Hi All,
I have been doing a lot of web research on high-end DAC's. Just ordered a Lavry DA11 to see some things for myself. But of the top 3 that I have found -- Mytek, Lavry, and Benchmark, I hear consistent comments about how much they open up the soundstage, compared to lesser systems like my MOTU 828 and Ultralights.
I do research into full-spectrum musical hearing correction for musicians and audiophiles. Many of us have lost portions of our hearing due to a variety of factors -- industrial noise, illness, misspent youth...
I wonder if anyone has tried an audio shootout among these high end converters using explicitly restricted bandwidth, say a steep lowpass filter at 10 kHz. For those with hearing impairment, there isn't much to be done about losses above 10 kHz. So apart from my anticpated comparison between a Lavry DA11 and the MOTU 828mk3, I would be interested to hear the impressions from avid, careful, listeners under conditions of restricted bandwidth playback. Can any difference be detected? or do you really need the "air" above 10 kHz to realize the differences so often noted for low-jitter accurate playback systems?
- DM
What is needed to hear the difference is a good imagination.
What is needed to hear the difference is a good imagination.
So you believe that the MOTU converters are already quite good?
You're going to have a hard time finding objective evidence showing that much of anything including units priced at $100 is audibly less than quite good.
Interesting... When I feed some worst-case waveforms through a MOTU UltraLight, round-trip from D/A to A/D, and record, I see the artifacts in the results. But they appear to be very slight, about -115 dBFS, and the noise floor seems pretty uniform at around -140 dBFS. So offhand, I would have to agree on an objective basis.
Since I have never listened to anything any better than the MOTU units, I'm looking forward to my impressions of the Lavry DA11. And I will attempt the same kinds of measurements with it to see if there is any objective difference.
- DM
Before going any farther in your pursuit of presenting objective evidence to this forum you need to realize that you agreed to our terms (http://www.hydrogenaudio.org/forums/index.php?showtopic=3974), specifically rule number eight (http://www.hydrogenaudio.org/forums/index.php?showtopic=3974#entry149481).
Here's a small part of the rule that cuts to the chase:
Graphs, non-blind listening tests, waveform difference comparisons, and so on, are not acceptable means of providing support.
I do agree, although "Graphs" is puzzling to me... but I'm not trying to present anything. I'm asking for contributions from other knowledgeable people.
Puzzling or not, the point is that we want to ensure that you understand that an objective basis of the assessment of the quality of sound is ultimately determined by the ears rather than the eyes, as far as this forum is concerned; though many here are ready willing and able to argue that this idea extends well beyond the confines of this forum.
You had mentioned performing measurements, implying that they may include viewing the noise floor from a loop-back test. This was the reason for my interjection.
OP: the default hypothesis here is that all DACs that aren't defective by design, sound the same and are all transparent, even if they measure more or less differently, beyond audibility.
Also, graphs (or pretty much any kind of measurement) aren't accepted here as an indicator of sound quality, because while they may show physical (real) differences, they don't say anything about the audibility of measured attributes.
To add, we do grant a certain degree of leniency when it comes to hardware (the rule itself was initially intended for the development of perceptual coding), though this leniency generally extends to situations that are widely held as uncontroversial (eg. the response of speakers and headphones).
Interesting... When I feed some worst-case waveforms through a MOTU UltraLight, round-trip from D/A to A/D, and record, I see the artifacts in the results. But they appear to be very slight, about -115 dBFS, and the noise floor seems pretty uniform at around -140 dBFS. So offhand, I would have to agree on an objective basis.
FWIW, make sure you test with whatever you usually have hooked up when using the device. So if you use headphones or speakers, get a splitter and leave them plugged in while you test. Often the results with just a loop back are much better then real world when equipment is hooked up.
If e.g 10 kHz is the upper threshold of once hearing you won’t hear anything above 10 kHz.
But if the IMD of signals above 10 k maps into below 10 k, band limiting the signal might have an audible effect.
A steep lowpass filter at 10 kHz might yield audible pre-ringing too.
Personally I wonder if the difference between low priced and higher priced DACs got anything to do with the frequency response.
Maybe you should look at the linearity of DACs.
The cheap ones are linear up to 12 bits, the more expensive ones up 18 bits and the $4000,- up to 22 bits. I wouldn’t be surprised if “opening up”, “air”, etc. etc. is about the linearity of a DAC
Interesting... When I feed some worst-case waveforms through a MOTU UltraLight, round-trip from D/A to A/D, and record, I see the artifacts in the results. But they appear to be very slight, about -115 dBFS, and the noise floor seems pretty uniform at around -140 dBFS. So offhand, I would have to agree on an objective basis.
FWIW, make sure you test with whatever you usually have hooked up when using the device. So if you use headphones or speakers, get a splitter and leave them plugged in while you test. Often the results with just a loop back are much better then real world when equipment is hooked up.
My statements were referring purely to a measured loopback test -- no listening. But my opinion isn't the point here. What I wanted to know is whether or not you need to be able to hear frequencies above 10 kHz for the "improvements" of a high-end low-jitter DAC to become apparent.
