RX950 Classic AD/DA Converter

[Marco Raaphorst]

amazing how when the peak stays the same, the RMS goes up. so the sound perception will be louder because of more complexity to signal without compression.

it is still amazes me that saturation adds more interesting tones to a signal.

What control on the RX950 does this? I want to check it out myself!!!

IMO saturation is the "salt" of the music world. A little bit will enhance almost any flavor, but too much and it just tastes "salty"…

saturation via input level. Adding odd harmonica in a nice way and adding to rms level.

[Marco Raaphorst]

amazing how when the peak stays the same, the RMS goes up. so the sound perception will be louder because of more complexity to signal without compression.

it is still amazes me that saturation adds more interesting tones to a signal.

What control on the RX950 does this? I want to check it out myself!!!

IMO saturation is the "salt" of the music world. A little bit will enhance almost any flavor, but too much and it just tastes "salty"…

low input level is interesting as well for that quantisation noise. it's a nice device which creates a sound I don't think I can mimic with Scream or any other VST plugin. it's low tech yet complex

[joeyluck]

What makes me curious is, since we can adjust bandwidth (which i assume is downsampling), what setting gives the sound of the Akai ? Is it the default setting ?

Listening to the demos, it seems to be doing something similar to when I send my sounds to my Korg ES-1 and then record straight back to the computer ! (except the volume going up) They become nastier, crunchy and have more presence. Will have to trial it when time allows.

The Akai could adjust the sample rate from 40kHz down to 7.5kHz, so I'm guessing that knob does the same - does it not read "Sample Rate" as you adjust it (shame if it doesn't, for those who are familiar with the original)?
So any setting is "the sound" of the original - just depends on which sound you're after!

The Audio Bandwidth, like the S950, adjusts from 3000 Hz to 19200 Hz

[inphonik]

Hi there,

Just thought I could shed some light on this topic:

Nowadays, we're kinda more used to the sample rate notion, but this should not be confused with audio bandwidth here. It's all about the famous Nyquist–Shannon sampling theorem. And as we're not in a perfect world, Akai chose to be even more careful: on the S950, the internal sample rate equals the audio bandwidth setting times 2.5. Do the math: 19200 x 2.5 = 48000

Cheers everyone, and thank you all for your kind words.

Best,

Mathieu

[nickb523]

Hi there,

Just thought I could shed some light on this topic:

Nowadays, we're kinda more used to the sample rate notion, but this should not be confused with audio bandwidth here. It's all about the famous Nyquist–Shannon sampling theorem. And as we're not in a perfect world, Akai chose to be even more careful: on the S950, the internal sample rate equals the audio bandwidth setting times 2.5. Do the math: 19200 x 2.5 = 48000

Cheers everyone, and thank you all for your kind words.

Best,

Mathieu

Super RE. Mais pouvez-vous corriger le réglage par défaut? - La valeur par défaut est trop forte. L'audio est très chaud. Je vous remercie.

I'm not a native French speaker.

The default setting is very, very loud. I'm surprised that it's only myself and Navi that mentioned it.

[selig]

Hi there,

Just thought I could shed some light on this topic:

Nowadays, we're kinda more used to the sample rate notion, but this should not be confused with audio bandwidth here. It's all about the famous Nyquist–Shannon sampling theorem. And as we're not in a perfect world, Akai chose to be even more careful: on the S950, the internal sample rate equals the audio bandwidth setting times 2.5. Do the math: 19200 x 2.5 = 48000

Cheers everyone, and thank you all for your kind words.

Best,

Mathieu

Sorry if I fed any confusion - as a past S900 owner I assumed the sample rates were the same.

About bandwidth, it’s tied to sample rate (all that was being said, right?). But the actual bandwidth depends on the filters used, not just the sample rate. For bandwidth to equal Nyquist, you would have to have perfect brick wall filters, which do not exist (certainly not in the analog domain). So it’s safe to say that audio bandwidth will be less than half the sample rate.

But as a recreation, I’d assume the bandwidth control would read the sample rate since that’s how the original worked, correct? At least that’s how it worked on my S900 - is the S950 different?


Sent from some crappy device using Tapatalk

[joeyluck]

Hi there,

Just thought I could shed some light on this topic:

Nowadays, we're kinda more used to the sample rate notion, but this should not be confused with audio bandwidth here. It's all about the famous Nyquist–Shannon sampling theorem. And as we're not in a perfect world, Akai chose to be even more careful: on the S950, the internal sample rate equals the audio bandwidth setting times 2.5. Do the math: 19200 x 2.5 = 48000

Cheers everyone, and thank you all for your kind words.

