aussievintage Posted March 10 Posted March 10 I am sure reactions to this video will be polarised, as it is an audio engineering take on the subject. Hope you can watch it to the end though, especially as there is a blind test to try there, and it contains lots of good info, even if you disagree. Also the demonstration of the effect of sight on hearing is good. 5
frankn Posted March 11 Posted March 11 Why then do mixing and mastering engineers use DAC’s costing multiple thousands of dollars then? I understand his points but he starts out wrong by showing a supposedly digitised waveform as stair steps - it’s not how it works and he should know this , especially as he was at pain to point out his knowledge of the Shannon-Nyquist theory (improperly referred to as Nyquist theory) - I’m nitpicking I know. Not a bad video ……. There are better ones about A-D sampling, by an American engineer but I can’t find it at the moment. All commercial products end up being priced at about x10 the sum of the parts, labour, development, advertising, legal and administrative costs, be it DAC’s , amplifiers, speakers, servers, etc. That is why there is such a large DIY community out there. If you are a believer then purchase a Topping DAC (or all their products ) and be done with it I suppose. 3
aussievintage Posted March 11 Author Posted March 11 9 hours ago, frankn said: supposedly digitised waveform as stair steps - Which, while it is on the screen, he says you do not hear due because "low pass filters exist". I think he presented it poorly. No comments on the blind test, or the example of sight influencing what is heard?
frankn Posted March 12 Posted March 12 2 hours ago, aussievintage said: Which, while it is on the screen, he says you do not hear due because "low pass filters exist". I think he presented it poorly. No comments on the blind test, or the example of sight influencing what is heard? The staircase representation is false. Sampling is not change and hold until the next sample. Either he believes this or he is too lazy to get the correct graphics in his presentation. You don’t hear it because it doesn’t exist. The interpolation filter joins the “dots”, then, finally a LP filter separates the high frequency noise from the “lower” frequency signal. Nitpicking , I know. I agree that sighted bias is real so no need to comment.
georgehifi Posted March 12 Posted March 12 The only way to musically test 2 x dacs is to play a 1khz sine wave of a test disc through each and get an exact level match. Then A/B with music, this will also sort out any dynamic differences between the 2. Cheers George 1
aussievintage Posted March 12 Author Posted March 12 1 hour ago, georgehifi said: The only way to musically test 2 x dacs is to play a 1khz sine wave of a test disc through each and get an exact level match. Then A/B with music, this will also sort out any dynamic differences between the 2. Cheers George Which is what he did, then asked how many times you hear him switch (A/B) between the two.
Keith_W Posted March 12 Posted March 12 13 hours ago, frankn said: Why then do mixing and mastering engineers use DAC’s costing multiple thousands of dollars then? Because those DAC's have features that your average home audio DAC does not have. For example, 16 channels of DA conversion, 64 MADI input/outputs, 16 channels of AD conversion, network audio support (such as AES67, Dante, Ravenna). Ability to gang dozens of DAC's together to make a super-DAC with thousands of DA channels (you might want this in an airport for example). You can slave all the DAC's to one master DAC and receive clocking signal through the network. You can route any channel into another channel. Some of those DAC's have on-board DSP. Oh, and if you buy from a reputable company - you get lifetime software support. My RME and Merging DAC's are still getting firmware and driver updates even though both DAC's are close to 10 years old. I'll tell you what those pro audio DAC's don't have - no linear power supplies, no fancy machined enclosure, no custom DAC chips (e.g. Ring DAC or FPGA or ladder). They all use off-the-shelf DAC chips from ESS or AKM. 5 1
aussievintage Posted March 12 Author Posted March 12 1 hour ago, frankn said: Sampling is not change and hold until the next sample. Actually, going back to basics, yes it is. Now I am nitpicking, but that in it's purest form, that's exactly what sampling in a scientific sense, is. But modern DACs are not that basic/simple. Back in the 70s (maybe earlier) we (DIYers) were building DACs that did just this, then relied on filtering to clean it up, but it was indeed a staircase before the filter.
