Topic: Edirol R09 - ADC not good enough for 24bit, right?

[Brian Skalinder]

I see a lot of people suggesting that it's not worth running 24-bit because the dynamic range of their source / environment / recording gear is lower than the dynamic range achievable with 24-bit (or even 16-bit).  But is there not more to the value of 24-bit than dynamic range, i.e. greater precision through an increase in the number of vertical axis samples used to describe the waveform?  And this extra precision more closely approximates the analog waveform independently of the noise floor.

For some reason, I find this easier to follow using sample rate as an example:  the sample rate is analagous to bit-depth and the frequency range is analagous to dynamic range.  (I know this is not a perfect analogy, but it illustrates - I hope - the value of increased samples.)  A sample rate of 24 kHz provides a frequency range of 0 - 12 kHz.  A 48 kHz sample rate provides a frequency range of 0 - 24 kHz.  But even if our audio source only produces a frequency range of 0 - 12 kHz, at 48 kHz we're still using more samples (twice as many in this case) to define the waveform, i.e. we've achieved greater precision.  And the greater precision produces sonic improvements beyond merely frequency range, like better time coherence and spatial cues.

So, going back to bit-depth, the increased precision produces sonic improvements (detail) beyond merely expanded dynamic range.

More discussion here:  http://taperssection.com/index.php/topic,77804.0.html

[Javier Cinakowski]

That makes sense Brian, thanks.  I am still not sure it is worth it to me though.....  Now, if I get a DVD-A player for my home theater and car, then I will get a 24bit recorder....

[BayTaynt3d]

And that precision becomes even more important as you add digital gain (aka normalizing/compressing) doesn't it? Because you are effectively digitally zooming in on the waveform.

[Gutbucket]

FWIW, I 'think' my 24bit recordings with the R-09 sound better than the handfull of 16 bit ones I tried when I first got a hold of it.  I could be fooling myself though because I haven't done serious critical listening comparisons.  Keep in mind that I'm talking about recording acoustic music or more intimate scale electric stuff up close (stage lip or nearly so) DPA4060>MMA6K>R09 and using computer based play back where various bit depth files aren't a problem.  For me the peace of mind knowing that I can set levels to peak at -6db and not have to worry about overs and that I may be capturing some extra resolution that my current playback system may not resolve anyway is worth the slightly larger file size. [shrug]

edit-
I honestly am not as convinced it would make as much difference with more racous stuff where most of the sound is filtered through a club's PA and the dynamic range of the music isn't so large.  I tape that stuff too but that's not where my rig really shines or where my recordist soul soars.. Gets my butt groovin' though.

[poorlyconditioned]

FWIW, I 'think' my 24bit recordings with the R-09 sound better than the handfull of 16 bit ones I tried when I first got a hold of it.  I could be fooling myself though because I haven't done serious critical listening comparisons.  Keep in mind that I'm talking about recording acoustic music or more intimate scale electric stuff up close (stage lip or nearly so) DPA4060>MMA6K>R09 and using computer based play back where various bit depth files aren't a problem.  For me the peace of mind knowing that I can set levels to peak at -6db and not have to worry about overs and that I may be capturing some extra resolution that my current playback system may not resolve anyway is worth the slightly larger file size. [shrug]

I run at -6dB or so too.  That is just good sense.

The problem with going 24 bit, though, is the last nine bits are so are all noise!  I put a 1k resistor on the input, set to record line-in, at the lowest gain, and I still get a -90dB noise floor!  The noise floor starts to increase a bit on mic in (low sens), but only after about #25/30 or so.  Anything before that and you're still getting 90dB of SNR.

OK, 16 bits is -96dB noise floor.  So why grab another eight bits of noise floor running 24bit?  Maybe there is some signal in there, but it is pretty much buried in the noise if you ask me.

Anyway, like I said, this thing sounds good at 16bit, and the interface is awesome.  So it is definitely recommended.