Best,

Mathieu

But as a recreation, I’d assume the bandwidth control would read the sample rate since that’s how the original worked, correct? At least that’s how it worked on my S900 - is the S950 different?


Sent from some crappy device using Tapatalk

Here are some videos showing Audio Bandwidth on each unit. Maybe you did the math in your head? Mandela Effect?

S950

S900

[inphonik]

Yes, it is safe to say that (as expressed with the aforementioned formula, or the other way around: bandwidth=samplerate*0.4 which is less than half as you said). Actually, the S950 use the audio bandwidth parameter to define the pre-filter cutoff frequency and the internal sample rate.

[inphonik]

As for the loud default setting: this is a very unfortunate issue that slipped in our code while we were dealing with an even more unfortunate acceptance issue from PH (let's call it even ). We fixed that issue and submitted version 1.0.3 for acceptance a few hours ago. We're sorry for the inconvenience. Will keep you posted.

Good night

Mathieu

[miscend]

Blamsoft make Resampler, which is a straight (but very tweakable) rate and bit crusher.
This seems to model all the audio effects of using an Akai S950 sampler. Input, AD/DA converters, bandwidth, output, filter.

So they're only loosely related products.

Resampler isn’t a straight bit crusher. It emulates different converters. When you switch between the different modern and classic modes you are switching between different converter types.

[miscend]

Is it as good advice the Blamsoft one?

What's the Blamsoft one?

Blamsoft Resampler.

[dioxide]

I'm looking forward to demoing this as a former S950 owner. I always wondered how much the S950 sound was to do with clipping, considering most people then didn't understand clipping like people do now (well me anyway). How much did this contribute to the sound and can it be recreated with this RE?

[EdGrip]

I was having a play with the trial last night - it's a lovely RE! It's simple and lightweight (1mb download!) and just samplers everything right up. The bandwidth control lets you fine-tune the low-pass sit-in-the-mix crunchy sampler quality, while there's saturation and dirt available from the input. The filter is also lovely!

The only thoughts I had for improvement were:

- it would be really cool to be able to use the filter without the rest of the processing. If a switch on the front would mess up the super-neat design, a pair of "filter input" jacks on the back would be just fine! I love filters.

- the filter knob doesn't seem very linear. It changes drastically from 12 o'clock to 2 o'clock and then doesn't change as much from there until the end of its range. It makes smooth filter sweeps a bit tricky. Maybe it was just me being rubbish with the mouse - I'll try it with automation.

I'm pretty sure I'll end up buying this even though I have TAL Sampler, just because it sounds great and has more tweakability than TAL's sampler models. And I can use it on any source!

Any more fun technical or development facts you can share with us, Mathieu?

[inphonik]

Yes, there will be more in one way or another

As for the filter knob, it's definitely non-linear as well, with respect to the original. Went through intense curve fitting for this

[mtbh]

The before/after soundexamples on the soundcloud are not normalized and jump in volume. It is not a very good way to show how it sounds, since louder will always be better .. I think many would appreciate a normalized preview. Just letting you know, mathieudemange. Looks interesting, though

[selig]

I'm looking forward to demoing this as a former S950 owner. I always wondered how much the S950 sound was to do with clipping, considering most people then didn't understand clipping like people do now (well me anyway). How much did this contribute to the sound and can it be recreated with this RE?

Excellent point - I even remember reading an article in Keyboard Mag (late 1980s era) about intentionally clipping the input on transient sounds to get more level. The advice was actually not that bad - they mentioned that clipping of less than a few ms was virtually undetectable to the human ear. I tried it on some percussion samples and was happy with the results (was previously was unable to get them as loud as the other samples in the perc kit).

I also used Alchemy (and later Sound Designer) to get some samples into the device (in addition to editing), bypassing the front end (A/D) entirely, which IIRC was a cleaner sound. This leads me to believe the A/D (where initial sample rate/bandwidth was determined) was a big part of the Akai “sound”.

Which reminds me: an “S900 mode” would be pretty sweet - always preferred the sound of the S900 over the ‘cleaner’ S950 and the super-clean S1000 series.


Sent from some crappy device using Tapatalk

[mreese80]

This is a beast. I have the vst and the RE Version of the RX950. If you used the Akai S950 back in the day you already know this is a must buy. The price is so good...so good. After messing with the filter for no more than 10 seconds, this should be an instant buy for all who are trialing the device. I'm in love with it.