frankn Posted March 12 Posted March 12 20 minutes ago, aussievintage said: Actually, going back to basics, yes it is. Now I am nitpicking, but that in it's purest form, that's exactly what sampling in a scientific sense, is. But modern DACs are not that basic/simple. Back in the 70s (maybe earlier) we (DIYers) were building DACs that did just this, then relied on filtering to clean it up, but it was indeed a staircase before the filter. The only value the digital circuit knows is the value for the sample at time-x, x1, x2 etc. The interpolation comes 1st, then filtering.
aussievintage Posted March 12 Author Posted March 12 13 minutes ago, frankn said: The only value the digital circuit knows is the value for the sample at time-x, x1, x2 etc. The interpolation comes 1st, then filtering. We are saying the same thing then. That first bit "The only value the digital circuit knows is the value for the sample at time-x, x1, x2 etc" is what forms the staircase
frankn Posted March 12 Posted March 12 It’s a wrong analogy, it is discrete sampling. If you step off the sample you fall to the floor. if it was a staircase you could move to new sample point in between and have the same value. Then upsampling and other things wouldn’t work.
aussievintage Posted March 12 Author Posted March 12 (edited) 6 minutes ago, frankn said: It’s a wrong analogy, it is discrete sampling. If you step off the sample you fall to the floor. if it was a staircase you could move to new sample point in between and have the same value. Then upsampling and other things wouldn’t work. You aren't sampling a staircase (although you could) so you won't fall off. The staircase just represents the value of the samples and the time periods they nominally represent. If you put digital numbers representing the sample values, one after the other, into an R2R network (for example) , a staircase comes out. I don't understand why that worries you. Anyway, too far down a rabbit hole, enough for me Edited March 12 by aussievintage
Volunteer Volunteer Posted March 12 Volunteer Posted March 12 Provocative posts and the responses have been hidden. They add nothing. Consider this a warning. Next steps do not need to be spelled out, so behave. 1
andyr Posted March 12 Posted March 12 Given you mentioned "blind test", av ... I have a question for you. AIUI, blind tests are supposed to remove bias - IOW, if I see that a $10k DAC is in play ... I will probably think it sounds better than my $200 Topping E30. But what if the listeners can't really hear any difference between the sounds of two DACs. They know which one is playing ... but as they can't pick any meaningful difference - why would this matter?
aussievintage Posted March 12 Author Posted March 12 50 minutes ago, andyr said: Given you mentioned "blind test", av ... I have a question for you. AIUI, blind tests are supposed to remove bias - IOW, if I see that a $10k DAC is in play ... I will probably think it sounds better than my $200 Topping E30. But what if the listeners can't really hear any difference between the sounds of two DACs. They know which one is playing ... but as they can't pick any meaningful difference - why would this matter? Not sure I understand the question. The test offered in the video seemed a simple way to remove bias - i.e. by asking how many times he switched between DACs, which of course are unseen. I might be wrong, but seemed a simple and elegant way to go about it. If they truly sound the same, then you'll not be able to hear any change and will not be able to tell how many times he did it.