  Richard

[Gutbucket]

I run at -6dB or so too.  That is just good sense.
...OK, 16 bits is -96dB noise floor.  So why grab another eight bits of noise floor running 24bit?  Maybe there is some signal in there, but it is pretty much buried in the noise if you ask me...

Call it a healthier than normal dose of dither if you like  .  I'm quite happy with the results too, and that's what counts most.. until I get a chance to to a serious listening comparison on material where I think it may make a difference I err on the side of caution.

That's the pragmatist in me talking, elbowing out the theorist for now.

edit-
The pragmatist would like me to add that sometimes I end up closer to -12db than -6db depending on the situation.  The meters are pretty compressed up there and sometimes I don't care to look at them anyway .

[it-goes-to-eleven]

I agree with Brian on 24 bit being more about the detail and less about the increased dynamic range. Sort of like coloring with 8 shades for every color vs. 128 shades of every color..

One of the sound devices guys once said measuring with a resistor is bad because you're just measuring the tolerance of the resistor.  I'm not sure if that is true..

[Church-Audio]

FWIW, I 'think' my 24bit recordings with the R-09 sound better than the handfull of 16 bit ones I tried when I first got a hold of it.  I could be fooling myself though because I haven't done serious critical listening comparisons.  Keep in mind that I'm talking about recording acoustic music or more intimate scale electric stuff up close (stage lip or nearly so) DPA4060>MMA6K>R09 and using computer based play back where various bit depth files aren't a problem.  For me the peace of mind knowing that I can set levels to peak at -6db and not have to worry about overs and that I may be capturing some extra resolution that my current playback system may not resolve anyway is worth the slightly larger file size. [shrug]

I run at -6dB or so too.  That is just good sense.

The problem with going 24 bit, though, is the last nine bits are so are all noise!  I put a 1k resistor on the input, set to record line-in, at the lowest gain, and I still get a -90dB noise floor!  The noise floor starts to increase a bit on mic in (low sens), but only after about #25/30 or so.  Anything before that and you're still getting 90dB of SNR.

OK, 16 bits is -96dB noise floor.  So why grab another eight bits of noise floor running 24bit?  Maybe there is some signal in there, but it is pretty much buried in the noise if you ask me.

Anyway, like I said, this thing sounds good at 16bit, and the interface is awesome.  So it is definitely recommended.

  Richard

Just got a software update from German company that makes my sound analysis software its the shit.. Check this out.. It has an auto test feature now that will do a bunch of tests for me automatically. This is my sound card running at 96k sampling rate at 24 bit they are working on a 192k 24 bit version they are going to give me soon here is the results of my M-Audio Audiophile 192k sound card.
There is a new Asio Loopback update I am waiting for so we can compare the digital end with the analog that is represented here in this graph then you can see what I was talking about with the real problem being the front end of all these devices.. I am stoked..

http://h1.ripway.com/churchaudio/SweepReport.htm

[poorlyconditioned]

FWIW, I 'think' my 24bit recordings with the R-09 sound better than the handfull of 16 bit ones I tried when I first got a hold of it.  I could be fooling myself though because I haven't done serious critical listening comparisons.  Keep in mind that I'm talking about recording acoustic music or more intimate scale electric stuff up close (stage lip or nearly so) DPA4060>MMA6K>R09 and using computer based play back where various bit depth files aren't a problem.  For me the peace of mind knowing that I can set levels to peak at -6db and not have to worry about overs and that I may be capturing some extra resolution that my current playback system may not resolve anyway is worth the slightly larger file size. [shrug]

I run at -6dB or so too.  That is just good sense.

The problem with going 24 bit, though, is the last nine bits are so are all noise!  I put a 1k resistor on the input, set to record line-in, at the lowest gain, and I still get a -90dB noise floor!  The noise floor starts to increase a bit on mic in (low sens), but only after about #25/30 or so.  Anything before that and you're still getting 90dB of SNR.

