"Consumer" versus "Pro" Sampling: an Alesis SR-16 "Case Study"

[Elteto]

For those of us who worry about having to spend big bucks on home studio kit, let me encourage you with this project I have just finished. The outcome is reminiscent of a comparison CD I found online that points out how surprisingly small, if any, differences are between "pro" studios and project studios in certain recording aspects.

Let us get the technical background out of the way. I used an actual Alesis SR-16 to satisfy the purists (as there are other options as well to acquire the SR-16 sample set, including buying it online). I still got it cheap on eBay, for $60, including shipping. This particular one was manufactured in 1996, and except for cosmetic signs of use, was in a fully working state. I connected the main 1/4" outputs with RadioShack 6' shielded audio cables to an E-MU 1212M audio interface, which I set to -10 on the inputs instead of +4 to get more gain on the original signal. I recorded preset kit 00 "Rock 1", triggering the sounds through a MIDI cable from an M-Audio Trigger Finger, set to full velocity. This was to prevent any possible circuit noises or velocity inconsistencies physical hits on the body of the SR-16 might cause. I used Audacity 1.3 Beta, set to 24-bits/96kHz. Even though the original samples were in 16-bit, 44.1kHz (I think), I hoped that the higher sampling rates would capture all nuisances of the machine's sound. Once finished capturing all 12 samples of the kit, I sliced them into individual .WAV files in Audacity. I built a Battery 3 kit from these samples, imitating the MIDI note and pad assignments of the SR-16. I sampled each hit in its original volume, pan, and assigned the open and closed hi hats to the same voice group to copy their original setup. Basically, the closed hat sample would cut off the open hat when triggered. Sampling the kit as a whole, maintaining its original settings, also helped preserve the dynamic differences between the individual samples. The hi-hat or the cabasa would sound much softer than the kick or the snare, as expected. I whipped up a minute-long drum sequence in Cubase, using Battery 3 as a VST plug-in. I exported the mixdown in 24/96. Then, maintaining the MIDI note and pad assignments, I switched my samples with the Killa Beat lineup I had purchased commercially, and exported the same sequence in the same resolution. To have more to compare these works to, I also used the sequence to trigger the SR-16 through MIDI and recorded the machine's output, again at 24/96.

I placed all three files in Audacity, making sure they are exactly lined up, and I normalized each track. For those not familiar with the process: all normalizing does is bring the loudest part of an audio file up to 0.0dB (or whatever the engineer designates), but the relative dynamics are maintained. In short: normalizing is NOT compression.

I reset the zoom settings in Audacity, so all three tracks were the exact same size on screen, and I took a screen shot.

Here are some interesting finds:

1. Killa Beat normalized individual samples, erasing drum kit dynamics.

2. Killa Beat's samples were 16/44.1, which still sounded crisp, but due to the normalization, a bit on the excess side.

3. Looking at the wave files, the Killa Beat track showed the reduced dynamics due to the normalization.

4. The crash and ride samples were audibly cut short in the Killa Beat set.

5. Panning exactly matched the machine in my sample kit.

So what does it all mean? For starters, one may crack OCD and red stapler anecdotes about my project, but at the same time, there is some encouraging results here for us "consumer"-grade musicians.

1. $15 electronics store cables are not that inferior to over-priced monsters, no pun intended.

2. $150 will get your analog electronic instruments recorded about as well as a multi-hundred-Dollar interface.

3. Having Pro Tools with fancy plug-ins does NOT mean you have to use it all, and there is no guarantee your work will sound any better just because of the software.

4. I know I used Cubase, but I would have had the exact same results with Reaper. A VST host is a VST host.

5. The samples would also have sounded the same in any other VST sampler, Battery just happened to be installed on this studio PC.

6. 16/44.1, 24/48, 24/192... There ARE some differences, but if your music makes no sense, it will do so at a higher sample rate, occupying a larger hard drive footprint.

Here is the screen shot, please click on it to see a larger version. The top two tracks are my sample kit, the second two are the direct output of the SR-16, and the last two are the Killa Beat version. Notice that my version even though a bit thicker, is nearly identical to the machine output, while the beefiness of the Killa Beat track shows the reduced dynamics.

Here are the files in .MP3 format, 320kbps/48kHz.