brodricj Posted March 13 Posted March 13 On 11/03/2025 at 6:58 AM, aussievintage said: Also the demonstration of the effect of sight on hearing is good. The sense of hearing is always better tuned in the absence of the sense of sight. That is uncontroversial. Blind A/B comparison testing of anything in audio is a pointless exercise. 2
aussievintage Posted March 13 Author Posted March 13 Is that all anyone is getting from the video? An opportunity to discuss blind testing? 1
frankn Posted March 13 Posted March 13 47 minutes ago, aussievintage said: Is that all anyone is getting from the video? An opportunity to discuss blind testing? What did you expect? This is just another take on an often discussed topic, there will never be a consensus no matter what is brought to the internet. If the event/demonstration had had SNA members in attendance there might be more discussion. 2
rocky500 Posted March 13 Posted March 13 Just did read this "On the surface, it seems like blind A/B testing is an indisputable method for establishing the superiority of one audio item over another. But there’s a flaw in that method. I argue that the increased enjoyment one experiences from one audio element over that from another cannot be verified with an analytical comparison activity. Music is very much a “right brain”, emotional activity. So much so that it gives your mind a break from a spate of “left brain”, analytical activity. When the mind is exercised in an A/B comparison, the left brain is dominant because it’s an analytical activity, with the right brain subdued. Problem is, music is appreciated more in a right brain sense. Hence the left brain results of an A/B test say little about the quality of the experience by the right brain when just listening. Only on reflection by the left brain of the right brain experience can one say whether it was good or not. Those who worship at the blind A/B altar attribute the inconsistencies of the results between what people say they prefer and what is chosen as better in A/B comparisons to things like expectation bias or to there being no differences between the two choices. Maybe people who are dominated by their left brain are those who cannot hear some things that they write off, while others truly prefer these things?" Could be why when ever I have done blind tests in the past, I always can never pick a difference. When I just listen casually for a few days like I normally do and then swap over and repeat for some time. Then I seem to appreciate the differebt components and how they work for me. Not saying the above quote is right, but it seems to fit with how I have gone in this hooby over the many many years. 5 1
pete_mac Posted March 13 Posted March 13 6 hours ago, rocky500 said: Just did read this "On the surface, it seems like blind A/B testing is an indisputable method for establishing the superiority of one audio item over another. But there’s a flaw in that method. I argue that the increased enjoyment one experiences from one audio element over that from another cannot be verified with an analytical comparison activity. Music is very much a “right brain”, emotional activity. So much so that it gives your mind a break from a spate of “left brain”, analytical activity. When the mind is exercised in an A/B comparison, the left brain is dominant because it’s an analytical activity, with the right brain subdued. Problem is, music is appreciated more in a right brain sense. Hence the left brain results of an A/B test say little about the quality of the experience by the right brain when just listening. Only on reflection by the left brain of the right brain experience can one say whether it was good or not. Those who worship at the blind A/B altar attribute the inconsistencies of the results between what people say they prefer and what is chosen as better in A/B comparisons to things like expectation bias or to there being no differences between the two choices. Maybe people who are dominated by their left brain are those who cannot hear some things that they write off, while others truly prefer these things?" Could be why when ever I have done blind tests in the past, I always can never pick a difference. When I just listen casually for a few days like I normally do and then swap over and repeat for some time. Then I seem to appreciate the differebt components and how they work for me. Not saying the above quote is right, but it seems to fit with how I have gone in this hooby over the many many years. Yes, I'm in the same boat. 4
aussievintage Posted March 13 Author Posted March 13 7 hours ago, rocky500 said: I argue that the increased enjoyment one experiences from one audio element over that from another cannot be verified with an analytical comparison activity. Not sure "increased enjoyment" is what comparison tests are striving to analyse. They tend to be asking simply which sounds better. Increased enjoyment brings in many other factors.
POV Posted March 14 Posted March 14 10 hours ago, aussievintage said: Not sure "increased enjoyment" is what comparison tests are striving to analyse. They tend to be asking simply which sounds better. Increased enjoyment brings in many other factors. Really it's best to start with can i reliably pick a difference between A and B; or A, B, and X. If you can't reliably pick a difference then it's pointless seeking a 'better'.
aussievintage Posted March 14 Author Posted March 14 9 minutes ago, POV said: Really it's best to start with can i reliably pick a difference between A and B; or A, B, and X. If you can't reliably pick a difference then it's pointless seeking a 'better'. Could not agree more, and that's why I thought the video's method was a nice easy way to start. Just prove you can hear a difference by answering the question, how many times was it switched.
aussievintage Posted March 14 Author Posted March 14 WAV OF BLIND TEST: https://drive.google.com/file/d/1JY8x...
Recommended Posts