U-Turn playback mode, a great "echo", and useful AM too, (FM not so much.)
• Sine shape for LFOs (something imperfect and slightly drifty of course.)
  
  Layers:
• User ROM folder to add custom simple waveforms.
• Solo layer in individual tabs. Modifier-click on/off to solo in ALL tab.
• Layers to individual outs.
• Snap-X option in layer view.
• Zoom/finer control over velocity scaling, limit highest/lowest vel (like Omni2 or Pianoteq)
• Automap root key from file name(s). Drag and drop multisamples, map by filenames.
  
  Bugs:
• The 6dB filter sometimes defaults to LP at 0.0Hz or the BP defaults to 22kHz, so no audible output.

Some great suggestions here, thank you. I've suggested some of these in the past but it won't hurt for me to pass on this full list to Patrick. I think there's plenty of room to add more flexibility without making it feel bloated or that we're chasing Kontakt in any way. Particularly leaning into the vintage aspect and adding more DAC emulations I think would be great and make total sense.

The filter bug I reported previously, I'm surprised that's still the same, I'll check up on that.

Big thanks for that

[audiojunkie]

Rather than more AC emulations, I'd rather have an ultra high quality interpolation option--Sinc 128 would be nice! I'm OK with the other options as well, but I'd really like the ultra high option. Since we aren't competing with Kontakt and we aren't trying giant libraries, at least we could make the most of what we've got. I don't think a ultra high quality interpolation option is out of the scope for users. We really need this. I'm sick of being reliant on DiscoDSP's Bliss for higher quality interpolation. Also, I use Linux. An Ultra High option is needed in the Linux realm by a quality developer. Please, emptyvessel, encourage this request.

EDIT: Imagine only needing one sample per octave to achieve the sound quality that it would normally require a sample every third note. If that isn't making things easier and putting the fun back into sampling, I don't know what is!