Alesis_SR-16_Machine_Direct.mp3

Alesis_SR-16_Killa_Beat.mp3

Alesis_SR-16_Elteto.mp3

And if you read all the way to the end, here is the entire Battery 3 Rock 1 kit as a reward.

Alesis_SR-16_Rock_1_by_Elteto.kt3

[Dave B]

Eh?

So ... the practical upshot of all this is that you can sample a drum machine and it sounds like a drum machine, but if you buy drum samples online, they still sound like drum samples ... yes?



Are the KB samples (never come across these .. sorry) advertised as being 'natural' sounding? Or are they designed to give a certain sound?

I don't understand the point that you are 'proving' here. Typically, such experiments have an introduction to explain what they are testing. Could we have one of those please...

[coool]

!!!!!!!!!!!!!!!!!!

[sharpeye]

i dont know if im missing something but where is the comparison between pro and consumer?

what high end gear did you compare these tests with?

[Tomás Mulcahy]

Agreed, there's no pro v consumer here. You've sampled a set of existing samples. You haven't made your own samples of a real drum kit.

I agree with your conclusions, except that they are not conclusions because they have nothing at all to do with your test, so they are just opinions!

You're also a little confused about dynamics. If you trigger the SR-16 with full velocity, then you've eliminated the multiple samples the unit uses for lower velocities. The SR-16 actually cross switches between samples to imitate the change of timbre that happens with a real instrument when you hit it at different "velocities". All you've done is sampled the cabasa at a lower level relative to the other instruments in the kit. That's a fixed volume level, not a dynamic (i.e. changing) volume level.

To sample an SR-16 properly, you need to send notes at varying velocities, and listen to each instrument to hear which ones use velocity cross switching. The cabasa has at least 2 samples, and the hi hats have 3 samples, if memory serves.

You're also a bit confused about what -10dBV and +4dBu is for. This has been explained many times on the forum, do a search.

You should also look into gain structure for a 24 bit system v a 16 bit system, and AES-18.

Finally, there is nothing to be gained by recording the output of an early nineties 16 bit DA at a higher resolution than 16 bit 44.1kHz, given that the Emu has vastly superior converters.

[Aural Reject]

I think the implication is that he made his own samples and compared them with a sample CD that he got from somewhere....although without rereading the whole thing I don't see the chain that the 'professional sample CD' was made with.

Or something.

[Doublehelix]

Finally, there is nothing to be gained by recording the output of an early nineties 16 bit DA at a higher resolution than 16 bit 44.1kHz, given that the Emu has vastly superior converters.

Ummm... why not? Maybe I am missing something here, but he has an analog signal, it matters not what source resolution it came from. If recording an analog signal, why wouldn't I would use the best resolution I could? Now if he was transferring a digital signal directly, then yes, I could see your point.

When I first read the original post, I wondered why he even brought that topic up in the first place.

Maybe you could clarify your answer for me a bit Tomas. Thanks!

[Tomás Mulcahy]

Ummm... why not?

IMO, the SR16 uses budget 16 bit 44.1kHz converters from c. 1990. The Emu has excellent modern converters, which are far closer to the theoretical ideal resolution that can be attained with that sampling rate and bit depth. So they can capture more detail than the SR16 can provide, especially if you consider the fact that all level changes on the SR16 are done in the digital domain, with a reduced DSP power compared to what we are used to today.

IMO converter technology has moved on drastically since the SR16 was made, especially the reconstruction filter side of things, and to a lesser extent the A to D side. I doubt that the SR16 is reconstructing the lower bits with full precision, wheareas the Emu can certainly achieve greater precision capturing at 16 bit resolution- so even though all that's being captured is quantizing error, the Emu is capturing that accurately

Anyway, the sampling rate and bit depth are the very least of the OP's problems

[stinkfinger]

well there's 60 seconds i'll never get back, thanks.



edit: and take the link down to the battery 3 rock kit before i report you to FACT.

[hollowsun]

all that's being captured is quantizing error, the Emu is capturing that accurately

Perhaps those are the "nuisances" the OP was keen on capturing

[Tomás Mulcahy]

Ha! I missed that. Very funny!

