What's new

Virtual vs Reality - The Potentials of Sampling

Which of the two versions A or B is a fully virtual orchestration?


  • Total voters
    23

jamess199

New Member
Hey guys.

Studying at university, I am currently undergoing my major project. As you can tell by the title, the research I am carrying out is analysing the realism of virtual instruments for composers and if a fully 'virtual' workflow could compare to traditional methods of recording an orchestra.

It is a tricky topic, and full of opinions and preferences but hopefully my findings will be positive! Any discussion on this topic will be massively appreciated as it will help towards researching my project.

To start the topic, I have re-scored two pieces of music, fully virtually. Excerpt 1 is a minimal piece of film music. I have included both the originally recorded piece and my re-scored version.

Excerpt 2 is the same concept, but instead a different piece of classical music. The aim is to identify from A and B which is virtual and reality. Any comments or discussion is really appreciated, regarding determining the two excerpts.

Here is the folder with the appropriate samples - https://www.dropbox.com/sh/r6d9cue692vumef/AAAOp4a6vKBAOtmWh8vMX1pma?dl=0
 
Hi jamess199,
welcome to this forum and have a good time here.
The topic samples vs. real has been very hot here for obvious reasons and we have had several A/B tests in the recent years. It seemed to have cooled down a bit lately since many practicioners have come to the conclusion to combine the best of both worlds since live recordings are not astronomically expensive any more.

Regarding your test here are some comments: You did quite well in mocking up the music! Of course the outcome of such tests is strongly dependent on how much effort is invested in the test material, so if you look at this in a scientific sense there you have a strong variable.

That being said I hear quite an amount of distortion in all versions of the music. Also I believe that you have altered the gain of the original recordings, and this is a non-neglectable factor.

Then: While the verdict seems to be clear with the first example so far which is puzzling since it is not much more than a drone plus a harp it is a tie with the Barber. I believe the reason in this particular case is that the live version is played with a very mellow tone and sort of expressionless/introverted which makes much sense in conjunction with the whole piece, but some hearers obviously take that as the sampled version in contrast to the brighter samples. I bet it would be different if we had a longer version with ups and downs of the music.

The last comment is: How many here would be able to know which is sampled just by looking at the waveform? Now that would be an interesting test. :)

Anyways, much success for your thesis!

realvsvirtual_waveforms.jpg
 
Last edited:
Hello James,

Your topic of Virtual Vs Reality has taken a real interest in me, and I appreciate the amount of time and effort you have clearly gone into to create realistic and excellent sounding virtual recreations.

As far as my views, in both cases I have gone with excerpt A as being virtual, though found it to be quite a difficult decision, especially in the case of excerpt 2.

With excerpt 1, my decision was swayed towards A simply because of the tone of the recording. I find that live composed recording has a certain tone that is created, maybe due to the emotion, care and attention that musicians and composers associate themselves with when playing the music (not to say you haven't put in an equal amount of care and attention). I also was swayed due to the background noise featured in B.

Excerpt 2 was a much more difficult task, and took several listens before I came up with a decision (though am still pondering on the idea to if I am correct). I came to the decision of A in the end due to one significant factor, the sound of the basses at around 19 seconds. I felt the tone of them stood out uncharacteristically more than in A, and feel that your decision would be more towards blending the basses in more, that trying to produce an uncharacteristic tone. This also added in with the extra noise featured in B swayed me towards picking A.

Very well done on the project though and the test, and thank you for sharing it around, very good quality work and I hope your Major Project goes well.

All the best
 
Very nice mockups!

Excerpt 1 was A. The English horn sounded more realistic and you could hear the emotion in the strings.

Excerpt 2 was B. This one was close, but the 1st violins gave it away with the odd vibrato and note transitions around 0:13-0:17.

If you look at the waveform, the real ones have much more little swells and decreases as a real orchestra naturally puts them in. Also, samples are usually normalized a least a bit which helps keep the waveform still.
 
Really appreciate all the comments and feedback, thank you guys!


What I've been exposed to learning through this research is the amount of variables. It really is dependent on the job in hand and no ultimate answer.

What about the future of sampling? Is it feasible to believe we could compose fully 'in the box', whist maintaining emotion, dynamics, style etc...? It just seems the detail sample libraries, notably orchestral, are recently providing us are ever increasing in quality and versatility?!

