Thoughts on a bunch of DACs (and why delta-sigma kinda sucks, just to get you to think about stuff)
Mar 7, 2015 at 8:24 PM Post #2,911 of 6,500
 
that might be part of why they aren't doing 16x oversampling, they seem to only be doing 8x oversampling according to what mike said about things being oversampled to 352 or 384khz, though that could also have been a limitation in the computational ability of the DSP they are using...who knows


Ah thanks for that, I dunno where I got the idea they're doing 16X OS from. Makes more sense.
 
Mar 7, 2015 at 11:48 PM Post #2,912 of 6,500

Analog to Digital Converter Shootout

http://www.analogplanet.com/content/analog-digital-converter-shootout-which-sounds-best
 
 
Lowest to highest price, the A/D converters are the HRT Linestreamer+, ($299.95), the P.S. Audio NuWave Phono Converter ($1,899), the Channel D Seta Piccola ($1,899, or $2098 as supplied including optional RIAA module), the M2TECH Joplin ($2,499), the Lynx Hilo ($2499) and the Ayre QA-9 ($3995).
 
Mar 8, 2015 at 10:01 AM Post #2,913 of 6,500
It does get a bit technical - for more technical background, go see my blog on DIYA which I linked to further up.

Here goes for a non-techy explanation. The Yggy DAC's designed to give the best performance with low frequencies - that's the application its intended for, where the output doesn't change at a fast rate. Think of it as a low sample frequency DAC where it only gets its stellar performance when updated relatively slowly (less than 10,000 times per second). Because of this, when the DAC's output changes, its spending a fair bit of time on its journey to the new commanded output. A deglitcher is used to remember the DAC's output and provide a copy of it to the following circuit while the DAC itself is changing its output. Think of the deglitcher as a switch and a memory of the last sample out of the DAC - when the DAC's told to update, the switch is opened and the memory element (normally a capacitor) provides the voltage to the downstream circuit. Then when the DAC's output has fully settled, the memory is updated with the new value.

Thx, makes a lot more sense now. Kinda like a buffer.
Your OS abbreviation was also pretty confusing...in my IT-centered world that stands for Operating System :)
 
Mar 8, 2015 at 10:20 AM Post #2,914 of 6,500

Analog to Digital Converter Shootout



http://www.analogplanet.com/content/analog-digital-converter-shootout-which-sounds-best


[COLOR=333333]Lowest to highest price, the A/D converters are the HRT Linestreamer+, ($299.95), the P.S. Audio NuWave Phono Converter ($1,899), the Channel D Seta Piccola ($1,899, or $2098 as supplied including optional RIAA module), the M2TECH Joplin ($2,499), the Lynx Hilo ($2499) and the Ayre QA-9 ($3995).[/COLOR]

An interesting test. The poll results are however already flawed because they allowed people to comment on the samples. Anyone who reads the comments will be flawed. I did not try but my bet is on the "all sound the same" option winning comfortably.

Here's another Dac shootout http://www.tomshardware.com/reviews/high-end-pc-audio,3733.html. From the $2 Dac on the PC mboard to the $2000 Benchmark. The results may surprise many but I cant really see any flaw in their test procedure.

Darko also keeps a nice Dac review index http://www.digitalaudioreview.net/the-darko-dac-index/. His review style is a bit too much "audiophile" for my taste but he is still quite moderate when it comes to hearing audio unicorns.
 
Mar 8, 2015 at 2:09 PM Post #2,915 of 6,500
- Fair enough on the stuff one notices / pays attention being people dependent (and you seem to have a particularly solid ear)
- About the ladder precision thing, let say I will remain skeptic until I can hear this for myself.
- 14bits effective resolution is 80dB plus of dynamic range.
- That means this "Y-axis jitter" term you're using comes into play for signals that are 80dB below the peak value at any given time block.
- You're thus saying you can hear these "fly farts within the orchestra


You are still not getting this. the y-axis jitter analogy applies to all bitcodes that represent the entire range, not just the low level signals.
 
In the end, it's not about dog farts in the orchestra. There is certain ambient, microdynamic, and harmonic information that disappears when amplitude data is "fuzzied". Recordings sound a little less live, less immediate, less real.

In the end, the audiophile's pursuit to have the recording fool him into thinking it's real.
 
Mar 8, 2015 at 3:23 PM Post #2,916 of 6,500
What's a deglitcher?! And generally, would you care to explain your post a bit? Not all of us have EE degrees


For resistor ladder D to A, the outputs for each bit are switched. When a zero gets switched to a one, there is a sudden inrush current. A spike, which then takes some time to settle. Kinda like your vacuum cleaner.

