system54,
may I ask about your intention here? You show up with some hypothesis and theories about DAC SQ but you are not willed to show details (several people asked for pictures of your mods you did not send) or explain your points more deeply and enter into a serious discussion. Now you come around with column speakers which is a completely different topic. Whats your point?
may I ask about your intention here? You show up with some hypothesis and theories about DAC SQ but you are not willed to show details (several people asked for pictures of your mods you did not send) or explain your points more deeply and enter into a serious discussion. Now you come around with column speakers which is a completely different topic. Whats your point?
Hi System54
If you have the same results with an headphone , the problems come from the DAC ,if not ......very strange
Mark,
BTW, did you compare your modded DAC to the Holo Audio Cyan DSD DAC which is VERY similar (AK4137 conversion to DSD) to your DAC ?
The Holo is around half the price of the DAC-3.
Matt
Mark, this is the DAC that uses the ESS9311 voltage chip.
http://www.project-audio.com/wp-content/uploads/2018/04/Pre_BoxS2Digital_EN.pdf
Pre Box S2 Digital – Pro-Ject Audio Systems
There is also another 9038pro DAC from somewhere in Shenzen, China that also uses it. Can't remember too clearly though
http://www.project-audio.com/wp-content/uploads/2018/04/Pre_BoxS2Digital_EN.pdf
Pre Box S2 Digital – Pro-Ject Audio Systems
There is also another 9038pro DAC from somewhere in Shenzen, China that also uses it. Can't remember too clearly though
Last edited:
Well, the previous version Katana still here for testing is up and running again with some help from Allo. Finally was able to try it with an external +-15v power supply. It did help to improve sound quality slightly, I would say, but not really enough to shift the most recent 3-way dac comparison rankings. What that says to me is the DC/DC converter on the microcontroller board is actually quite good for that type of thing.
If anyone missed it, the most recent 3-way rankings are those from just after the last AVCC changes to my modded Chinese dac a few days ago. I didn't say so then, but at the same time as the AVCC mod, I tried adding some additional output stage filtering after the differential summing stage which I think helped SQ a little bit, too. Overall, the changes produced much more of an improvement than I was expecting, although in absolute terms it wasn't really a lot. I just wasn't expecting all that much given there probably isn't that much more that can be done, at least aside from something like going to outboard interpolation filtering which would be more of an architectural change rather than a little fine tuning.
Moving along then, the forward looking news is that we are expecting Allo to ship a couple of boards for another take at product review after maybe another week of testing and whatever else their process involves. Also expecting there will also be a revised user manual, along with maybe some supporting material covering use with Volumio (not sure about that part although I have seen some something of that nature, just don't know the plans for it).
The new Kantana boards should be 1) an updated discrete opamp output stage board, and 2) a new for the first time Katana Isolator board. Both boards are expected to provide some sound quality improvements which means of course, a new 3-way dac comparison. As always, I would like to see Katana be the best it can be and it would fine and even good if it turns out to surpass what my modded dac can currently do. The goal is to get better dacs in more people's hands by whatever means, either by modding if that is what somebody wants to do, or by purchasing a very good quality one at a price that is affordable. Please don't forget that as dacs keep getting better and reveal more details than most people have probably ever heard on a record, it takes the rest of a reproduction system to be up to the task. I still strongly recommend modding a cheap $32 Chinese LME49600 headphone amp as a tool for comparing dac and preamp sound quality and for plain enjoyment of music. Even for people that don't like headphones, a really low distortion headphone amp turns out to be a very useful thing to have around. I consider it an essential tool in my technical tool box, as well as enjoyable to listen to for it's excellent sound quality.
If anyone missed it, the most recent 3-way rankings are those from just after the last AVCC changes to my modded Chinese dac a few days ago. I didn't say so then, but at the same time as the AVCC mod, I tried adding some additional output stage filtering after the differential summing stage which I think helped SQ a little bit, too. Overall, the changes produced much more of an improvement than I was expecting, although in absolute terms it wasn't really a lot. I just wasn't expecting all that much given there probably isn't that much more that can be done, at least aside from something like going to outboard interpolation filtering which would be more of an architectural change rather than a little fine tuning.
Moving along then, the forward looking news is that we are expecting Allo to ship a couple of boards for another take at product review after maybe another week of testing and whatever else their process involves. Also expecting there will also be a revised user manual, along with maybe some supporting material covering use with Volumio (not sure about that part although I have seen some something of that nature, just don't know the plans for it).