OK, 16 bits is -96dB noise floor.  So why grab another eight bits of noise floor running 24bit?  Maybe there is some signal in there, but it is pretty much buried in the noise if you ask me.

Anyway, like I said, this thing sounds good at 16bit, and the interface is awesome.  So it is definitely recommended.

  Richard

Just got a software update from German company that makes my sound analysis software its the shit.. Check this out.. It has an auto test feature now that will do a bunch of tests for me automatically. This is my sound card running at 96k sampling rate at 24 bit they are working on a 192k 24 bit version they are going to give me soon here is the results of my M-Audio Audiophile 192k sound card.
There is a new Asio Loopback update I am waiting for so we can compare the digital end with the analog that is represented here in this graph then you can see what I was talking about with the real problem being the front end of all these devices.. I am stoked..

http://h1.ripway.com/churchaudio/SweepReport.htm

Can you give a URL to that software?  Some of us here might try it out...

  Richard

[Church-Audio]

FWIW, I 'think' my 24bit recordings with the R-09 sound better than the handfull of 16 bit ones I tried when I first got a hold of it.  I could be fooling myself though because I haven't done serious critical listening comparisons.  Keep in mind that I'm talking about recording acoustic music or more intimate scale electric stuff up close (stage lip or nearly so) DPA4060>MMA6K>R09 and using computer based play back where various bit depth files aren't a problem.  For me the peace of mind knowing that I can set levels to peak at -6db and not have to worry about overs and that I may be capturing some extra resolution that my current playback system may not resolve anyway is worth the slightly larger file size. [shrug]

I run at -6dB or so too.  That is just good sense.

The problem with going 24 bit, though, is the last nine bits are so are all noise!  I put a 1k resistor on the input, set to record line-in, at the lowest gain, and I still get a -90dB noise floor!  The noise floor starts to increase a bit on mic in (low sens), but only after about #25/30 or so.  Anything before that and you're still getting 90dB of SNR.

OK, 16 bits is -96dB noise floor.  So why grab another eight bits of noise floor running 24bit?  Maybe there is some signal in there, but it is pretty much buried in the noise if you ask me.

Anyway, like I said, this thing sounds good at 16bit, and the interface is awesome.  So it is definitely recommended.

  Richard

Just got a software update from German company that makes my sound analysis software its the shit.. Check this out.. It has an auto test feature now that will do a bunch of tests for me automatically. This is my sound card running at 96k sampling rate at 24 bit they are working on a 192k 24 bit version they are going to give me soon here is the results of my M-Audio Audiophile 192k sound card.
There is a new Asio Loopback update I am waiting for so we can compare the digital end with the analog that is represented here in this graph then you can see what I was talking about with the real problem being the front end of all these devices.. I am stoked..

http://h1.ripway.com/churchaudio/SweepReport.htm

Can you give a URL to that software?  Some of us here might try it out...

  Richard

Yes sorry here it is http://www.dr-jordan-design.de/Downloads.htm this is a download of the basic version I am running the Pro EX version with all the plugins.. Its pretty dam expensive for the whole thing but you can pick and chose what you want. Here is a screen shot of the program.

[Church-Audio]

FWIW, I 'think' my 24bit recordings with the R-09 sound better than the handfull of 16 bit ones I tried when I first got a hold of it.  I could be fooling myself though because I haven't done serious critical listening comparisons.  Keep in mind that I'm talking about recording acoustic music or more intimate scale electric stuff up close (stage lip or nearly so) DPA4060>MMA6K>R09 and using computer based play back where various bit depth files aren't a problem.  For me the peace of mind knowing that I can set levels to peak at -6db and not have to worry about overs and that I may be capturing some extra resolution that my current playback system may not resolve anyway is worth the slightly larger file size. [shrug]

I run at -6dB or so too.  That is just good sense.

