Upsampling

[tuga]

And these from a RME ADI-2 DAC:

Any users of the RME ADI-2 DAC?

I currently have in my primary system an Ayre QB-9 (24/96 version, tried to get it upgraded, but after 3 years of the run around gave up). Have any of you compared it to an Ayre QB-9 or any of the better quality DAC's on the market today? What are your thoughts? At $1099 it appears to be a good b...

audiophilestyle.com

 

[tuga]

Unfortunately ASR's measurements are band limited to the audio range which makes them less capable of characterising performance.

 

[davidsrsb]

That may be due to poor implementation, poor algorithms or bcause the processing in the DAC is generating rubbish.
Measurements from the DAC in question, Holo Spring 3:

I was talking more about the ICs themselves, AD and TI (was BB) give very thorough datasheets.

 

[Alex S]

To the max (32/1.536PCM or DSD1024) will require an extremely powerful computer.

Yes. Max for my dacs is 32/768or 32/384 and that's no problem. I am yet to fiddle around with DSD. I have read that really powerful PCs can bug out using HQPlayer's most intensive settings so when I get round to it, I'll be interested how an M1Pro MacBook copes.

 

[tuga]

I was talking more about the ICs themselves, AD and TI (was BB) give very thorough datasheets.

HQPlayer's developer has published measurements of several DACs and they all measured betterwhen fed upsampled PCM and in the case of Sigma-Delta chips those performed even better when fed high-rate DSD (bypassing all chip internal DSP and also the SD modulator).

 

[Jim Audiomisc]

That may have been a problem in the past but these days most digital audio workstations and certainly all professional mastering engineers have easy access to true peak (dBTP) meters.

The level indicatiors have been able to do that for many decades. The problem tends to be 'recording gurus' who feel 'LOUDNESS SELLS' so they crank up the level and clip the output. Many Cds exhibit this. cf my webpages on the topic. I've had many arguments with record producers/gurus who have *defended* this behaviour on the basis that it is 'liked'. For them its an 'effect' which brings in more punters!

 

[tuga]

Yes. Max for my dacs is 32/768or 32/384 and that's no problem. I am yet to fiddle around with DSD. I have read that really powerful PCs can bug out using HQPlayer's most intensive settings so when I get round to it, I'll be interested how an M1Pro MacBook copes.

I have the standard 16GB M1 mini and I can easily do DSD256 with all filters and modulators and a bit of room EQ.
On the mini 16GB of RAM is mandatory to avoid pauses/drop-outs.

 

[Jim Audiomisc]

Rarely heard night and day differences with DACs.
.sjb

Same for me. For my main system I use a Benchmark ADC+DAC. But in other situations/cases I'm happy with a range from the early Cambridge Audio or even the Scarlett 2i2 upwards. Walking from one system/room to another has vastly more impact on what I hear.

 

[tuga]

Same for me. For my main system I use a Benchmark ADC+DAC. But in other situations/cases I'm happy with a range from the early Cambridge Audio or even the Scarlett 2i2 upwards. Walking from one system/room to another has vastly more impact on what I hear.

I wonder whether differences would be more audible if you had a single system and room as your reference.

 

[Jim Audiomisc]

Another pitfall to beware here is that the actual meaning of 'NOS' in practice will vary with real systems.
If you upsample 48k source material to a rate N times higher simply by repeating each sample N times then the resulting analogue output is a series of flat-topped 'stairs' which can generate a lot of hf crap. UNLESS you have a good analogue filter to run that though.

'Normal' upsampling sets out to smoothly regenerate what the samples define in terms of Nyquist. i.e. no such steps at the 'base rate', but smaller ones at the upsampled rate. Which means lower ultrasonic 'distortion' and an easier job for a following analogue fulter.

 

[tuga]

@davidsrsb on-board upsampling presents a few challenges, namely limited processing power and the processor generating noise inside the DAC or even chip.

AKM is now spliiting the D/A and the digital filter & modulator into separate chips:

AKM releases to production the new flagship D/A converter (DAC) solution AK4499EX and AK4191 for a complete separation of digital and analog circuitry. | Archives 2022 | News | About AKM | Asahi Kasei Microdevices (AKM)

AKM releases to production the new flagship D/A converter (DAC) solution AK4499EX and AK4191 for a complete separation of digital and analog circuitry.

www.akm.com

 

[Jim Audiomisc]

I wonder whether differences would be more audible if you had a single system and room as your reference.

FWIW I used to run two systems in different rooms using ESLs and two running LS3/5as. That's reduced these days. But I still swap kit about at times. My point was that compared with things like speakers and rooms the effect of choice of DAC is almost irrelevant once the DAC is half-decent. That wasn't true many decades ago. But things have moved on. YMMV

 

[tuga]

Another pitfall to beware here is that the actual meaning of 'NOS' in practice will vary with real systems.
If you upsample 48k source material to a rate N times higher simply by repeating each sample N times then the resulting analogue output is a series of flat-topped 'stairs' which can generate a lot of hf crap. UNLESS you have a good analogue filter to run that though.

