• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

PCM61P DAC chip info requested from a Dennon DCD-660

garbulky

Addicted to Fun and Learning
Joined
Feb 14, 2018
Messages
756
Likes
190
#21
Not that exact one but they are all similar and don't use a reconstruction filter, they produce as much distortion as a record player...
But how is that even possible? I thought a filter was essential to convert sound from digital to analog?
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,616
Likes
2,667
Location
Riverview, FL
#23
I thought a filter was essential to convert sound from digital to analog?
The DAC chip outputs voltage (or maybe current, which can be converted to voltage).

Given a high enough bandwidth of the circuitry after the DAC, a nice stair-stepped voltage "wave" would emerge.

Speakers (analog) play when varying voltage is applied.

Where's the filter?
 
Last edited:

gvl

Addicted to Fun and Learning
Joined
Mar 16, 2018
Messages
945
Likes
519
Location
SoCal
#24
The DAC chip outputs a voltage (or maybe current, which can be converted to voltage).

Given a high enough bandwith of the circuitry after the DAC, a nice stair-stepped voltage "wave" would emerge.

Speakers (analog) play when voltage is applied.

Where's the filter?
In your ear.
 

Frank Dernie

Major Contributor
Patreon Donor
Joined
Mar 24, 2016
Messages
1,313
Likes
1,401
Location
Oxfordshire
#25
Not really. Without a reconstruction filter the device isn't a properly engineered DAC and outputs a lot of spurious artefacts which are easily removed and should be. Nobody's ear can filter it out but since quite a lot of audio equipment fanatics seem to like a fair old bit of added distortion (LPs, SETs anybody?) plenty of people like it and cough up the cash.
 

gvl

Addicted to Fun and Learning
Joined
Mar 16, 2018
Messages
945
Likes
519
Location
SoCal
#26
Not really. Without a reconstruction filter the device isn't a properly engineered DAC and outputs a lot of spurious artefacts which are easily removed and should be. Nobody's ear can filter it out but since quite a lot of audio equipment fanatics seem to like a fair old bit of added distortion (LPs, SETs anybody?) plenty of people like it and cough up the cash.
That was just an answer to that specific question and not an argument in favor of omitting the filter. However the ear is a very good filter and given a 0% THD equipment with infinite bandwidth between it and the DAC it would be a good combo, sadly such equipment doesn't exist.
 

garbulky

Addicted to Fun and Learning
Joined
Feb 14, 2018
Messages
756
Likes
190
#27
Interesting. I didn't realize. I thought the PCM DA conversion and filters were inextricably linked. So if I'm looking at these measurements
https://www.stereophile.com/content...t-dac-21x-signature-da-processor-measurements
the lack of a reconstruction filter causes
1. high levels of ultrasonic distortion evidenced by a 25 khz mirror.
2. artifacts of -55 db and less.
3. Poor -90 db performance
4. Possibly unrelated poor channel matching at very low levels.

Though not competitive with other dacs, and potentially audible with certain types of recordings, this isn't a deal breaker for me. Most of my listening don't have dynamics that reach that much afaik or if they do my ear can't actually hear them. If I get a chance to listen to one, I think I shall try it. I have never heard a DAC without an actual filter. According to another forum user who has built the DAC 4.1
http://www.ankaudiokits.com/Non-Oversampling-Valve-Rectified-Tube-DAC.html
He thinks it sounds fantastic. However his listening preferences and mine tend to be different. He tends to love the sound of tubes. While my listening to the same products found me to not prefer them to a solid state amp.
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,616
Likes
2,667
Location
Riverview, FL
#29
Interesting. I didn't realize. I thought the PCM DA conversion and filters were inextricably linked.
The reconstruction filter (low pass) removes the high frequency content (what would be near vertical jumps between voltage levels become a ramp).