[Elteto]

Here is the link to the commercial sample pack I had bought from Killa Beat Productions.

http://www.wildstyle24.com/index.html?target=p_13.html&lang=en-us

This is their claim:
"THESE SOUNDS WERE RECORDED AND INDIVIDUALLY PROCESSED BY PROFESSIONAL ENGINEERS USING PRO TOOLS FOR THE BEST SAMPLE QUALITY AVAILABLE."

So the point is, Pro Tools and professional engineers do not always mean that the rest of us "mortal" music enthusiasts cannot produce quality sounds.

Correct, I sampled a sampled instrument. There is nothing unusual about it; there are all sorts of sample packages out there with various drum machine sounds, such as those of 808 or 909.

As far as the velocity layers go, I understand the concept. However, if sampled at a consistent 127 velocity, the hat and cabasa samples will STILL be softer, as each individual sample in the SR-16 has its own assigned sample volume in the kit, which is NOT the same as the MIDI velocity. You can get to those settings by getting in the drum kit menu, and flipping through the pages. If you look at the waveforms, you will see that the dynamic difference between cabasa and kick is still there, even though each note in the sequence was at a standard 127 MIDI velocity.

No confusion about the -10/+4dBu, I simply explained that I switched the E-MU over, as the input signal level was too low.

Regarding the point of sampling an SR-16 at a higher rate: just check out http://www.goldbaby.co.nz/. They record drum machines to tape, then re-sample the playback to digital. The process captures a certain tone that many musicians seem to enjoy, as these "tape packs" are quite popular. Getting the SR-16 digital samples recorded after the D/A conversion adds some tonal coloration.

In conclusion, again, the point was that a "consumer", such as myself, can still achieve good enough results using budget kit. There are no ulterior motives here, I do not sell anything, and unlike many MySpace artists, my site does not even have advertisements on it.

[Elteto]

Finally, there is nothing to be gained by recording the output of an early nineties 16 bit DA at a higher resolution than 16 bit 44.1kHz, given that the Emu has vastly superior converters.

Matter of opinion. These guys (and their happy customers) certainly seem to think otherwise:
http://www.goldbaby.co.nz/

[Bertyjnr2]

Here is the link to the commercial sample pack I had bought from Killa Beat Productions.

http://www.wildstyle24.com/index.html?target=p_13.html&lang=en-us

This is their claim:
"THESE SOUNDS WERE RECORDED AND INDIVIDUALLY PROCESSED BY PROFESSIONAL ENGINEERS USING PRO TOOLS FOR THE BEST SAMPLE QUALITY AVAILABLE."

So the point is, Pro Tools and professional engineers do not always mean that the rest of us "mortal" music enthusiasts cannot produce quality sounds.

The only problem is, it's a fairly crap amateur product. Professional engineers?! Well I never! That can mean anything. What reputation does this company have? Apart from the spelling mistakes on their website.

They don't look very professional at all, and are selling products claiming to be the sounds of famous musicians without the musicians' permission.

Also, they tell you nothing of the sampling process for their other products, e.g. signal path, mics. For how it could be done, compare it to this:
http://www.propellerheads.se/products/refills/rdk/index.cfm?fuseaction=get_article&article=what_is_it

Edit: stupid forum keeps ruining the link, can a mod fix this?

[Elteto]

edit: and take the link down to the battery 3 rock kit before i report you to FACT.

I fully appreciate your concern, but I respectfully submit for your consideration that the Battery kit is posted as fair use.

1. It is only a fractional sample of the whole sound lineup of the SR-16.
2. It is not an exact or a digital copy.
3. It contains sounds created by an instrument I legally purchased for the purpose of music production. There are plenty of electronic as well as analog instrument samples out there.
4. The sounds have been altered from their original form.
5. It is used as a part of a review for educational, non-profit use only.

[Elteto]

Also, they tell you nothing of the sampling process for their other products, e.g. signal path, mics.

Well, one thing you can deduct is that if they use Pro Tools, probably they have some sort of Pro Tools compatible hardware, which could be anything from HD, LE, or MP.

Their sample packages have been selling at a few different web stores for months now, and so far they do not seem to have had any legal action done against them or have not been reported to any of the credit card processing agencies or to PayPal.