Or maybe the virtual world has its footprint in the industry, and will simply stay put?
 
Hey James,

As someone who hasn't really studied in-depth music composition, scoring and engineering, I found this test quite challenging yet at the same time very interesting.

I believe that A is the virtual composition in excerpt 1, because I felt that the sound was more pure and organic to my ears with B.
I think that B is the virtual composition in except 2, as I felt A was a more richer and fuller sounding piece.

Very impressed with your technical ability, you have real talent and very good luck with your major project!
 
Very nice mockups!

Excerpt 1 was A. The English horn sounded more realistic and you could hear the emotion in the strings.

Excerpt 2 was B. This one was close, but the 1st violins gave it away with the odd vibrato and note transitions around 0:13-0:17.

If you look at the waveform, the real ones have much more little swells and decreases as a real orchestra naturally puts them in. Also, samples are usually normalized a least a bit which helps keep the waveform still.

Based on same reasons as Lawson, I think 1A and 2B are virtual. It was an interesting test. Very good mockups!!
 
I voted and then thought to myself: "Wait, I'm the ONLY ONE who thought 1B was the real one?!?!" And then I realized that the voting was for the opposite - oops!!! I fixed it, though.

I think the mezza di voce gives the Cor Anglais away in the first example - the real player nails exactly where the swell, the turn of the phrase, and the note timing goes, while the sample library sounds like a real player who isn't really paying attention to what's going on around him!

I actually thought the Barber was more obvious - the sound of those strings in 2B is really strident, with some accordion-y vibrato, and that way that fake strings often feel like the emphasis on a moving line is on the note transition, instead of giving gravity and depth to the body of the note. If I were conducting 2B, I'd be giving dirty looks to the first desks to tone down the clarity, and pushing the back desks - which gives a rounder, weightier tone. As an example, the sensitive phrase from 0:18-0:20 is played sensitively and 'like a string group' in 2A, while 2B plows into that profound note at the end of the phrase like it's on fire.

Should we do the guessing game? That vibrato on 2B sticking out sounds like it might be VSL Solo Strings pasted on top, while the tone color sounds to me like Kirk Hunter Concert Strings.
 
After listening a second time with headphones, 2A is definitely real therefore 2B is not.

Reasoning:

At ~0:23 you can hear microphone noise during the pause. At ~0:30 you can hear a player breathe.


As a result I have changed my vote.
 
Very good mockups. The big giveaway for me was the natural artefacts in the live versions - Rumble, mic hiss, room tone, breath sounds etc.
 
Ahh..that brightness on the mockups. :)

I remeber when i started composing i heard a comparison of the same line played by the 1st violins VSL and the same 1st violins on the apassionata series, and i could only hear a duller sound on aspassionata which made me think i prefered the regular VSL sound.
 
Ahh..that brightness on the mockups. :)

I remeber when i started composing i heard a comparison of the same line played by the 1st violins VSL and the same 1st violins on the apassionata series, and i could only hear a duller sound on aspassionata which made me think i prefered the regular VSL sound.

haven't we all been there in the one or other form? :P I call it the brightness war
 
Here are 2 extracts from Pirates of the Caribbean: On Stranger Tides. I think that for people using samples it's not a difficult thing to say which one was made using samples, but I post this just to show how realistic samples can be nowadays. And by the way, one of the extract is the real cue from the movie, and the other is the demo made by Hans Zimmer himself.

https://www.dropbox.com/sh/6lg0izp8pdznzd1/AAB3szquy28YKdZthnHCvW6na?dl=0
 
Hey guys! Thank you everyone for the feedback, really interesting discussions happening!
It is infact 1B and 2A that are real live recordings.
Does anyone else have any A/B comparisons of virtual and real? I find analysing the two so interesting!
 
I have lots of A/B comparisions, unfortunately the most virtual versions are under NDA.
 
I thought that may be the case! Not to worry, is there anything available for educational purposes, under no copyright laws etc?
 
Does anyone else have any A/B comparisons of virtual and real? I find analysing the two so interesting!
I have a mockup of the opening cue from Chicken Run that I made last year for practice.


And here's the original. -
 
Top Bottom