This behavior is not always the same. Changes in the bitcodes of your WAV files vary. Worse case glitch occurs when a lot of bits get flipped at once. More bits flipped, more switches flipped, sharper spike, longer to settle.

Glitching tends to be worse on non-audio chips.

One method is to build a sample and hold amplifier. But as Mike said, he feels sample and hold sounds like moose-butt.

Interestingly enough, Jason answered my question on how they dealt with the glitch, right before I asked the question. He probably figured I had read the spec sheet. I'm not telling; and I think it would be presumptuous to think that Mike and company didn't know what they were doing compared to us Internet armchair DAC designers.

But guessing is fun. I know I had fun trying to guess the chips used based on hints Mike left on the Internet. There are hints in the Yggy's architecture on how the glitch was dealt with.
 
Mar 8, 2015 at 4:13 PM Post #2,917 of 6,500
An interesting test. The poll results are however already flawed because they allowed people to comment on the samples. Anyone who reads the comments will be flawed. I did not try but my bet is on the "all sound the same" option winning comfortably.

Here's another Dac shootout http://www.tomshardware.com/reviews/high-end-pc-audio,3733.html. From the $2 Dac on the PC mboard to the $2000 Benchmark. The results may surprise many but I cant really see any flaw in their test procedure.

Darko also keeps a nice Dac review index http://www.digitalaudioreview.net/the-darko-dac-index/. His review style is a bit too much "audiophile" for my taste but he is still quite moderate when it comes to hearing audio unicorns.

My opinion on tomshardware dropped more than 50% because of that article. It doesn't deserve any attention and any link on this thread.
 
Mar 8, 2015 at 4:35 PM Post #2,918 of 6,500
  My opinion on tomshardware dropped more than 50% because of that article. It doesn't deserve any attention and any link on this thread.

 
I don't have have a problem with the results, but I have a problem with context and framing of the results (picture of expensive McIntosh tube amp to pander to the "overpriced expensive stuff" agenda (the McIntosh amp had nothing to do with the test), running from the built-in headouts, and comparing stuff which mostly really does sound more or less the same. But you have consider a the audience: It's rather trendy thing for people in their twenties who still live with their parents to rail against the "overpriced arena of ******** that is the field of 'high-end" audio'. Part of this an extension of the nwavguy movement, which in turn is a modern take on of the the Doug Self movement decades ago.
 
Lest we forget, we only need to go a few posts back in this very thread: http://www.head-fi.org/t/693798/thoughts-on-a-bunch-of-dacs-and-why-i-hate-chocolate-ice-cream/2850#post_11386526.
 
You won't see me going to Bimmer or Audi forums calling people out for being tards on buying "overpriced expensive stuff." To each his own.
 
Mar 8, 2015 at 4:55 PM Post #2,919 of 6,500
Quote:
 
- Fine, but there are concepts like level masking that long exist and prove this makes little sense.


 
Yes, the question whether these small inaccuracies make any difference when the bitcodes indicate higher levels is valid.
 
What I would ask you to do is examine the bitcodes in WAV files of good quality recordings with a lot of dynamic range. You would be surprised how low/small the absolute amplitude/bitcodes tend to be at. I certainly was. Also, you have to realize that even a high level 0dBFS sinewave is composed of bitcodes that run from 0 to 32767 back to 0 down to -32767 back up to 0, et. al. 
 
As far as your doubts whether 16-bits is necessary or not, I guess through your own powers of intuition that you know better than Sony and Philips who came up with the 16/44 Redbook spec, or Philips who realized the TDA1540 at 14-bits was too sucky and needed to make a more accurate 16-bit chip, or the professional recording industry who didn't like the 13/14 bits of information left over from processing in 16-bits, or Mike Mofatt who pioneered the standalone DAC / use of milspec chips (GAIN) two decades ago.
 
Personally, I'll place my in trust Sony, Philips, Mofatt, (and my own ears which tell me 18-bits is about right for me for playback) than your belief that 16-bits is unnecessary and 14 honest bits with 2 fuzzy fits is all that we need. Honestly, you just sound like an apologize for TotalDAC at this point. It is what it is. No need to defend it. I don't disbelieve you when you said the totalDAC sounded great with your 10 minutes or whatever gear and recordings you used during your audition.
 


 
Mar 8, 2015 at 6:30 PM Post #2,920 of 6,500
What I would ask you to do is examine the bitcodes in WAV files of good quality recordings with a lot of dynamic range. You would be surprised how low/small the absolute amplitude/bitcodes tend to be at. I certainly was. Also, you have to realize that even a high level 0dBFS sinewave is composed of bitcodes that run from 0 to 32767 back to 0 down to -32767 back up to 0, et. al. 


