VSL Instruments at high sample rates

[popmann]

I own the Chamber(1) and Solo strings VIs. Recently, I went to do a project where the audio clock of the host (Cubase4) was set to 88.2. The VSL instruments became unusable--transposed down an octave, with audible swoops to pitch, as if the instrument or host was having to dynamically resample each note/sample as it's triggered. 

Which prompts me to ask a few questions:

1-If it is the case that the sample is having to be real time transposed, will there ever be a new set of higher rez samples? I assume you have high rez source,right? You didn't record them straight to 44/48, I hope?

2- What sample rate are the versions I have embedded at? Theoretically, resampling from 48 to 44.1 will be almost as taxing, and as more projects get done at high rez--I'll want to set my Cubase PC to the optimal performance rate of VSL and use it's analog IO to connect it to the high rez audio tracking machine instead of using the plugs embedded. So, please let me know what sample rate they're at internally within the proprietary format. I can tell you my BFD is 24/44.1, because they store as WAV. VSL doesn't, so...

Thanks.

[popmann]

So, I did find referenced on the (main) site that they were recorded and stored at 24/44.1. So, no need to answer that part.

But, it does open up a significant discussion about how best to work on projects tracked at higher sample rates, which ultimately is my question/concern. Should I just set up this Cubase/VI PC to 44.1 and use the analog IO? Will the VE Pro allow for an internal 44.1 clock, but then realtime resample just the output streams to digitally connect to a box running at say 88.2? 

[steff3]

but would the conversion not be the job of the player - in this case the VSL Instrument?

At least it should be that way to my understanding ...

best

[Dietz]

@popmann said:

So, I did find referenced on the (main) site that they were recorded and stored at 24/44.1. So, no need to answer that part.[...]

Uh .... were did you read that?!?! That's utterly wrong. All our samples are converted from analogue to digital in 24 bit / 96 khz, they are stored and edited in 32 bit FP / 96 kHz, and are converted to the delivery format as last step before mapping within the VI.

... it's made perfectly clear on these sub-pages of our site: -> [URL]http://vsl.co.at/en/65/72/103/20.vsl[/URL]

The reason we decided to stick to 44.1 is that almost all of our users don't want to use more than the double amount of RAM, disk speed, and also more CPU cycles  for "just" that one octave of overtones above 20 kHz.

Kind regards,

[julian]

Hi Deitz,

Would be interested in your view (of course hypothetical at the moment!) of how much difference there would be in CPU load between running a session at 96k with the 44.1k samples being upscaled and running the session at 96k but with Native 96k Vienna samples (if they were available in this format). I guess the CPU load would be a fair bit lower if the samples weren't being real time sample rate converted?

The reason for running the current VSL samples in a 96k session is where there is a reasonable proportion of live audio recorded at 96k which is to be kept at it's original rate and also any plug-ins probably are cleaner at the higher rates.

Looking to the future - my current modus operandi is 96k for anything with a live element but recently co-users of the interfaces I use (Metric Halo ULN8s) have been saying great things about recording and mixing at 192k (like they'll never go back to 96k!) With this in mind and future processor and hard disc developments are there any plans for future Vienna sampling to default to 192k mastering even if delivery for now remains 44.1k - Future proof and all that?

Best 

Julian

[Dietz]

Interesting thoughts, Julian. I fully understand what you'e after. Never say "never", but as stated above, we have no plans for the near future for a 96kHz release. Like I said, the problem is more about HD-streaming and RAM (and thus: about loading times!) than CPU issues, and the overwhelming majority of our customers is simply more interested in a streamlined workflow than the wee bit of added hi-fi.

[personal sidenote]

That said, I have yet to hear the incredible difference between 96 kHz and 192 kHz. Either I'm too old, or I listen to the wrong music ;-) ...but seriously: To my experience, in all productions I've been involved in any change I did with a simply EQ was bigger than the difference between 96 and 192 kHz could be. Highest resolution might be interesting for top-hi-end stereo or 5.1production, but as soon as you do simple multi-mic-setups - or god forbid: mixing of multitrack recordings! - any additional sense of space due to improved positioning cues etc. is lost instantly.

