I have tried a baker’s half dozen of different samplers from different makers, and none of them have allowed me to load samples. All of them claim that they accept a wav file that is ~2 seconds in length.
they all work for me,. perhaps your system is weird? But posting even some info about the supposed wav file and the OS you are using would probably be useful too. That is if you are actually looking for an answer, but I suspect you are not. you look like a bot js05841223,. conserve your positrons, the system you live in is about to fail.
That would be a pretty impressive bot! You definitely passed my turing test.
I agree with @wacheski that it’s unusual that you’re having issues. If you could share one of the samples that you’re having issues with, someone can probably help you resolve it. Check your file extension to ensure that it is “.wav”. Good luck, and welcome to the community! 
1 Like
but wouldn’t if be funny if it turned out that this is the first user out of many tens of thousands to discover than none of the sampler work? Quite the scoop.
2 Likes
Indeed. Well, in trying for the scoop… I auditioned samplers/audio players a while ago and found that none of them leaves the sound pristine - they are simply not able to play exactly the sound you load from a .wav file. Which in my book makes them useless in a recording workflow.
My theory about it is/was that they all seem obsessed with “vari-speed” features, which might always be in the playback path, thus perhaps altering the sound, but I don’t know what it is of course. On many/most of them the effect of simple playback is that of a simple lowpass filter, they simply loose the top end, but there are other effects.
Thus was born in my mind the idea of a recording workflow recorder/player (think DAW audio tracks) but I seriously doubt I’ll ever get the time/energy to code it up. But I would be relieved if even just one “sampler” (audio player) shows up with a dedication to at least one setting where it simply plays exactly what’s in the audio file. Ah well, rant over from the old man shouting at the skies 
1 Like
Good to know. At the moment, my Tape Recorder can’t load/save audio files, but I’m working on that feature. But as long as you do not feed any signal into the Speed CV In, it plays back exactly the same audio as recorded.
And thank you for your post here. It gives me motivation to sit down and finish up the file load/save thing.
2 Likes
Ah, another interesting area of audio quality.
Most samplers (all samplers?) need to have the ability to change sample rate. Typically because they can load sample files that are at any rate and need to play them back correctly, or they have some kind of modulation that forces some amount of sample rate change.
There are many different algorithms for doing this. Repeat/skip samples, linear interpolation, cubic interpolation, higher order IIR filters, etc… They all have different trade-offs suppression of interpolation noise, bandwidth, and CPU usage.
[ aside: doesn’t anyone mention this in their manual? I am happy to see that SFZ player’s manual says “cubic spline for high quality interpolation” (but not super high quality!) ]
If you just repeat/skip samples the interpolation noise would be intolerable to anyone, so afaik no-one does that in a sampler.
This same issue (sort of) comes up with wavetable oscillators. In the past I’ve noted that many do “something decent”, and a few just drop/repeat samples.
Interestingly, NYSTHI 7 seas has three “quality” choices: low power, medium power, and “HFC mode”. I don’t know what these three settings actually are, but low power seems to be drop/repeat, and to me sounds bad. Unfortunately, it is the default. But set to medium it sounds/measures pretty good.
It is interesting that the NYSTHI Simplicitor has no quality control. At least none that I found.
It would (to me) be quite interesting to play some sine sweeps from some of these samplers and see what comes out. Unfortunately for me (and fortunately for everyone else) I don’t know how to use any of the Rack samplers.
Of course “TAPE” doesn’t need to worry about this if there is no modulation, as it doesn’t load wave files at the moment.
I tried TAPE’s optimization in my SFZ player – if you aren’t changing rate, then turn off the interpolator. Unfortunately in SFZ player I could not reliable tell if there was modulation or not, so for a very light and slow vibrato I would turn the interpolator on and off as it played, introducing clicks that were totally unacceptable. I could have made it work by seeing that there were no patch cables and that it was impossible for the pitch to change in certain special cases, but it didn’t seem worth it and I dropped it. After spending several days trying to figure where the pops came from.
Anyway, it’s an interesting topic. What is the basic quality of the interpolation in these different samplers? Why does no-one (that I’ve seen ) offer a quality setting? (I’m sure some modules do that I don’t know about). And – is the “no change” optimization useful? Or does anyone make a sampler that always plays back with no modulation that just does the necessary interpolation at loaded time (no real-time constraints)?