Well, I can discretize a sine wave to 64 bits and still see every bit active. That's does not prove any quantisation below -80dB FS is useful perceptively.

As far as your doubts whether 16-bits is necessary or not, I guess through your own powers of intuition that you know better than Sony and Philips who came up with the 16/44 Redbook spec, or Philips who realized the TDA1540 at 14-bits was too sucky and needed to make a more accurate 16-bit chip, or the professional recording industry who didn't like the 13/14 bits of information left over from processing in 16-bits, or Mike Mofatt who pioneered the standalone DAC / use of milspec chips (GAIN) two decades ago.


- 14 bits is over 80 dB of dynamic range
- Electronics 20 years ago could barely touch that in terms of background noise. I used to do a lot of n&v measurements as part of my work, still look / process test data every day.
- 16 bits was thus a reasonable value considering the typical clipping headroom.
- Just for the same reason we use 24 bit DAC in modern ADC because it makes life easier (we guarantee our background noise is due to mics / electronics rather than quantisation error).

Personally, I'll place my in trust Sony, Philips, Mofatt, (and my own ears which tell me 18-bits is about right for me for playback) than your belief that 16-bits is unnecessary and 14 honest bits with 2 fuzzy fits is all that we need. Honestly, you just sound like an apologize for TotalDAC at this point. It is what it is. No need to defend it. I don't disbelieve you when you said the totalDAC sounded great with your 10 minutes or whatever gear and recordings you used during your audition.


- Not preaching for TotalDAC in particular but I guess I can object to your claims of it sounding unresolved since it's in my stash for a month and proving you wrong to my ears.
- I am all about room cues / ambience / keeping the texture of instruments day, placing instruments in the room. And, despite your talk, the D1 stomps every other 20-24 bit D/S dac I have heard in that regard. Something doesn't add up..
- Conversely, all you've heard from the D1 is its spec sheet and extrapolation of what MSB gear sounds like...
- As for Mike Moffat's opinion, I'd be inclined to believe he has no interest in saying one needs any less than 18 bits as he's the one who decided on using a fairly expensive DAC part in the Iggy :wink:.

I don't disbelieve you when you said the totalDAC sounded great with your 10 minutes or whatever gear and recordings you used during your audition.


Now I see the true purrin: discredit people's authority, discredit people's ear, trash competitive gear, trash reviewers, do anything it takes to make your point. I have this image of you racing in the mud with many other people, stepping on people's face to stay out of the mud :p

At last though as we both have a life I suppose: fair enough in the required dynamic effective bit depth, I'd be curious to hear your altered wav files. Especially when I hear the Iggy as it would make sense only then, but even now just to validate it sounds identical on the D1.

Arnaud
 
Mar 8, 2015 at 7:14 PM Post #2,922 of 6,500
  Quote:
 
- Fine, but there are concepts like level masking that long exist and prove this makes little sense.


 
Yes, the question whether these small inaccuracies make any difference when the bitcodes indicate higher levels is valid.
 
What I would ask you to do is examine the bitcodes in WAV files of good quality recordings with a lot of dynamic range. You would be surprised how low/small the absolute amplitude/bitcodes tend to be at. I certainly was. Also, you have to realize that even a high level 0dBFS sinewave is composed of bitcodes that run from 0 to 32767 back to 0 down to -32767 back up to 0, et. al. 
 
As far as your doubts whether 16-bits is necessary or not, I guess through your own powers of intuition that you know better than Sony and Philips who came up with the 16/44 Redbook spec, or Philips who realized the TDA1540 at 14-bits was too sucky and needed to make a more accurate 16-bit chip, or the professional recording industry who didn't like the 13/14 bits of information left over from processing in 16-bits, or Mike Mofatt who pioneered the standalone DAC / use of milspec chips (GAIN) two decades ago.
 
Personally, I'll place my in trust Sony, Philips, Mofatt, (and my own ears which tell me 18-bits is about right for me for playback) than your belief that 16-bits is unnecessary and 14 honest bits with 2 fuzzy fits is all that we need. Honestly, you just sound like an apologize for TotalDAC at this point. It is what it is. No need to defend it. I don't disbelieve you when you said the totalDAC sounded great with your 10 minutes or whatever gear and recordings you used during your audition.
 

 
Not sure I understand how it works with bit depth.  So a high dynamic range 24 bit recording will use how many bits?  Is there any point having a DAC that can resolve more than this number of bits?
 
Seems to be some kind of relationship between DAC noise floor and dynamic range?
 
