Recall the bit where I proposed many years back that data appears across frames (which is/was complementary to what Teranex was doing for stills in the military sat image field) in the video world, when scaling DVD's 640x480 resolution up to 1080P. This, due to the original telecine process being complimentary to the up-scaling as the original was 1080P frame grabbing in the digital domain and then intelligently scaled down.
There is some extant but minor analogy to this situation here.
What you are seemingly getting here is the data sometimes being there and sometimes not.... but the ear/brain system is the most complex and capable real time re-construction system that the human body has and utilizes ..thus it is no surprise, to me, that such sonic characteristics can be 'heard' so deep into the noise floor.
There is some extant but minor analogy to this situation here.
What you are seemingly getting here is the data sometimes being there and sometimes not.... but the ear/brain system is the most complex and capable real time re-construction system that the human body has and utilizes ..thus it is no surprise, to me, that such sonic characteristics can be 'heard' so deep into the noise floor.
At 24 bits low level resolution is not an issue it's just a case of how close
sonically a 16 bit dithered down version comes to it's 24 bit counterpart, all
other things being equal.
I believe most mastering engineers will concur that with the best dithering
algorithms there's not a huge amount in it. Ref pic below of a VHQ DAC,
1kHz, -90dB tone at 16 and 24 bits.
cheers
Terry
Attachments
I would not take the whole Mastering community as an authority here. After all, they seemed to go silently along with the con that digital master tapes for 1980 LPs were an upgrade on the 1979 releases (thinking about the Decca and EMI classical releases, anyway). Comparing those LPs today shows just how much was thrown away in 16/44.1, for any kind of orchestral or choral music.
It was all the greater pity, since the 1979 Deccas, in particular, were spectacularly improved against the year before.
Last edited:
Take an input gain stage with the familiar TO220 MOSFETs at 60mA minimum, same +60mA biased output stage after volume attenuation, and multiply that by two for balanced operation.
Take a series stage, a parallel stage, add a push-pull regulating stage, another series stage (decide for yourself in which order), and bump up the raw Vdc of the powersupply by at least another 10 points.
Add a whole bunch of sealed relays for source switching and stepped balanced volume attenuation, the point where preamp kitchen things really start cooking and/or insanity becomes unsanitary.
I poked fun at the bit wars by asking about 64 bits. How many real 24 bit linear no missing code converters are there?
ES
Ah yes the devices intended for digital scales have great resolution! A bit sloooow.
None - of course. However quite a few people here, it appears, have the
misconception that only analog recording mediums, ie; tape machines
have the ability to store audio information well below the rms noise floor.
T
Nah, you would still need to capture and hold the samples with 24 bit resolution in level and time!
Not as easy as some would think.
Doesn't work that way. 48 converters sampling at 2kHz still don't see a 10kHz signal. I presume the 24 bits comes from a lot of noise BW limiting to prevent aliasing of the noise.
You would be better off simply averaging 48 "24" bit converters.
Last edited:
Yes, but they probably have some of the same error sources so it would not follow the normal improvement from parallel operation!
So Scott when can we see our 64 bit converters?
I'm tempted to say never, but, I'll play safe and say not in my lifetime.when can we see our 64 bit converters?
just one word... SQUIDs
I think if you extend some of the S/N, sample frequency graphs of the superconducting ADC tech back to audio range you start to see 24 bits
Hypres.com >technology>data converters
anyone got a few Mil US$ for a custom audio ADC or DAC?
the helium refrigerator keeps the Vodka cubes cool too
I think if you extend some of the S/N, sample frequency graphs of the superconducting ADC tech back to audio range you start to see 24 bits
Hypres.com >technology>data converters
anyone got a few Mil US$ for a custom audio ADC or DAC?
the helium refrigerator keeps the Vodka cubes cool too
Maybe, I hope you don't count me among them based on my previous post, if so, I am sorry to have not been clear. The point I was trying to make was that the RMS or any other noise floor specification is a broadband single number, and that masking criteria instead must be applied to frequency limited signals to determine their audibility above the noise. This applies to dithered digital systems as well as all analog systems.
I've experimented with the audibility of different dithering schemes since the late 1980's in my role as a replication engineer. Without dither 16 bit systems (all digital systems) have a nasty sounding noise floor for sure. With dither they can encode many more dB of noise-correlated signal...
Good topic...
In it's best current embodiment, when designed well, at high SR / bit depths,
it is way closer to perfect than analog could ever hope to be.
Whilst searching for elusive forms of low level distortion that somehow
correlate what we hear to measurements is valid, I think we first and
foremost, need to seek and aknowledge the correct answer to one simple
question:
Does the best analog sound fantastic in spite of it's various and complex non
linearities or because of them?
This interview with Kristian Zimerman reflects my views on a lot of this
stuff.
YouTube - Krystian Zimerman in interview (1/5)
YouTube - Krystian Zimerman in interview (2/5)
cheers
Terry
the idea that the digitalized reproduction via a DAC and some sort of filter arrangement is "closer to perfect" may not be quite correct. we've been spending a lot of time discussing exactly how ultra low distortion amps (and preamps, eh?) specifications do not correlate directly with the perceived subjective presentation?
I certainly wish for nearly perfect recording by whatever technological means - but the reality remains that the way that we hear things simply doesn't correlate well to measurement techniques just yet - so if and when it comes, I'll be quite pleased.
I think there is a very simple test that everyone who is motivated can try, any day, any time.
I did this years ago. It is fairly remarkable... take any two mics, string them OUTSIDE and listen inside with your speakers... it works best when the mics can be on opposite or adjacent sides of the building, but outside is the first criterion. Listen for a while...
Try to record it.
Get back to me on it... feel free to send copies of your outside if you wish...
Even using seemingly not-so-great mics and preamps, there is something that appears to me to be NOT captured in the recording process... YMMV, as always...
_-_-bear
I certainly wish for nearly perfect recording by whatever technological means - but the reality remains that the way that we hear things simply doesn't correlate well to measurement techniques just yet - so if and when it comes, I'll be quite pleased.
I think there is a very simple test that everyone who is motivated can try, any day, any time.
I did this years ago. It is fairly remarkable... take any two mics, string them OUTSIDE and listen inside with your speakers... it works best when the mics can be on opposite or adjacent sides of the building, but outside is the first criterion. Listen for a while...
Try to record it.
Get back to me on it... feel free to send copies of your outside if you wish...
Even using seemingly not-so-great mics and preamps, there is something that appears to me to be NOT captured in the recording process... YMMV, as always...
_-_-bear
- Status
- Not open for further replies.