MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

The void of 24bit recording environments
MUFF WIGGLER Forum Index -> Production Techniques Goto page 1, 2  Next [all]
Author The void of 24bit recording environments
boxxgrooved
I know this is a subjective point, but I can't get over how much better my 1999 Zoom Sampletrack sounds with its 32khz/18bit sampling engine compared to my newly aquired Elektron Digitakt running at 48khz/24 bit.

I have developed a fascination for older equipment with lower sampling rates and bit depths. When I listen to modern music and nearly all new digital synths and drum machines with higher sample rate engines and bit depths (not analogue) everything sounds too bright and plastic (Novation), and there is this underlying feeling of massive space in the sound which needs filling out.

The Digitakt is nearly 20 years ahead of the Sampletrack yet it sounds weak and thin in comparison. This is not an advancement in sound to my ears, despite the better specs. I have noticed this with other boxes too. MPC3000 which I had was especially warm (16bit) and the Korg Electribe Sampler running at 33khz/16Bit sounded nice too.

I do not know much about AD/DA converters but I know enough to hear that some of the cheap Casio keyboards from the 80's with their low bit depths and sampling rates will resonate my base chakra more than a Novation Mininova or Korg Microkorg of modern ilk.

What is going on here? Is it the quality of the converters used or maybe that higher sampling rates and bit depths is not everything we expect it to be? I would imagine the quality of the converters used in an old Casio keyboard are inferior to what is used nowadays in the Novation Mininova for example, yet despite that the old keyboard usually always sounds fuller and fatter.

I remember once using an old Minidisc multitrack recorder and the songs I made on that have more coherence and sound warmer than most of the songs I make in a 24bit environment today. I can't remember what format Minidisc was but there was some audio compression going and possibly a low sample rate.

I am no audio engineer by any means but I have developed a good sense of mixing instruments over the years and letting everything sit properly in the mix, having said that the massive headroom and space available in 24bit environments seems difficult to tame. I will admit that I do not enjoy mixing in modern 24bit systems. Its almost like there is too much space to fill. I don't know if I am explaining this well but you only have to listen to some kids doing remixes on youtube to hear how weak everything sounds in the mix. Going back 20 years that same kid remixing a dance track through an analogue mixer and recording it to 16 bit DAT would probably do better off.

I have drifted off point slightly but I would still like to know why some of the old school 12bit synths and samplers sound warmer and punchier than most newer stuff today
Oldstench
It's the user, not the specs.
dubonaire
Oldstench wrote:
It's the user, not the specs.


Not necessarily, I've found that Elektron gear has poor converters. No idea why.
subbasshead
Is what you describe not predominantly a matter of taste? eg to take it to a further extreme some people love 8 bit sounds and yet I mostly cant stand to listen to them (or to bit crushers/aliasing or to warbly tape) - so why is that true for me and not them? Its simply taste isnt it? And there is no accounting for taste.....
Oldstench
dubonaire wrote:
Oldstench wrote:
It's the user, not the specs.


Not necessarily, I've found that Elektron gear has poor converters. No idea why.


And I'm sure both of us have heard great music made on Elektron gear and didn't know it was Elektron gear and who cares because it's the user, not the specs.
hsosdrum
dubonaire wrote:
Oldstench wrote:
It's the user, not the specs.


Not necessarily, I've found that Elektron gear has poor converters. No idea why.


Exactly. When I read the OP's first sentence I said to myself "It's the difference in ADCs." When comparing two completely different digital recording systems that are running on completely different pieces of hardware, you cannot automatically attribute a difference in sound to being caused by the difference in sampling frequency and bit depth. The FAR more likely culprit is the analog-to-digital converter stage — there's just way more different ways to handle the ADC process, and since it still deals with an AC signal that represents an original sound wave, there's lots of ways that signal can get altered that will affect the sound coming out at the end of the process.

Once you've converted that AC signal into digital code, you have to make enormous changes in it before you audibly affect the sound coming out at the end of the process. The sound quality differences between 44.1kJHz/16-bit and 192kHz/24-bit are extremely difficult to reliably and repeatedly hear, even through high-resolution playback equipment.

My DAW can record at both 192/24 and 44.1/16, and as an experiment I recorded live acoustic music (drums, so it was plenty dynamic) in my studio at 192/24, converted it to 44.1/16 and compared the recordings. The sonic difference was very small, but audible after repeated listening. And this was using identical hardware for all stages of the test: Same recording, same ADC & DAC (PreSonus 1818VSL), same DAW, same playback room, same monitors, same playback volume. The difference was there but it was very small and required concentration to be able to reliably hear.

If you're clearly hearing large differences between recording system A and recording system B, those differences are almost certainly being caused by something other than a difference in sampling rate and bit depth.