The new Kantana boards should be 1) an updated discrete opamp output stage board, and 2) a new for the first time Katana Isolator board. Both boards are expected to provide some sound quality improvements which means of course, a new 3-way dac comparison. As always, I would like to see Katana be the best it can be and it would fine and even good if it turns out to surpass what my modded dac can currently do. The goal is to get better dacs in more people's hands by whatever means, either by modding if that is what somebody wants to do, or by purchasing a very good quality one at a price that is affordable. Please don't forget that as dacs keep getting better and reveal more details than most people have probably ever heard on a record, it takes the rest of a reproduction system to be up to the task. I still strongly recommend modding a cheap $32 Chinese LME49600 headphone amp as a tool for comparing dac and preamp sound quality and for plain enjoyment of music. Even for people that don't like headphones, a really low distortion headphone amp turns out to be a very useful thing to have around. I consider it an essential tool in my technical tool box, as well as enjoyable to listen to for it's excellent sound quality.
Last edited:
Mark, how about describing what aspects were lacking to make you make the final rankings. Also what external power supply did you use or the power supply throughout the chain.
At some point, as the system improves the artifacts of the power supply begin to show itself. Especially so on analog. You might be a disbeliever in op amp sounds, but the influence of the power supply should not be put as a minor issue and it will significantly influence the end result. Especially so as the circuit itself gets better.
At some point, as the system improves the artifacts of the power supply begin to show itself. Especially so on analog. You might be a disbeliever in op amp sounds, but the influence of the power supply should not be put as a minor issue and it will significantly influence the end result. Especially so as the circuit itself gets better.
There is not much point in me trying to describe too much SQ details of a product that will not be the released design. Haven't heard that yet. But, generally speaking, we are just talking about minimizing distortion and noise which keeps getting better and better, and which is already down at a level making further improvements more challenging each time.
For Katana power, it needs 3 or 4 power supplies for best SQ. The first 3 are 5v and the if there were a 4th it would be +-15v. In the Katana design a lot of attention has been paid to power supply regulation and filtering, so it seems much more tolerant to PS quality than most designs.
Regarding opamps, sometimes they sound different. I wouldn't argue that they don't or can't in some situations. However, the reason people roll opamps to try to improve sound probably has more do with them being in sockets than anything else. That means if the problem causing less than stellar sound quality is unrelated to opamps, people will still try to compensate for it with the thing that is easiest to swap out and try. It is easy to get stuck in a false optimum solution where what is probably really the wrong opamp is compensating for some other problem that isn't getting fixed.
Probably, it will be best to wait patiently and see what the next Katana iteration brings. When Allo settles on the final design will be the time to talk about what customers can expect. Partly by writing things about my dac and the old Katana I still have I am just keeping the story going of how they may leapfrog each other as the designs continue to be fine tuned. All that will really matter to consumers is what they end up getting if they buy something. All the talk while we are waiting for that is for the technical types who are interested in the ongoing story in the background.
Last edited:
We know that Katana is a design in transition at this time. So judgement there is meaningless to some degree at this point but it would be interesting to hear what you hear are the effects of each change. That way people might learn what they can possibly try when they have the same issue. But what about your modded DAC to this point and the DAC you use as a reference What qualities do you ascribe to each...that way no judgement is called but purely as an opinion of what you hear not what is better overall. This is not a contest but a sound character description.
Unfortunately, I don't find that words are very good at conveying sound quality impressions. I could describe how something very subtle sounds to me and somebody could say, "oh yeah, my $10 usb dac was just like that and I made it perfect with a new opamp!" How do you explain to them, no, that is not what I meant? Especially if they never heard something like a $2,000 dac and only their computer sound card?
It just never seems to work satisfactorily when I have tried it, I guess is the bottom line.
It just never seems to work satisfactorily when I have tried it, I guess is the bottom line.
Last edited:
Yes, I know about $10 DAC sound. Sorry, I think I probably exaggerated what I was trying to get at which I shouldn't have.
Maybe I should have said something like, I can't describe something to someone that has no point of reference for it. Words can't really do it justice. For example, suppose someone had never seen the color yellow. How could you describe it to them in a way so they could really understand what it was like? I don't think you can, really. The only way we learn what yellow is when someone points at a bus and says yellow, then points to a banana and says yellow again. We start to figure out that they must be referring to something that is the same about them. If they keep pointing to things and saying yellow pretty soon we can be sure we know they are referring to the mental experience of seeing yellow.
But, you have to understand yellow light is not really yellow. In other words, yellow is not actually a property of light waves, but a type of shared mental experience humans can have, and can give a name to. If fact, the experience of yellow can be generated by light waves that are not yellow frequency at all. The only colors in a computer monitor are red, green, and blue. There is no yellow, the only yellow is the experience inside our heads.