The problem with going 24 bit, though, is the last nine bits are so are all noise!  I put a 1k resistor on the input, set to record line-in, at the lowest gain, and I still get a -90dB noise floor!  The noise floor starts to increase a bit on mic in (low sens), but only after about #25/30 or so.  Anything before that and you're still getting 90dB of SNR.

OK, 16 bits is -96dB noise floor.  So why grab another eight bits of noise floor running 24bit?  Maybe there is some signal in there, but it is pretty much buried in the noise if you ask me.

Anyway, like I said, this thing sounds good at 16bit, and the interface is awesome.  So it is definitely recommended.

  Richard

Just got a software update from German company that makes my sound analysis software its the shit.. Check this out.. It has an auto test feature now that will do a bunch of tests for me automatically. This is my sound card running at 96k sampling rate at 24 bit they are working on a 192k 24 bit version they are going to give me soon here is the results of my M-Audio Audiophile 192k sound card.
There is a new Asio Loopback update I am waiting for so we can compare the digital end with the analog that is represented here in this graph then you can see what I was talking about with the real problem being the front end of all these devices.. I am stoked..

http://h1.ripway.com/churchaudio/SweepReport.htm

Can you give a URL to that software?  Some of us here might try it out...

  Richard

I was talking to Dr Jordan tonight and he sent me the 192k 1k notch filter update so now I can measure noise at 24bit 192khz here is what my test of my sound card looked like.. I thought you might be interested in this since we were talking about digital and how -120 seems to be ok.. I was wrong about nothing being able to have a -140 and be analog this is a simple loop back test of the analog input connected to its own output and it shows what 24 bit can really do at 192khz.

[SparkE!]

FWIW, I 'think' my 24bit recordings with the R-09 sound better than the handfull of 16 bit ones I tried when I first got a hold of it.  I could be fooling myself though because I haven't done serious critical listening comparisons.  Keep in mind that I'm talking about recording acoustic music or more intimate scale electric stuff up close (stage lip or nearly so) DPA4060>MMA6K>R09 and using computer based play back where various bit depth files aren't a problem.  For me the peace of mind knowing that I can set levels to peak at -6db and not have to worry about overs and that I may be capturing some extra resolution that my current playback system may not resolve anyway is worth the slightly larger file size. [shrug]

I run at -6dB or so too.  That is just good sense.

The problem with going 24 bit, though, is the last nine bits are so are all noise!  I put a 1k resistor on the input, set to record line-in, at the lowest gain, and I still get a -90dB noise floor!  The noise floor starts to increase a bit on mic in (low sens), but only after about #25/30 or so.  Anything before that and you're still getting 90dB of SNR.

OK, 16 bits is -96dB noise floor.  So why grab another eight bits of noise floor running 24bit?  Maybe there is some signal in there, but it is pretty much buried in the noise if you ask me.

Anyway, like I said, this thing sounds good at 16bit, and the interface is awesome.  So it is definitely recommended.

  Richard

Just got a software update from German company that makes my sound analysis software its the shit.. Check this out.. It has an auto test feature now that will do a bunch of tests for me automatically. This is my sound card running at 96k sampling rate at 24 bit they are working on a 192k 24 bit version they are going to give me soon here is the results of my M-Audio Audiophile 192k sound card.
There is a new Asio Loopback update I am waiting for so we can compare the digital end with the analog that is represented here in this graph then you can see what I was talking about with the real problem being the front end of all these devices.. I am stoked..

http://h1.ripway.com/churchaudio/SweepReport.htm

Can you give a URL to that software?  Some of us here might try it out...

  Richard

I was talking to Dr Jordan tonight and he sent me the 192k 1k notch filter update so now I can measure noise at 24bit 192khz here is what my test of my sound card looked like.. I thought you might be interested in this since we were talking about digital and how -120 seems to be ok.. I was wrong about nothing being able to have a -140 and be analog this is a simple loop back test of the analog input connected to its own output and it shows what 24 bit can really do at 192khz.