[willbe]

The before/after soundexamples on the soundcloud are not normalized and jump in volume. It is not a very good way to show how it sounds, since louder will always be better .. I think many would appreciate a normalized preview. Just letting you know, mathieudemange. Looks interesting, though

Hi everyone!

I've made the example audio clips. I admit there's an intended difference in loudness between dry&wet, simply because that's what the RX950 provides in my opinion.

Full story : I've been working with Mathieu for a very long time on the design of the RX950 before I could get an actual version to test, in order to start working on the demosong in Reason. I've been blown away by the easiness of making each instrument louder and warmer in my mix, something I was struggling to obtain properly for years. This is something I did want to put into the audio clips.

[KirkMarkarian]

Here's a short track using the RX950 on the entire track:

I used:
Kong (all synthetic drums)
Parsec

The audio chain:
Distillery > RX950 > Kuassa

I like the warmth it imparts, as well as the ease of use. In this case, it's got the sound I want, and few, concise knobs. Love it!

[selig]

The before/after soundexamples on the soundcloud are not normalized and jump in volume. It is not a very good way to show how it sounds, since louder will always be better .. I think many would appreciate a normalized preview. Just letting you know, mathieudemange. Looks interesting, though

Hi everyone!

I've made the example audio clips. I admit there's an intended difference in loudness between dry&wet, simply because that's what the RX950 provides in my opinion.

Full story : I've been working with Mathieu for a very long time on the design of the RX950 before I could get an actual version to test, in order to start working on the demosong in Reason. I've been blown away by the easiness of making each instrument louder and warmer in my mix, something I was struggling to obtain properly for years. This is something I did want to put into the audio clips.

Quick sanity check…
The RX950 adds 9.25 dB of gain - are you suggesting the original S950 added that much gain as well? I'm pretty certain my S900 didn't add any gain by default. So yea, it's going to make your instruments sound louder for sure!

Input in Blue, output in Yellow:

Screen Shot 2018-03-22 at 10.59.37 AM.png (144.36 KiB) Viewed 3172 times

That's my only disappointment with this otherwise stellar device. Additionally, it would have been nice to see the original input meter to help with intentional clipping for effect.

And although I have tons of sample from my S900 that already have "the sound", I'm sure to find many uses for this device in my rack!

[inphonik]

Hi everyone,

Update version 1.0.3 is available and resolves the Output Level issue. Sorry again for the inconvenience. Time to sync your REs.

Cheers,

Mathieu

[mtbh]

Full story : I've been working with Mathieu for a very long time on the design of the RX950 before I could get an actual version to test, in order to start working on the demosong in Reason. I've been blown away by the easiness of making each instrument louder and warmer in my mix, something I was struggling to obtain properly for years. This is something I did want to put into the audio clips.

Ok, well it is not a dealbreaker the real way to hear it is of course to "try" it. I agree that it adds "warmth" or let's just say, awesomeness, to every thing I put in front of it. Nice RE. Job welldone

[Marco Raaphorst]

I bought it. It's a great "more RMS and more sweetness device". Adds pressure to sound, more RMS in a unique way. Nice saturation and low tech data.

[joeyluck]

Hi everyone,

Update version 1.0.3 is available and resolves the Output Level issue. Sorry again for the inconvenience. Time to sync your REs.

Cheers,

Mathieu

[joeyluck]

Hi everyone!

I've made the example audio clips. I admit there's an intended difference in loudness between dry&wet, simply because that's what the RX950 provides in my opinion.

Full story : I've been working with Mathieu for a very long time on the design of the RX950 before I could get an actual version to test, in order to start working on the demosong in Reason. I've been blown away by the easiness of making each instrument louder and warmer in my mix, something I was struggling to obtain properly for years. This is something I did want to put into the audio clips.

Quick sanity check…
The RX950 adds 9.25 dB of gain - are you suggesting the original S950 added that much gain as well? I'm pretty certain my S900 didn't add any gain by default. So yea, it's going to make your instruments sound louder for sure!

Input in Blue, output in Yellow:
Screen Shot 2018-03-22 at 10.59.37 AM.png


That's my only disappointment with this otherwise stellar device. Additionally, it would have been nice to see the original input meter to help with intentional clipping for effect.

And although I have tons of sample from my S900 that already have "the sound", I'm sure to find many uses for this device in my rack!

There was a fix that is now available. This behavior wasn't intended and wasn't present during beta testing, but as Mathieu explained, was a bug introduced sometime during the acceptance process. I think what willbe is describing, is the use of the input gain knob, and the deliberate, manual adding of loudness.