Contrived illustrative example:

upload_2018-4-24_17-43-25.png



Unfiltered and Low Pass Filtered Spectrum

upload_2018-4-24_17-38-51.png
upload_2018-4-24_17-39-56.png
 

garbulky

Addicted to Fun and Learning
Joined
Feb 14, 2018
Messages
756
Likes
190
#30
The reconstruction filter (low pass) removes the high frequency content (what would be near vertical jumps between voltage levels become a ramp).

Contrived illustrative example:

View attachment 12317


Unfiltered and Low Pass Filtered Spectrum

View attachment 12315 View attachment 12316
That's what I was wondering about. Without a filter the sine wave actually has steps in it right? versus a smooth wave? I'm imagining that's not good and that's why I assumed you HAD to have a filter. Also in the bottom two lowpass, what does the large "mountain" before the spikes indicate?
 

DonH56

Major Contributor
Technical Expert
Patreon Donor
Joined
Mar 15, 2016
Messages
2,654
Likes
2,668
Location
Monument, CO
#31
A delta-sigma DAC, which is most of the audio DACs these days, puts out a high-frequency pulse stream so looks even worse before filtering.
 

restorer-john

Major Contributor
Joined
Mar 1, 2018
Messages
1,770
Likes
2,456
Location
Gold Coast, Queensland, Australia
#32
As far as I see it, the removal or omission of a correctly engineered LPF simply creates an inherently faulty audio D/A converter. Not fit for purpose in any way. Dangerous to downstream components, amplifiers and speakers.

There has been much stupidity with dumb-ass 'mods', made by unqualified clowns masquerading as 'experts' and amassing mindless followers, more so in the last decade or two, than any other period I can think of.

The relative perfection of digital audio and its complete lack of 'tweakability' has led to the current level of misinformation and stupidity we find ourselves wading knee-deep through.

ASR- Draining The Audiophile Swamp of Misinformation, Lies and Deception. :)
 

DonH56

Major Contributor
Technical Expert
Patreon Donor
Joined
Mar 15, 2016
Messages
2,654
Likes
2,668
Location
Monument, CO
#33
^^^ Ah, the voice of one who has repaired "improvements" made by removing the coupling capacitors before power amps, realizing the dream of DC-coupled with no nasty capacitors in the signal path. DC Hz. ;) And of course the op-amp mods leading to oscillations and all that jazz. Now DAC image filters bad, better to pour ultrasonics into our tweeters, great.
 

gvl

Addicted to Fun and Learning
Joined
Mar 16, 2018
Messages
945
Likes
519
Location
SoCal
#34
Realistically the level of ultrasonics from a DAC w/no filter is fairly low, but I'd still resist the temptation to crank the volume up.
 

restorer-john

Major Contributor
Joined
Mar 1, 2018
Messages
1,770
Likes
2,456
Location
Gold Coast, Queensland, Australia
#35
...Realistically the level of ultrasonics from a DAC w/no filter is fairly low...
I would say the exact opposite. The significant amount of HF spuriae results in risk to the stability of amplifiers and the destruction of tweeters, and also inter modulation products in the audible range.

Have you actually FFTd the spectrum out of a D/A chip (after the IV stage)? Ray's depiction isn't far off the reality. It isn't pretty and it's not remotely musically related content- it is faulty design spuriae.

And if you couldn't safely 'crank the volume up', what apart from a completely faulty system have you created? It is the very antithesis of the pursuit of High Fidelity reproduction.
 

gvl

Addicted to Fun and Learning
Joined
Mar 16, 2018
Messages
945
Likes
519
Location
SoCal
#36
For the record, I'm not advocating filterless DACs. However, even though they are fairly wide-spread we don't see horror stories about exploding amplifiers and melting tweeters too often, if at all. That makes me think that while the dangers are technically there the probability of bad things happening is pretty low. The reason being the HF content in music is at fairly low levels, and the energy of resulting images is even lower. Sure, if you send a synthetic 0dB 20kHz digital tone to such a DAC all bets are off, but fortunately this doesn't happen with pretty much all real life music material. I agree, intermodulation distortion products may fold back into the audible band, that's probably the main issue IMO.
 