[Elteto]

You're also a little confused about dynamics. If you trigger the SR-16 with full velocity, then you've eliminated the multiple samples the unit uses for lower velocities. The SR-16 actually cross switches between samples to imitate the change of timbre that happens with a real instrument when you hit it at different "velocities". All you've done is sampled the cabasa at a lower level relative to the other instruments in the kit. That's a fixed volume level, not a dynamic (i.e. changing) volume level.

My apologies, I should probably clarify my point: uniform velocity triggering for all sounds in a kit will maintain volume differences between the samples, while individual normalizing will erase this. This drum machine has "8 loudness levels" (on the pads, as claimed by Alesis, whether it means just velocity zones or multi-layer samples), but you can adjust sample loudness INDEPENDENTLY from MIDI velocity. When all pads are triggered at the same velocity, their individual volume settings still cause them to have volume differences between the samples.

Press the "Drum Set" button, then press the pad for the sound you want to see/adjust settings for, then flip through the menu pages using the "Tempo/Page" up or down buttons. Notice that in the waveforms, the cabasas are softer compared to the kicks, despite the fact that all three sample sequences were triggered at uniform max velocity to ensure comparability.

To make a fair comparison though, I could only use one sample per sound, as the Killa Beat product only had the same.

I could certainly divide each sound into eight velocity zones, and record eight samples for every pad, recreating the layers in Battery. (Technically, I could do 128 sample layers if I wanted to) That would just further prove my point that "consumers" can achieve "pro" results.

[Elteto]

I don't understand the point that you are 'proving' here. Typically, such experiments have an introduction to explain what they are testing. Could we have one of those please...

Short summary:
Original question: Could I produce a better sample set than what I paid a "professional" publisher for? (Then my tech geek side took over and went into detail both in the experiment and the explanation of it, so I offer my apologies.)

Outcome: Yes.

Any claims of professional scientist credentials or this being a professional scientific experiment in the first place using the scientific method, or the results being scientific findings instead of perceptions and opinions: None.

[hollowsun]

Ok Elteto, I think I can see where you're coming from and you have a point ... up to a point.

The fact of the matter (as a professional samplist and sound designer of some experience and reputation) is that it's not difficult to sample a 90s drum machine and there is a lot of bollox and snake oil and Emperor's New Clothes out there regarding certain 'professional' sound libraries and the hype used to describe them.

Sampling a 90s 44/16 drum machine at anything other than 44/16 is a compete and utter waste of time, memory and disk space. Fact.

And in fact, you could probably sample a lot of digital beatboxes at 22/16 and not notice a difference. Fact!

There can only be some added value if those samples went via some 'boutique' outboard processing (or, indeed, some prestige plug-ins) to create some magical sound that transformed the originals in some way. Frankly (and with no disrespect), any arse can sample a raw SR16 and get good results - it really isn't rocket science.

I would venture to suggest that you have not highlighted the difference between 'pro' and 'consumer' sampling, just that you have maybe exposed a particular commercial library for not living up to its hype. Which is fair enough in its own right.

However, this does also bring up the thorny subject of copyright and IP and so on. Unless otherwise allowed and licensed from Alesis with whatever arrangement, the commercial library you are comparing with is technically illegal. I doubt very much they have that license (though I could be wrong). And technically speaking, that 'taster' of the SR16 samples you made available for Battery in your exercise is also technically illegal. You can cite 'fair use' by all means but Alesis can also cite 'lost sales' (SR16 is still a current product in Alesis' catalogue as I understand it) in a court of law where their IP is protected under international law.

Not that anything would come of this (I doubt Alesis would even spot this) but I suggest you withdraw the link to the Battery download to be on the safe side.

And you can argue all you like about that... I happen to know about these things on several levels.

I think you have presented an interesting case - and maybe even proved - that some hyped up commercial sound libraries are not all they seem. And that's not a bad thing - there are charlatans and cowboys out there selling crap. But I don't think you have presented a convincing argument for "pro" vs "consumer".

Sorry!

[Elteto]

Hollowsun,

Thanks for the time end effort you put into writing a professional response.

You actually got the point: an SR-16 "fanboy" tinkering around on a quest to get a better sample set than the one he paid for and is not satisfied with, proving that what he paid for was not as good as it could have been. That really is all there is to it. The SR-16 is my musical "red stapler".