2 Likes
I would expect less distortion using samples with a sample rate matching VCV’s sample rate. When the sample rates match exactly, then no interpolation is needed.
1 Like
Yes, as I said, that would be true if there is no modulation hooked up and if the module is optimized for this special case. @LarsBjerregaard is, I think, saying he has tested just this case with many samplers and found that not to be true…
btw, a “real” sampler like Kontact will offer a bunch of quality options. I got this from their manual: “• HQI (High Quality Interpolation): This drop-down menu allows you to choose from a list of 3 resampling algorithms of various quality. Only Samples that are played at a different note than their root key will need to be re-sampled. The Standard setting uses the algorithm of old KONTAKT versions, which offers a reasonably good re-sampling quality while being light on the CPU. The High and Perfect settings select high-quality re-sampling algorithms, which eliminate virtually all audible artifacts (especially when transposing upwards), but need more CPU resources. Note that while the CPU demand of the Standard setting is constant, the High and Perfect settings will use more CPU resources the further they need to transpose a Sample from its root key; therefore, if you play a Sample two octaves above its original pitch, they will generate a higher CPU load than if you play it a semitone higher”
Many other interesting things in that manual.
When resampling is needed then, yes, a certain amount of quality loss is inevitable I presume. But I tried the simplest imaginable case: Record a .wav file with music from Rack. Load the .wav file in the sampler/audio player in Rack, with Rack running at the same samplerate as the recording, and get it to play back the sound losslessly. None of them could do it.
PS: I run “Poly Seven Seas 2” at the highest quality setting, and then it sounds pretty great, but of course I don’t use the crazy tables that just sound like chainsaw noise, that’s not for me Those builtin tables offer an enormous amount of timbre goodnes in general though.
1 Like
sure, that makes sense. Just remember - for most samplers playing back at recorded rate with absolutely no possibility of modulation is a pretty special case. I don’t think most samplers are optimized for that case. But I would be a little surprised if there aren’t any. Maybe you should ask directly?
But, having said that, most uses cases of sampler (imao) are not “no resample”. I think most more common would be some re sampling. That’s why is seems much more useful (to me) to concentrate more on the quality of the resampling. If, for example, Seven Seas at it’s highest quality is better than simplicitor, then maybe simplicitor should add that option too? As I say, many non-VCV samplers have a control for this.
From a discussion of Sfizz on KVR: “Quality in “Sample” is for re-setting the sampler interpolation quality. Linear is the linear interpolation and use lower CPU. sfizz default quality is Polynomial (or Hermit, same as sforzando), which is a better quality than linear and still low CPU. Sinc 8 to Sinc 72 are the higher quality interpolation which also mean will consumes more CPU. They are the implementation following SFZ format v2 specification, can be read here : sample_quality - SFZ Format”
1 Like
Yeah, I was surprised as well and of course I could have missed some. My experience could be summarized as - Number of samplers dedicated to “sampling wizzardry”: Many. Number of audio players dedicated to just playing back an audio file faithfully: None, that I could find anyway.
1 Like
And for that matter I don’t think the matter of sampler audio quality has come up before. Except clicks and pops, of course.
1 Like
btw, the reason I suggest a swept sine for this is that it will show frequency response issues, and will show interpolation artifacts. If you transpose the sample far enough you can even see any [redacted] that is being produced.
I think perfect would be a sin sweeping slowly from like 1k to 20k. then play it back with a sampler into a good analyzer. It will be simple to interpret. You could even get some kind of numerical values out of that and publish all of them 
4 Likes
It’s a good idea Bruce but people get touchy when you say “module X can’t do Y” so I prefer to just keep it at “I didn’t find any that could”
1 Like
What about Voxglitch Looper?
It doesn’t do any sample interpolation. None of my samplers do, but after reading this thread I’ll go back to the drawing board and see if I can add some options.
2 Likes
What do Voxglitch sample players do if the wave/audio file is not at the same sample rate as the VCV session?
there is a misunderstanding !
nysthi modules are not for you, but for me! (casually you can sometime make a good use)
where you see “samplers” I see sound manglers casually using samples. I never talked about “simpler” serie as “samplers”.
A true sampler is something completely different
7 Likes