So it is with sound, too. I can't tell someone what a really good-for-the-price $2,000 dac sounds like. What it sounds like is a mental experience, that's all. If someone else has heard a bunch a dacs while I was there too we could point at them and say some words to describe the sound and pretty soon we could talk about the sound in a meaningful way.
With Mikett, I think he and I share some word meanings about sound quality even though we never met. Some but not all of his descriptions of how his dac sounds have some meaning to me that I am guessing is about what he intends me to think. To the extent people have learned some of the commonly used words people use to describe the difference between hi-end audio equipment and lesser equipment, there can probably be some useful communication. Thing is though, now that there are no longer any hi-fi stores where people can go to listen to the good stuff and learn some of the words from the salesmen or friends with similar interests, there is much less opportunity for many or most people to ever hear any of that level equipment. What they have for reference is closer to cell phones and computer sound card level audio sound quality.
For people that haven't heard the good stuff there is no way for me to explain the subtle differences, maybe like subtle differences in different shades of yellow. But, that doesn't keep people's brains from producing a belief of understanding based on what they have experienced.
This brings me back to the comment about the $10 dac. I probably should have said $500 dac or something. Even at that, I can't really explain very well in words something like what graininess means, or how subtle shades of graininess differ. Whenever I have tried, I always seem to get comments back from people who think they know exactly what I mean, but they also say things that imply they didn't understand what I was trying to say anywhere near as well as they felt that they did. Whether I failed as a communicator or they assumed maybe too much, it doesn't matter. It just didn't work at conveying what I wanted to try to convey.
There is more to it than that, but hopefully I have said enough to help clarify why I don't like trying to describe what something sounds like with much attempt at exact detail. I am happy to give a recommendation, or maybe even to rate something as an approximate percentage sound quality difference relative to something else, but don't want to get into trying to communicate a mental experience in detail.
Maybe I should have said something like, I can't describe something to someone that has no point of reference for it. Words can't really do it justice. For example, suppose someone had never seen the color yellow. How could you describe it to them in a way so they could really understand what it was like? I don't think you can, really. The only way we learn what yellow is when someone points at a bus and says yellow, then points to a banana and says yellow again. We start to figure out that they must be referring to something that is the same about them. If they keep pointing to things and saying yellow pretty soon we can be sure we know they are referring to the mental experience of seeing yellow.
But, you have to understand yellow light is not really yellow. In other words, yellow is not actually a property of light waves, but a type of shared mental experience humans can have, and can give a name to. If fact, the experience of yellow can be generated by light waves that are not yellow frequency at all. The only colors in a computer monitor are red, green, and blue. There is no yellow, the only yellow is the experience inside our heads.
So it is with sound, too. I can't tell someone what a really good-for-the-price $2,000 dac sounds like. What it sounds like is a mental experience, that's all. If someone else has heard a bunch a dacs while I was there too we could point at them and say some words to describe the sound and pretty soon we could talk about the sound in a meaningful way.
With Mikett, I think he and I share some word meanings about sound quality even though we never met. Some but not all of his descriptions of how his dac sounds have some meaning to me that I am guessing is about what he intends me to think. To the extent people have learned some of the commonly used words people use to describe the difference between hi-end audio equipment and lesser equipment, there can probably be some useful communication. Thing is though, now that there are no longer any hi-fi stores where people can go to listen to the good stuff and learn some of the words from the salesmen or friends with similar interests, there is much less opportunity for many or most people to ever hear any of that level equipment. What they have for reference is closer to cell phones and computer sound card level audio sound quality.
For people that haven't heard the good stuff there is no way for me to explain the subtle differences, maybe like subtle differences in different shades of yellow. But, that doesn't keep people's brains from producing a belief of understanding based on what they have experienced.
This brings me back to the comment about the $10 dac. I probably should have said $500 dac or something. Even at that, I can't really explain very well in words something like what graininess means, or how subtle shades of graininess differ. Whenever I have tried, I always seem to get comments back from people who think they know exactly what I mean, but they also say things that imply they didn't understand what I was trying to say anywhere near as well as they felt that they did. Whether I failed as a communicator or they assumed maybe too much, it doesn't matter. It just didn't work at conveying what I wanted to try to convey.
There is more to it than that, but hopefully I have said enough to help clarify why I don't like trying to describe what something sounds like with much attempt at exact detail. I am happy to give a recommendation, or maybe even to rate something as an approximate percentage sound quality difference relative to something else, but don't want to get into trying to communicate a mental experience in detail.
Last edited:
Okay. There are very detailed measurements of DAC-3 available at Stereophile and in the owner's manual which can be downloaded at the Benchmark Media website.
Regarding my dac and Katana, the premise of the dac mod project by me in this thread has been that it was to see how good a DIY dac could turn out if done by DIYer's without access to expensive specialized test equipment (which there is none of here).
