EPS 16+ 2KXL S3000 Freq curves

This is not scientific whatsoever.
I have not used any spectral analyzer hardware to do those test.
It has been done from a user perspective.
Signal path:
Studio One
chirp Sine wave 20hz to 20khz
Audient ID22 Alt Out (Mono Left)
Monster Studio Pro 2000
Sampler Left input
(EPS 16+ = 44.6khz 13 voices, S3000 = 20khz, 2000XL = 44.1khz)
Sampler left output
Audient ID22 Return ( Directly hitting the AD converter, bypassing the pre )
Voxengo SPAN to analyze.

Original signal:


ID22 loop back:


EPS 16+


S3000


2KXL


Take note that Saturation per frequency or THD havent been captured so this doesnt reflect the unit behavior but its actual frequency response and thats all there is to see here.
Also note that what you're seeing here is an extreme close up of 4db of dynamic range.

Now shove this down your throat and choke on it

Debate , whine, throw a tantrum in 1,2,3 GO !!!

The s3000 looks like a monster on the low end...

Cockdiesel wrote:The s3000 looks like a monster on the low end...

Well, it certainly feel heavy but not 5db heavier for sure.

Like I said in the post, theres at least 2 parameter that arent shown that would also help us understand those unit behavior.

Also, lets keep in mind that all of them have mostly +/- 1 or 2 dbfs difference from 200hz to 20khz.

A human perceived level is 3db for 1 unit of gain reduction ( see that as stage ).

So the difference are REAL but the frequency curve really caught me by surprise.

I mean, the one that make the most sense to me is really the EPS.

That drop from 2k to 8k is really what I'm hearing. Also that warmth in the 500hz tome is even more prominent. Might be the saturation at that frequency.

The 2K on the other hand, that smooth 2k to 19k LPF curve is really surprising. Again, there must be a SH!T load of saturation happening near 2K or higher. The agressive behavior of the 2k is undeniable.
Also if you look closely, you see a roll of from 30hz to 20hz. That also might give a hint to some architecture that might be made maybe with some 15ips R2R in mind.

And now the S3000....all around the place hahaha.
I mean, it make sense for real. The S3000 sound is REALLY special. That cut at 350 where some of the meat usually happen compensated with that huge 500hz alias is very very insteresting.
If you guys remember in most of the article where people were talking about dilla sound, they always said he wasnt scared to boost or cut frequency with very narrow Q.
Looking at this freq graph, it seems like a huge part was commin from the hardware itself.

Again, Very very interesting but there'snot much to understand here.
I would really need to capture the saturation accross the spectrum for us to understand the nature of each units. And sadly, that I dont have a single clue how to do that.

how long is your chirp? the longer the better.
id be really curious your process to do this, mostly because these readings look so low, why is everything so quiet?
honestly, the 3000 looks like an error, but i dont have one, and the signal appears so low, who knows.

you could do a similar test by sampling white noise and seeing what it looks like.

tapedeck wrote:how long is your chirp? the longer the better.
id be really curious your process to do this, mostly because these readings look so low, why is everything so quiet?
honestly, the 3000 looks like an error, but i dont have one, and the signal appears so low, who knows.

you could do a similar test by sampling white noise and seeing what it looks like.

The sweep was 10 seconds long.

The sweep gave me a more accurate reading on AVG hold.

The white noise I got was moving all accross randomly with a shelving at 45hz and lower.

I've tested the S3000 4 times because I had the same impression you have.
But the reading is correct and the setup was the same as others.
The audio I got also gave me the impression that the signal wasnt processed the same at all.

All received signal was calibrated at -18dbfs and the sent signal was calibrated at the optimal input stage of the samplers.

yea i think the chirp is a better choice, whitenoise is just another option.
good call on the length of the chirp too.

i guess i'm just understanding the db meter then? i would think they would be closer to the 0 mark than -36db down

tapedeck wrote:yea i think the chirp is a better choice, whitenoise is just another option.
good call on the length of the chirp too.

i guess i'm just understanding the db meter then? i would think they would be closer to the 0 mark than -36db down

To hit 0dbfs, I would have needed some kind of preamps to boost my signal out of the samplers without introducing too much noise from the sampler when his amp is kicking in.

I found that in both Akai units, its when I cross the 12 oclock line. Then I start to see the noise floor rising.

Well, I'm not sure how to calculate this but it was a mono signal sent back in a mono input ( hitting A/D converter directly , no preamps involved here ) and recorded in the daw on a mono channel.

-36dbfs would be half as much as -18dbfs in this stereo Voxengo SPAN

Or maybe I'm totally off hahaha.

I really dont know why SPAN read it at -36dbfs to be honest.

EDIT: looking back at those results, its quite panicking to see that I seem to have a ground problem in my ID22 and my 2kXL ( That 60hz peak tho.... )