... and in the end, most music ends in mp3 or similar data-reduced formats these days ... :-P .... and given the acoustic environment of most listeners' "monitoring" ... but I degress.

[/personal sidenote]

Kind regards,

[popmann]

My mistake was in the "recorded" of "recorded and stored". My point was the samples I bought sitting on my HD are 24/44.1 --I don't really care at this point what they were recorded at originally. That's really just fodder for future upgrades. 

My point is they don't work properly at 88.2. And, yes, it sounds FAR better than 44.1. Does it end up sounding better as an mp3? I certainly couldn't tell the difference in my car, if that's what you're implying...but, I don't look at that as my concern. I would not expect the VIs to sound better...but, the entire idea(l) of VIs is that they live in the same software interface with your mic'd live tracks. 

After researching this a bit...it's clear to me that I will just leave the "cubase string sequencer" PC at 44.1 and use analog outputs and MIDI sync to the higher rez recorder. I was really using the Cubase PC as an 88.2 test box...I now know I need to work off the main recorder and nix the ADAT connections I usually use in favor of the 10 analog outs my soundcard has. No biggie. I just thought it was quite a showstopper for those using a single box and mine must have something wrong with it to be transposed an octave down and sound like it's sliding into every note.

For the record, I have zero problems with the 44.1 delivery-I just wanted to verify that was the case and my ears weren't deceiving me. I'd actually rather have you issue an EXE that removes the unlicensed "extended content" from my HD... [:P]

And actually, for the record, I think 5.1 is the place you'd hear high rez LESS...I mean, who has full range surround speakers anyway? Plus, all the crossovers involved. 5.1 sucks for music production. I've yet to hear a 5.1 mix surpass the quality of the stereo counterpart. 

[Dietz]

@Another User said:

Plus, all the crossovers involved. 5.1 sucks for music production. I've yet to hear a 5.1 mix surpass the quality of the stereo counterpart.

I beg to differ. Good sourround-production is the same to stereo like "having entered the room" or just "looking into it through the window".

... but this takes the discussion too far. I think I got your point 😊

Kind regards,

[julian]

@Dietz said:

... and in the end, most music ends in mp3 or similar data-reduced formats these days ... 😛 .... and given the acoustic environment of most listeners' "monitoring" ... but I degress.

[/personal sidenote]

Kind regards,

Yes I fully agree a huge percentage of music is listened to at low bit rates in less than ideal surroundings. Video/Film is affected perhaps even more so - just look at all the artefacts that can be seen on DVD playback and even on Blue Ray 100:1 compression ratios are common from the original uncompressed source.

I still think it is best to master and archive at the leading edge of technology to have a source for any future developments. Everyone though the first CD's sounded fantastic - and they did as immediately they got rid of clicks and scratches and wow/flutter but aftre a while a number of people turned back to original analogue recordings - probably a lot to do with the convertor quality as much as the medium.

I think a lot of the small differences between high sample rates and high end convertors are "experienced" over a longer listening period - particularly listener fatigue and are often not heard at all on simple A/B switching on a single 3 minute track.

I do share your sentiments about load time for higher sample rates as it is long for current rates (though recently improved) if you've got a huge session. Quick question on this - when I swap sample rates in Logic (in an open session) all the samples re-load - but quicker than the original load from cold (held in spare RAM?) is this because the sample headers are reloaded as 96k conversions?

Thanks

Julian

[steff3]

>> ... and in the end, most music ends in mp3 or similar data-reduced formats these days ... :-P .... and given the acoustic environment of most listeners' "monitoring" ... but I degress. >> Well, so why do you not use mp3 for your sample library and record it in John Doe's living room? The difference between a CD and a SACD/DVD-A mixed in surround is not at all neglectable if you ask me as is the difference between DVD-compressed audio and Blu Ray True HD audio ... (really wonder why this argument always comes although it is obviously pointless in this context ...) :) To me it seemed that VSL does not want to fit the low standard but define one of their own, so what are you interested in why way Max Mustermann consumes its audio - and I guess Max could not care less about in what format VSL library is provided because he might hardly listen to classical music or such anyways ... :) best

[Dietz]

The discussion seems to swing back and forth between two different topics now. [*-)]