P.S. The final result of my experiment was that I record all live tracks at 192/24 (since hard-drive space is so cheap these days) and do all my mastering at 44.1/16.
dubonaire
Oldstench wrote:
dubonaire wrote:
Oldstench wrote:
It's the user, not the specs.


Not necessarily, I've found that Elektron gear has poor converters. No idea why.


And I'm sure both of us have heard great music made on Elektron gear and didn't know it was Elektron gear and who cares because it's the user, not the specs.


I'm sure I have heard great music. That doesn't change my direct observations of those machines. I'm not saying it's the specs because the specs don't tell the whole story. The specs just tell the story of the converter chip. I'm also not one of those people who thinks I need to buy $4,000 converters. I do know that audio I recorded into the Octatrack didn't sound great and it wasn't operator error and I'm not the first person to notice this.

Also I'm not impressed with the sound coming out of my digitone, and I'm also not the only person to notice this.

Anyway, I'm not going to round in circles arguing over your simplistic comment.
boxxgrooved
hsosdrum wrote:

If you're clearly hearing large differences between recording system A and recording system B, those differences are almost certainly being caused by something other than a difference in sampling rate and bit depth.


This is what I suspect. So what is this 'other' principle that makes a 1999 Zoom Sampletrack sound better than a 2017 Digitakt for example?

I also agree with the other comments above that the Octatrack sounds weak and some of their other gear like the Monomachine. The Digitone sounds thin compared to a cheap Volca FM and don't even get me started when comparing it to old Yamaha DX line, the DX sounds way better than the Digitone.

Does this conclude that Zoom used superior converters in a budget 8 voice sampler in 1999 compared to Elektron with a vastly more expensive boxes in 2017? Why would Elektron go to all that length to create brilliant devices and then put crap converters inside them? Sound is more important than ANYTHING.

I think the comment about the quality of the converters has more to do with it than 24bit vs 12/16bit etc.
Oldstench
dubonaire wrote:

Anyway, I'm not going to round in circles arguing over your simplistic comment.


[edited to remove snark]

You're correct. My comment was "simplistic" as well as pointless. I apologize for adding white noise to the thread.
matthewjuran
I’ve wondered if the mixing software algorithm has distortion tradeoffs.

What speakers are you using? I’ve also wondered if the speaker material and design has an effect on response time and accuracy that would affect how the DAC signal is distorted. Maybe your DAC and speaker combination do sound better with the reduced quality digital format?
xenosapien
most of these ADC-centric debates are full of half-truths and snake oil.

that being said, the octatrack does have a gain staging/clipping issue that is nearly undocumented, there´s a recent and ongoing discussion about that over at Elektronauts:
there seems to be a -12dB mismatch of what is indicated as being "clipping" and when it actually hard-clips the waveforms.

i.e., if you manage to (accidentally?) get the gain staging right when sampling directly into the OT, you will have ZERO problems and it WILL sound "normal" = good.

I never use the Octatrack for actual sampling, I do that via my DAW and prepare/cut/effect my material there and then just dump into the CF card.
this way, I have never had any problems of getting stuff to sound great - and if I wanted it super distorted, I would do that via the Octa´s bitcrusher FX or before that in my DAW/through outboard gear.

/end of OT-offtopic-rant ^^
----------


@topic:

I´d rather have the OPTION of "clean", high bit-depth /sample-rate and dirty up the material on purpose, than being stuck with 1990s technologie that gives me "that" sound - but I couldn´t sample a violin through that without it sounding like, well... a 90s violin sample.

wink

I did love my Korg Electribe ES-1and its 12bit (or was it 13bit/14bit? I vaguely remember it being something "uneven") sound though, first piece of gear I ever regretted selling.
dubonaire
boxxgrooved wrote:
hsosdrum wrote:

If you're clearly hearing large differences between recording system A and recording system B, those differences are almost certainly being caused by something other than a difference in sampling rate and bit depth.


This is what I suspect. So what is this 'other' principle that makes a 1999 Zoom Sampletrack sound better than a 2017 Digitakt for example?

I also agree with the other comments above that the Octatrack sounds weak and some of their other gear like the Monomachine. The Digitone sounds thin compared to a cheap Volca FM and don't even get me started when comparing it to old Yamaha DX line, the DX sounds way better than the Digitone.

Does this conclude that Zoom used superior converters in a budget 8 voice sampler in 1999 compared to Elektron with a vastly more expensive boxes in 2017? Why would Elektron go to all that length to create brilliant devices and then put crap converters inside them? Sound is more important than ANYTHING.

I think the comment about the quality of the converters has more to do with it than 24bit vs 12/16bit etc.