Mar 8, 2015 at 7:31 PM Post #2,923 of 6,500
I don't have have a problem with the results, but I have a problem with context and framing of the results (picture of expensive McIntosh tube amp to pander to the "overpriced expensive stuff" agenda (the McIntosh amp had nothing to do with the test), running from the built-in headouts, and comparing stuff which mostly really does sound more or less the same. But you have consider a the audience: It's rather trendy thing for people in their twenties who still live with their parents to rail against the "[COLOR=333333]overpriced arena of ******** that is the field of 'high-end" audio'. Part of this an extension of the nwavguy movement, which in turn is a modern take on of the the Doug Self movement decades ago.[/COLOR]

[COLOR=333333]Lest we forget, we only need to go a few posts back in this very thread: [/COLOR]http://www.head-fi.org/t/693798/thoughts-on-a-bunch-of-dacs-and-why-i-hate-chocolate-ice-cream/2850#post_11386526.

You won't see me going to Bimmer or Audi forums calling people out for being tards on buying "overpriced expensive stuff." To each his own.


I am pretty sure noone from tom's will come here to complain about your dacs :)
And I wont either. Got no problem paying 5-10k for a component. But if I pay that kind of money, I wanna be sure it sounds better than the $2 mboard chip. And I want that 'better' to be easily audible. Same as easy as seeing the price diff in my account.
Did tom's have an agenda? Maybe. But everyone has. And their test procedure was more than ok. Many people knew already that the benchmark and the odac were quite similar. Nwavguy said that too and many confirmed. But a $2 soundcard chip!? That's not funny anymore
 
Mar 8, 2015 at 7:48 PM Post #2,924 of 6,500
I am pretty sure noone from tom's will come here to complain about your dacs
smily_headphones1.gif

And I wont either. Got no problem paying 5-10k for a component. But if I pay that kind of money, I wanna be sure it sounds better than the $2 mboard chip. And I want that 'better' to be easily audible. Same as easy as seeing the price diff in my account.
Did tom's have an agenda? Maybe. But everyone has. And their test procedure was more than ok. Most people knew aready that the benchmark and the odac were quite similar. Nwavguy said that too and many confirmed. But a $2 soundcard chip!? That's not funny anymore

Dude, I've listened to a bunch of dacs starting from onboard dacs to msb diamond dac. If you cannot hear the difference between the 2$ soundcard and a 5000$ dac there could be many issues:
 
1. The other components in the system are not up the the task
2. The dac is truly BS
3. You don't hear the difference because:
  1. Hearing problems -> Go see a doctor 
  2. Audio memory problems -> untrained ear most probably 
 
As I said before we rely most on our vision, so the hearing is not our best sensory input, thus most of us don’t have a very good audio memory. This is an interesting test for audio (rhythm) memory. If you score badly and cannot tell the differences between different rhythms, it will be very hard to tell the differences between some audio gear ( as soundstage, details, dynamics, transients differences are harder to distinguish than rhythm changes). However, in time your audio memory gets better and your ears more sensible to these aspects. This is why this hobby is appreciated better in time and gets even more addictive. 
 
You may also want to join the Philips Golden Ears Challenge which has very good training and tests. 
 
Mar 8, 2015 at 8:16 PM Post #2,925 of 6,500
Dude, I've listened to a bunch of dacs starting from onboard dacs to msb diamond dac. If you cannot hear the difference between the 2$ soundcard and a 5000$ dac there could be many issues:

1. The other components in the system are not up the the task
2. The dac is truly BS
3. You don't hear the difference because:
  • Hearing problems -> Go see a doctor 
  • Audio memory problems -> untrained ear most probably 

As I said before we rely most on our vision, so the hearing is not our best sensory input, thus most of us don’t have a very good audio memory. This is an [COLOR=801D1D]interesting test [/COLOR]for audio (rhythm) memory. If you score badly and cannot tell the differences between different rhythms, it will be very hard to tell the differences between some audio gear ( as soundstage, details, dynamics, transients differences are harder to distinguish than rhythm changes). However, in time your audio memory gets better and your ears more sensible to these aspects. This is why this hobby is appreciated better in time and gets even more addictive. 

You may also want to join the [COLOR=801D1D]Philips Golden Ears Challenge[/COLOR] which has very good training and tests. 


Are you sure you are talkin to me? I did no test, I just posted a link. Not my Dacs, not my ears, not my results. And I tried to find a fault in that test too because it sounds kinda funny. No success.

You wanna prove that tom's test was flawed?! Find the flaw. Or do a better test. Dont just parade your 'magic' ears and/or insult other people's ears. Oh and btw, neither I nor my audiologist like to be called 'dude'.

P.S.
You may also find this thread interesting http://www.head-fi.org/t/754695/asrock-z97-e-itx-ac-outperforms-my-bifrost-and-crack-hd-600
 

Users who are viewing this thread

Back
Top