Careful, now... What you are showing there is spectral density.  You have to integrate over frequency to know what noise level that corresponds to and I'm betting you are looking at something more in the -110 to -115 range when you integrate across the spectrum you've shown.

[Church-Audio]

FWIW, I 'think' my 24bit recordings with the R-09 sound better than the handfull of 16 bit ones I tried when I first got a hold of it.  I could be fooling myself though because I haven't done serious critical listening comparisons.  Keep in mind that I'm talking about recording acoustic music or more intimate scale electric stuff up close (stage lip or nearly so) DPA4060>MMA6K>R09 and using computer based play back where various bit depth files aren't a problem.  For me the peace of mind knowing that I can set levels to peak at -6db and not have to worry about overs and that I may be capturing some extra resolution that my current playback system may not resolve anyway is worth the slightly larger file size. [shrug]

I run at -6dB or so too.  That is just good sense.

The problem with going 24 bit, though, is the last nine bits are so are all noise!  I put a 1k resistor on the input, set to record line-in, at the lowest gain, and I still get a -90dB noise floor!  The noise floor starts to increase a bit on mic in (low sens), but only after about #25/30 or so.  Anything before that and you're still getting 90dB of SNR.

OK, 16 bits is -96dB noise floor.  So why grab another eight bits of noise floor running 24bit?  Maybe there is some signal in there, but it is pretty much buried in the noise if you ask me.

Anyway, like I said, this thing sounds good at 16bit, and the interface is awesome.  So it is definitely recommended.

  Richard

Just got a software update from German company that makes my sound analysis software its the shit.. Check this out.. It has an auto test feature now that will do a bunch of tests for me automatically. This is my sound card running at 96k sampling rate at 24 bit they are working on a 192k 24 bit version they are going to give me soon here is the results of my M-Audio Audiophile 192k sound card.
There is a new Asio Loopback update I am waiting for so we can compare the digital end with the analog that is represented here in this graph then you can see what I was talking about with the real problem being the front end of all these devices.. I am stoked..

http://h1.ripway.com/churchaudio/SweepReport.htm

Can you give a URL to that software?  Some of us here might try it out...

  Richard

I was talking to Dr Jordan tonight and he sent me the 192k 1k notch filter update so now I can measure noise at 24bit 192khz here is what my test of my sound card looked like.. I thought you might be interested in this since we were talking about digital and how -120 seems to be ok.. I was wrong about nothing being able to have a -140 and be analog this is a simple loop back test of the analog input connected to its own output and it shows what 24 bit can really do at 192khz.

Careful, now... What you are showing there is spectral density.  You have to integrate over frequency to know what noise level that corresponds to and I'm betting you are looking at something more in the -110 to -115 range when you integrate across the spectrum you've shown.

Oops... I knew it looked too good. You are correct sorry about that. The figure is -108 you were very close. I was WAY OFF.. There is still a huge learning curve with this program I just got the noise plugin a few days ago.

[Gutbucket]

One of my concerns is that I don't understand the stuff enough in depth to really know what I'm measuring, if I'm measuring the right thing, what is relevent data & what is not and the true implications of all the various measurment options.  Not to mention what the measurement really means.

Powerful tools indeed, but tricky to wield with insight & understanding.

At least I can take comfort in Socrates (filtered through Plato), "The wize man is he who admits he does not know."  So I consider it a healthy concern.

[Church-Audio]

One of my concerns is that I don't understand the stuff enough in depth to really know what I'm measuring, if I'm measuring the right thing, what is relevent data & what is not and the true implications of all the various measurment options.  Not to mention what the measurement really means.

Powerful tools indeed, but tricky to wield with insight & understanding.

At least I can take comfort in Socrates (filtered through Plato), "The wize man is he who admits he does not know."  So I consider it a healthy concern.

That is so true I made a mistake that Sparke pointed out! and I own the dam program lol... Its a learning curve but its a very important tool. And if used properly it can tell you lots of things. But in the end there is no tool more powerful then your ears.

Chris Church