There is a lot more to DAC design than just the chip or chips used, it's the whole design of the circuitry. I know this, but I am not a DAC genius, so I don't know what that entails.

I mentioned it elsewhere but I was quite shocked when I put the Digitone through my good sound system.

Here is something interesting, this is the inside of a hifi DAC:



And here is the inside of a Digitone (the two big chips are the processors that handle the software synthesis):

boxxgrooved
Interesting pics above. On the Hifi board I imagine the signal passing through all those discrete components warms up considerably before it reaches the output jacks.

On the Digitone I bet the signal path is short from those chips to the outputs.

Maybe the old process of running a signal through many different components before it reaches the outputs helps to warm things up? I'm no electronic engineer by any means but I would be interested to know why some modern units sound thin given their better specs.
dubonaire
boxxgrooved wrote:
Interesting pics above. On the Hifi board I imagine the signal passing through all those discrete components warms up considerably before it reaches the output jacks.

On the Digitone I bet the signal path is short from those chips to the outputs.

Maybe the old process of running a signal through many different components before it reaches the outputs helps to warm things up? I'm no electronic engineer by any means but I would be interested to know why some modern units sound thin given their better specs.


Power conditioning and jitter suppression make up a lot of that DAC in the picture. When it comes to jitter there are two camps. There are the people who say jitter doesn't matter and there are those that say it does matter. I think it does.

The Digitone is a 2-in, 4-out ADC/DAC and microprocessor with midi and USB that costs less than some low-mid range audio interfaces, and less than most dedicated DACs. So either we are all being taken for a total ride by the DAC makers, or Elektron has cut corners to keep the price down.
Hainbach
I have not researched into the topic since my university days, but when I took digital signal processing classes the consensus was that the design of the analog filter stage before the conversion process plays a big role. If you have a lower sample rate the analog lowpass filter needs to be more agressive thus coloring the sound. My acoustic professor always brought up the early Lexicon reverbs and showed us the insides, stating that tuning the analog stages was important as the digital conversion. This goes both ways, AD/DA.

Thus something like the Sampletrack (lovely machine, also loves being recorded hot which adds pleasant drive) will have been tuned more to sound good with the limited specs of the ADC available at the time, while newer machines have ADC with more bandwith which makes them sound good in a "what comes in comes out sense" but lacks the oomph of "vintage" tuned analog input/output stage.
dubonaire
Hainbach wrote:
I have not researched into the topic since my university days, but when I took digital signal processing classes the consensus was that the design of the analog filter stage before the conversion process plays a big role. If you have a lower sample rate the analog lowpass filter needs to be more agressive thus coloring the sound. My acoustic professor always brought up the early Lexicon reverbs and showed us the insides, stating that tuning the analog stages was important as the digital conversion. This goes both ways, AD/DA.

Thus something like the Sampletrack (lovely machine, also loves being recorded hot which adds pleasant drive) will have been tuned more to sound good with the limited specs of the ADC available at the time, while newer machines have ADC with more bandwith which makes them sound good in a "what comes in comes out sense" but lacks the oomph of "vintage" tuned analog input/output stage.


Makes a lot sense.
ndkent
Some observations from experience. 24 bit is not going to sound amazingly different if you record your 16 bit recording at good levels, no clipping, playback more or less as is.

Where you hear huge differences are when you amplify lower level signals, mix lower level stuff together, hard to improve sound that lacks detail.

With the older gear there must be a little something going on somewhere. For instance I had a cassette recorder that sounded exceptional, turns out it had a decent limiter I didn't realize. Digital gear often has a rough and irritating high end, lower sample rate = more rolled off top due to input filters. Less familiar input metering can sometimes lead you to record low
boxxgrooved
Hainbach wrote:
I have not researched into the topic since my university days, but when I took digital signal processing classes the consensus was that the design of the analog filter stage before the conversion process plays a big role. If you have a lower sample rate the analog lowpass filter needs to be more agressive thus coloring the sound. My acoustic professor always brought up the early Lexicon reverbs and showed us the insides, stating that tuning the analog stages was important as the digital conversion. This goes both ways, AD/DA.

Thus something like the Sampletrack (lovely machine, also loves being recorded hot which adds pleasant drive) will have been tuned more to sound good with the limited specs of the ADC available at the time, while newer machines have ADC with more bandwith which makes them sound good in a "what comes in comes out sense" but lacks the oomph of "vintage" tuned analog input/output stage.


Does this simply mean that some devices are "tuned" better than others? A simple case of taking the time to tweak the various stages to achieve something pleasant sounding ... or not at all in some cases!