One answer so far, talks about IMD, which is not restricted to highest frequencies. Great answer. Does anyone have any kind of acceptable evidence that you can hear a difference if you restrict the audio to frequencies below 10 kHz?
I wouldn’t be surprised if “opening up”, “air”, etc. etc. is about the linearity of a DAC
I would be completely surprised if it has anything to do with something other than the expectation bias of those offering these descriptions.
To your knowledge is there even a single double-blind listening test that was verified by a third-party indicating this to be the case?
The cheap ones are linear up to 12 bits, the more expensive ones up 18 bits and the $4000,- up to 22 bits.
The iPod classic is good to about 16 bits in the tests I've seen, which is all you need. Certainly not bad for a DAC that costs a few dollars.
BTW, 22 bit linearity? What DAC are you looking at that has that many effective bits? And why would you care about linearity so far below the noise floor?
My statements were referring purely to a measured loopback test -- no listening. But my opinion isn't the point here.
What I mean is that your results are likely to be over-optimistic. There is very little difference in the output of DACs in loopback configuration. If you test in a more realistic situation, you will likely find higher distortion. This is particularly so with headphones.
What I wanted to know is whether or not you need to be able to hear frequencies above 10 kHz for the "improvements" of a high-end low-jitter DAC to become apparent.
Most modern DACs are low jitter these days, so I don't think you'll find much difference, regardless of how high you can hear. You have to go quite low end before clock jitter becomes a significant issue.
Hi All,
I have been doing a lot of web research on high-end DAC's. Just ordered a Lavry DA11 to see some things for myself. But of the top 3 that I have found -- Mytek, Lavry, and Benchmark, I hear consistent comments about how much they open up the soundstage, compared to lesser systems like my MOTU 828 and Ultralights.
I do research into full-spectrum musical hearing correction for musicians and audiophiles. Many of us have lost portions of our hearing due to a variety of factors -- industrial noise, illness, misspent youth...
I wonder if anyone has tried an audio shootout among these high end converters using explicitly restricted bandwidth, say a steep lowpass filter at 10 kHz. For those with hearing impairment, there isn't much to be done about losses above 10 kHz. So apart from my anticpated comparison between a Lavry DA11 and the MOTU 828mk3, I would be interested to hear the impressions from avid, careful, listeners under conditions of restricted bandwidth playback. Can any difference be detected? or do you really need the "air" above 10 kHz to realize the differences so often noted for low-jitter accurate playback systems?
- DM
I applaud the sentiment. There are no well-controlled blind listening comparisons of high end DACs vs mass market, that I know of. From what I've read of Dan Lavry's writings, I think he too would expect people to find the differences difficult if not impossible to detect by ear in a fair comparison.
Interesting... When I feed some worst-case waveforms through a MOTU UltraLight, round-trip from D/A to A/D, and record, I see the artifacts in the results. But they appear to be very slight, about -115 dBFS, and the noise floor seems pretty uniform at around -140 dBFS. So offhand, I would have to agree on an objective basis.
Since I have never listened to anything any better than the MOTU units, I'm looking forward to my impressions of the Lavry DA11. And I will attempt the same kinds of measurements with it to see if there is any objective difference.
There's no doubt that artifacts that are 100 dB down or better can't be reliably detected by nobody, nowhere, no how.
In fact that's almost as true for stuff that is 80 dB down. On a really bad day you might just maybe hear something a little, but probably not even that.
Artifacts have to be -60 dB down or worse to be interesting to listen to.
And BTW I promise you that you can spend the rest of your life measuring objective differences between DACs. Did that for about a decade. They are there. Reliably detecting them in a good listening test is just a little bit harder! ;-)
I applaud the sentiment. There are no well-controlled blind listening comparisons of high end DACs vs mass market, that I know of.
I would say that DBT comparisons of mass market DACs versus absolute and total perfection would be even better than comparisons to mere high end DACs.
The ingredients for those tests are any of the several software ABX comparators (e.g. Foobar2000's ABX plug-in) and the files labelled "Soundblaster generations" from here:
http://www.ethanwiner.com/aes/ (http://www.ethanwiner.com/aes/)
Cutting to the chase, the results are that even multiple generations of even mid-priced consumer-grade ADC/DAC pairs are sonically transparent.
DACs and ADCs are among the most perfected of all kinds of audio components.
I also would be interesting in this proposal.
I have a MOTU Ultralite already. When I first bought it I put it to the question by looping it 4x (all I had cables for at the time). Couldn't tell the difference. In fact it was so difficult I gave up trying as a waste of time very quickly. I also recorded vinyl rips and compared the rips to the original. Couldn't tell the difference there either.
Not saying it's impossible for everyone but is is with my ears.
That said, if I had my time again I would have saved up a bit longer and bought something from RME. Better design, better support, better software and particularly better ongoing software support are worth paying for in the long run even if there are truly no sonic differences. Dunno if Lavry, Benchmark et al meet those same criteria.