Last edited:

restorer-john

Major Contributor
Joined
Mar 1, 2018
Messages
1,770
Likes
2,456
Location
Gold Coast, Queensland, Australia
#37
I think a lot of the ancilliary equipment people tend to use with those faulty D/As tends to perform much of the LPF's job in the front end of the amplifier. What's left is killed by the combination of the HF feedback rolloff network, stopper caps on drivers/outputs, the zobel network and some HF shunting (hopefully) in their crossovers.

It's akin to every individual wearing a radiation suit, instead of shielding the reactors in the first place. :)
 

gvl

Addicted to Fun and Learning
Joined
Mar 16, 2018
Messages
945
Likes
519
Location
SoCal
#38
I think a lot of the ancilliary equipment people tend to use with those faulty D/As tends to perform much of the LPF's job in the front end of the amplifier. What's left is killed by the combination of the HF feedback rolloff network, stopper caps on drivers/outputs, the zobel network and some HF shunting (hopefully) in their crossovers.

It's akin to every individual wearing a radiation suit, instead of shielding the reactors in the first place. :)
Here is a pseudo-scientific article that touches on these and other points: https://www.criticalsound.co.nz/assets/Uploads/Design-Philosophy-Metrum-Acoustics.pdf
 

DonH56

Major Contributor
Technical Expert
Patreon Donor
Joined
Mar 15, 2016
Messages
2,654
Likes
2,668
Location
Monument, CO
#39
Realistically the level of ultrasonics from a DAC w/no filter is fairly low, but I'd still resist the temptation to crank the volume up.
The images follow a sinc (sin(x)/x) curve so signals in the upper midrange get folded around Nyquist (one-half the sampling frequency) and can appear above 20 kHz in "significant" amounts. The levels may indeed be "fairly low" but then again tweeters aren't designed to take a lot of power. I have an article on it, somewhere, but in any event have given up trying to convince anyone of anything in audio these days. More to the point, oversampled DACs (currently the vast majority of them) move the images well above the audio band where they are less likely to modulate a driver back into the audio band. Still likely to heat up the tweeter, of course.

This article touches on images: https://www.audiosciencereview.com/...ital-audio-converters-dacs-fundamentals.1927/
 
Last edited:

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,616
Likes
2,667
Location
Riverview, FL
#40
That's what I was wondering about. Without a filter the sine wave actually has steps in it right? versus a smooth wave? I'm imagining that's not good and that's why I assumed you HAD to have a filter. Also in the bottom two lowpass, what does the large "mountain" before the spikes indicate?
DonH gently informed me I don't really know what I'm talking about...

A delta-sigma DAC, which is most of the audio DACs these days, puts out a high-frequency pulse stream so looks even worse before filtering.
The actual modern delta-sigma DAC output (unfiltered) might look something like the red trace here if observed on an oscillosope:

upload_2018-4-25_12-23-40.png


I figure the traces are not "to scale" in the illustration above...

Without a filter you won't see the filtered result (which should closely resemble the original analog waveform). Stairsteps, or high-frequency varying width pulses, it won't match the original until further processing (the low pass filter).

In either case, fed to a speaker (which has low-pass properties), you'll get the sound, with added (unfiltered) HF content (if you can hear it or its effects), I would guesstimate.

---

The "mountain" is the 400Hz tone - FFT spreads the displayed width of lower frequencies when the window type and size are varied. Maybe I could have picked better, but it clearly showed the high frequency content of the steps vs the lack of them when the signal is tightly filtered just above the displayed fundamental frequency.

---

Maybe the "stairsteps" are applicable to a DAC type that would feed an R2R ladder? Or the output (unfiltered) of the ladder...

Don?
 
Last edited:

Similar threads

Top Bottom