Excuse my lack of knowledge about AD/DAC's but on the Sampletrack for example you say there there is an agressive analogue filter rolling off the top end to reduce the bandwidth to 32khz. What happens on the output stage, will there be some more aggressive tuning or will it be left fairly open?

After reading your post it seems like to me some of the magic of the warmth of older units is the due to the more aggressive analogue filtering stage. I imagine the type of filters used in the conversion process also plays a massive role, after all different filters have totally different sounds.
Panason
Quote:
I can't get over how much better my 1999 Zoom Sampletrack sounds with its 32khz/18bit sampling engine compared to my newly aquired Elektron Digitakt running at 48khz/24 bit.


It has nothing to do with bitrate. Elektron's sound engines are just not very good.

Quote:
Why would Elektron go to all that length to create brilliant devices and then put crap converters inside them? Sound is more important than ANYTHING.


Money is more important than anything else for Elektron. They are expert at corner-cutting while maintaining a hi-tech image. There is nothing hi-tech about their boxes, just some clever software (at times a bit too clever for its own good). The hardware quality has actually decreased over the years ...
matthewjuran
Panason wrote:
It has nothing to do with bitrate. Elektron's sound engines are just not very good.

I’ll mention that I don’t share this opinion even though I questioned the mixing implementation. I haven’t listened to my Octatrack on large systems but in my small recording arena it’s been great. I’ve had the thought that putting synths on an external mixer instead of through the Octatrack inputs might lead to crisper synth tones but I haven’t even had arbitrary perceptions that prove that. It's set to 24bit.

What year was the hifi DAC design made? Availability of components matters and I’d expect there would be a quality improvement in surface mount signal processing the closer to today you get.
boxxgrooved
dubonaire wrote:


The Digitone is a 2-in, 4-out ADC/DAC and microprocessor with midi and USB that costs less than some low-mid range audio interfaces, and less than most dedicated DACs. So either we are all being taken for a total ride by the DAC makers, or Elektron has cut corners to keep the price down.


I do think decicated sound cards are expensive for what they do, I don't think the Digitone/Digitakt will necessarily have to use very low quality converters to make it pay. That being said it does sound like Elektron are using low grade converters or making a bad job of the entire conversion process as described above.

I doubt Zoom would have used better quality converters in 1999 in a budget sampler, it wouldn't make sense for them financially. If Zoom can milk a nice sound out of low grade converters why can't Elektron in 2017? There are only 2 answers to this question: either they don't care about it or the converters they use are total garbage.
dubonaire
matthewjuran wrote:
What year was the hifi DAC design made? Availability of components matters and I’d expect there would be a quality improvement in surface mount signal processing the closer to today you get.


That's a good point, it's not new. But here is the inside of a state of the art Schiit DAC which interestingly doesn't use delta-sigma D/A converters.

dubonaire
boxxgrooved wrote:
dubonaire wrote:


The Digitone is a 2-in, 4-out ADC/DAC and microprocessor with midi and USB that costs less than some low-mid range audio interfaces, and less than most dedicated DACs. So either we are all being taken for a total ride by the DAC makers, or Elektron has cut corners to keep the price down.


I do think decicated sound cards are expensive for what they do, I don't think the Digitone/Digitakt will necessarily have to use very low quality converters to make it pay. That being said it does sound like Elektron are using low grade converters or making a bad job of the entire conversion process as described above.

I doubt Zoom would have used better quality converters in 1999 in a budget sampler, it wouldn't make sense for them financially. If Zoom can milk a nice sound out of low grade converters why can't Elektron in 2017? There are only 2 answers to this question: either they don't care about it or the converters they use are total garbage.


I think Elektron will be just be using an ADC/DAC chip like a Cirrus which does everything on the chip and not doing anything else to improve DAC quality.
mousegarden
I think older digital gear can sound better than modern stuff, I'm using an old 80's 16 bit Philips CD player simply because I haven't found anything current that sounds as good.
Also weirdly, my old DX7 sounded warm in comparison to modern synths, despite the fact that I thought it sounded thin back in the day.
Also, discreet components are the key to a good sound IMO, no chips.
The inside of my lovely 70's Trio amp.

boxxgrooved
dubonaire wrote:


I think Elektron will be just be using an ADC/DAC chip like a Cirrus which does everything on the chip and not doing anything else to improve DAC quality.


This sounds like the equivalent in the analogue synth world to the Curtis chip. A whole synth voice on a small chip, convenient but thin sounding.

Like mousegarden said I think I'm starting to understand the importance of a signal passing through lots of discrete components before hitting the outputs.
MUFF WIGGLER Forum Index -> Production Techniques Goto page 1, 2  Next [all]
Page 1 of 2
Powered by phpBB © phpBB Group