'Normal' upsampling sets out to smoothly regenerate what the samples define in terms of Nyquist. i.e. no such steps at the 'base rate', but smaller ones at the upsampled rate. Which means lower ultrasonic 'distortion' and an easier job for a following analogue fulter.

This topic is about external upsampling which makes sense when using a NOS DAC; it's not about reproducing NOS Redbook (16.44.1).

Some D/A chip allow bypassing of DSP and SDM stages when fed DSD which may be considered NOS.
Other chips like the ESS will further upsample any input.

 

[Tony L]

The level indicatiors have been able to do that for many decades. The problem tends to be 'recording gurus' who feel 'LOUDNESS SELLS' so they crank up the level and clip the output. Many Cds exhibit this.

Annoyingly the really loud mastering trend mostly remains, though it does seem they are getting far better at it. Most of the recent pop/rock CDs I’ve bought (not many, I tend to buy vinyl) have been really loud, e.g. Kendrick Lamar Mr Morale, Sufjan Stevens Javelin etc, but the awful harshness and glare of the typical ‘loudness wars’ crap of the late-90s and early-00s seems not to exist anymore.

I haven’t looked at either of the discs I cite in Audacity, but since say Daft Punk’s Random Access Memories I think loud CDs have got a lot better (the Daft Punk being a very loud CD that subjectively sounds very good, certainly no harshness, digital glare etc). My guess is the industry has learned to use far better compressors and limiters etc at the digital mastering stage to prevent overshoot. The days of something sounding like absolute shite like say a Red Hot Chilli Peppers or Muse CD is hopefully behind us!

I’d still prefer they cut stuff a fair few db quieter though. The problem is humans equate ‘louder’ with ‘better’. That’s just a fact, I’ve seen it in dealer dem rooms so many times.

 

[tuga]

FWIW I used to run two systems in different rooms using ESLs and two running LS3/5as. That's reduced these days. But I still swap kit about at times. My point was that compared with things like speakers and rooms the effect of choice of DAC is almost irrelevant once the DAC is half-decent. That wasn't true many decades ago. But things have moved on. YMMV

I agree that room effects and speaker impact on reproduction is much more significant.
But, for me, a half-decent DAC is not enough if I can have better. I only have a single system and my goal is to get the best performance for a given budget (and taking into account family constraints as well as the fact that I have been moving house every 3 years).

 

[gustav_errata]

The level indicatiors have been able to do that for many decades. The problem tends to be 'recording gurus' who feel 'LOUDNESS SELLS' so they crank up the level and clip the output. Many Cds exhibit this. cf my webpages on the topic. I've had many arguments with record producers/gurus who have *defended* this behaviour on the basis that it is 'liked'. For them its an 'effect' which brings in more punters!

Yes, they "crank it up" via digital mastering limiters which have the threshold set in dBTP. I suspect your CD examples are outdated but I'm happy to be proven wrong. I'm sure there are still examples of people doing it wrong, but it should be significantly less common now.

 

[tuga]

Not that recent but this is the first track of R.E.M.'s Green album, "Pop Song '89".

From the original 1988 Green CD (DR12)

From the 2011 Part Lies, Part Garbage compilation CD (DR5)

 

[Tony L]

Not that recent but this is the first track of the Green album, Pop Song '89.

I don’t understand how to read the second chart, it doesn’t look possible, the plot looks like just half the AC waveform! It’s an asymmetrical clipping fuzz-box on full!

 

[Alex S]

I wonder whether differences would be more audible if you had a single system and room as your reference.

I have a single room with two very different systems in the same place - valves/ZUs and Genelec monitors. I can use five dacs, three relatively similar and two very different in terms of build (2 ESS, one AKM, one TDA 1541 transformer coupled and one PCM 1704 with valve buffer). They all sound far more alike than different. Bizarrely, the systems sound quite similar as well although the bigger speakers fill the room more, but I sit 2m away anyway. Sitting closer to all of my speakers is by far the biggest (and best) change I've made over recent years.

 

[tuga]

I have a single room with two very different systems in the same place - valves/ZUs and Genelec monitors. I can use five dacs, three relatively similar and two very different in terms of build (2 ESS, one AKM, one TDA 1541 transformer coupled and one PCM 1704 with valve buffer). They all sound far more alike than different. Bizarrely, the systems sound quite similar as well although the bigger speakers fill the room more, but I sit 2m away anyway. Sitting closer to all of my speakers is by far the biggest (and best) change I've made over recent years.

If you have 2 systems in a single room and 5 DACs then my assumption regarding not having a reference still applies.

I agree about very farfield listening, although constant directivity speakers such as horns or dipoles are not particularly affected by room interference above the threshold (Schroeder) frequency due to their constant & narrow directivity (also D&D and Kii).