Inspired by the GoldBaby tape samples, I wanted to capture everything that was coming out of the analog output of the SR-16. However, I agree with you on the bit-depth/resolution issue, and I even mentioned it in my findings that they did NOT make too dramatic a difference.

I would venture to suggest that you have not highlighted the difference between 'pro' and 'consumer' sampling, just that you have maybe exposed a particular commercial library for not living up to its hype.

As far as the whole market is concerned, absolutely not. I guess I considered a company using Pro Tools and publicly selling sample packs as "pro". I certainly considered myself a consumer compared to their scale of operations. Which leads me to the next point:

any arse can sample a raw SR16 and get good results - it really isn't rocket science

Well, apparently not everyone, as the Killa Beat example shows (even if they DO use Pro Tools). I feel triumphant in that aspect that the "little consumer guy with Audacity" could get it done decently.

I appreciate the input about IP, and to elaborate on that, that is something I specifically contacted Alesis customer service and press relations about (how is that for being a thorough "fanboy"?). I am not in the habit of upsetting people (especially SOS forum readers), so I will disable the link for now to prevent getting assaulted with multi-smiley peppered warning notices from fellow members.

But I don't think you have presented a convincing argument for "pro" vs "consumer".

With technology and distribution channels becoming easily accessible, and after the above discussions, it seems there are no clear definitions of "pro" and "consumer" (I think I will start another thread on that subject), but fair enough. I guess my next geeky project will be to take on a "pro" (and hopefully licensed) loop title, as I am sure you agree, the range of quality of those products can be rather wide.

[Dave B]

Thanks for the clarification. I now understand that your initial premise was flawed.

An arguably more interesting example of 'pro' vs 'hobbyist' (I would hesitate to use the 'consumer' label) would be the case study of the Natural Studio drum samples. These started out as a lone engineer in a studio deciding to sample drums (and some other instruments - I loved his upright piano samples) initially for his own use, and then later make them available on the web. At the time, multi-sample kits were fairly expensive and fairly limited - he was _way_ ahead of the curve. By the time his mk6 set of samples came out, it was a pretty hot set of samples.

However all was not well. First there were costs associated with this kind of online distribution and the download sizes were immense. So for v7 it became logical for him to switch to a paid product model - dirt cheap at 50quid for half a dozen DVDs - and become a 'pro' supplier.

So you are right : anyone can do this, but to do it well takes skill, time and patience. And that deserves payment. Maybe some wierdos do silly stuff like existing simplistic drum machines (although someone may still want those samples in a modern format), but to go beyond that is more complex.

I look forward to hearing your multi-sampled kits!

[forumuser732935]

*waves "hello" to hollowsun*

And technically speaking, that 'taster' of the SR16 samples you made available for Battery in your exercise is also technically illegal. You can cite 'fair use' by all means

Except he can't in Britain (well, England and Wales), because under English law there's no such thing. The law does allow for something it terms "fair dealing", which is basically only either (a) excerpting for academic / non-commercial research purposes when accompanied by sufficient acknowledgement, or (b) quoting for review or criticism, so long as accompanied by sufficient acknowledgement and the work is already public; but there's no generalised right of fair use similar to the one in America (in particular, one may note the lack of protection for parody). (b) doesn't apply here, but on (a) I'd guess that although the SR-16's continued availability was a complicating factor a case could probably be made that the entire project was only undertaken for non-commercial research purposes, and the offering of the Battery 3 kit is consistent with the academic principle of full disclosure.

...That's if I were a lawyer, and were asked to come up with a reply to a threatening letter from Alesis. But I'm not; I only did a year and a bit of law school. (Let's just say I pulled out when I realised that there was more to becoming a lawyer than the ability to come up with cleverly rhetoricised convincing arguments...)

Mind, this is all beside the point. I'd have to agree with you on the broader point, that the author of the original post hasn't really demonstrated what he thought he had.

[Elteto]

I'd have to agree with you on the broader point, that the author of the original post hasn't really demonstrated what he thought he had.

To clarify, the only thing I set out to demonstrate was that I could make a more precisely sampled SR-16 drum kit than the one I paid guys who make money off it could. That was easily proven simply by the over-normailzed KB waveform and by actually hearing their sample tails cut off, but probably due to the length of my original post, people read more into it than was really there. Upon review, I should have titled my post "SOS Forum Member vs. Killa Beat".