I have tried before, and I may try again, to take some very basic measurements here with a sound card. But there is a problem that I described in detail before about being located on the top of a hill not too far away from three airports. My location is awash in RF sources of all types which tends to cause spurs in FFTs taken with a sound card. I have put my dac boards inside a steel file server computer case which might help some, but it is a difficult problem I have already worked on a fair amount. Bottom line is I may have to go somewhere to get even harmonic distortion @1kHz (which is hardly enough to characterize a dac). More information about the local RF environment about half way down the following post: http://www.diyaudio.com/forums/digital-line-level/314935-es9038q2m-board-129.html#post5445103
However, DAC-3 has been measured by multiple people and it is the reference for sound quality here where I am. I can compare the sound of it to the other dacs which I do. So far it has worked as a sound quality reference to help make my modded dac currently sound more like DAC-3 than the previous version of Katana did. We will have to see about the next version of Katana.
Once Allo determines the final version of Katana is done, I think there are plans to for measurements to be posted over at audiosciencenews.com . Perhaps some other review sites will also post measurements, which is nice because it means I can point to people to where to find that information. In the meantime, I am not aware of any detailed measurements of Katana having been released by Allo. It may be we will have to wait longer to find out that type of information.
Regarding my dac and Katana, the premise of the dac mod project by me in this thread has been that it was to see how good a DIY dac could turn out if done by DIYer's without access to expensive specialized test equipment (which there is none of here).
I have tried before, and I may try again, to take some very basic measurements here with a sound card. But there is a problem that I described in detail before about being located on the top of a hill not too far away from three airports. My location is awash in RF sources of all types which tends to cause spurs in FFTs taken with a sound card. I have put my dac boards inside a steel file server computer case which might help some, but it is a difficult problem I have already worked on a fair amount. Bottom line is I may have to go somewhere to get even harmonic distortion @1kHz (which is hardly enough to characterize a dac). More information about the local RF environment about half way down the following post: http://www.diyaudio.com/forums/digital-line-level/314935-es9038q2m-board-129.html#post5445103
However, DAC-3 has been measured by multiple people and it is the reference for sound quality here where I am. I can compare the sound of it to the other dacs which I do. So far it has worked as a sound quality reference to help make my modded dac currently sound more like DAC-3 than the previous version of Katana did. We will have to see about the next version of Katana.
Once Allo determines the final version of Katana is done, I think there are plans to for measurements to be posted over at audiosciencenews.com . Perhaps some other review sites will also post measurements, which is nice because it means I can point to people to where to find that information. In the meantime, I am not aware of any detailed measurements of Katana having been released by Allo. It may be we will have to wait longer to find out that type of information.
Last edited:
Markw4, thank you. What I really want to see is, how those cheap Ebay boards can be better after modification, and what modification solves what problem. As I said before, I only have had very unlucky experience with those Ebay ESS boards, and measurements were strange looking, so the measurements of this particular board before/after modification is very interesting.
plasnu, I think it is fairly easy to see how the boards can be better. ESS has a downloads page on their website with some advice on getting the best out of their dac chips. I started by following the advice very carefully and literally. It says to use certain parts, put output stages and AVCC power very close to the dac chip, etc. That's why I made such tiny SMD boards for AVCC power and I/V stages, and put them on the other side of the ground plane. I cleaned up the power quality on the board, and since there are no power planes, I tried to keep power supply impedances to opamps low. I put in a very low jitter clock, too. I then experimented with external SRC including SRC4392 and AK4137, including improved grounding with extra ground pins for I2C bus interconnections. That was mostly after reading what I could find at the benchmark website about some of the techniques they use. The story is really that simple when you come right down to it. Other people don't take as many little details into consideration when they do mods, or they think some things that would inconvenient or hard to do can be ignored and probably won't matter, whereas I think a lot about my layouts before I start soldering and figure out how to do hard, small assembly if necessary, for when it matters to do it that way. I also made a summary sheet of recommendations made by ESS because they are spread out and can be hard to keep in mind all at once, then I checked my layouts and parts choices to verify I wasn't missing details. I will attach the sheet below so you can see.
Another thing I did was finally ask ESS for a data sheet. I figured out a way to tap into I2C bus and configure the dac registers. ESS doesn't say much about some things, but to be thorough I tried adjusting DPLL bandwidth and found it helped sound quality. I found that upsampling to DSD rather than PCM sounded better by trying it, etc.
So, it comes down to meticulously following manufacturer recommendations, following well-known good engineering practices, and basically leaving no stone unturned in trying to address every variable I could that might affect sound quality, excepting external interpolation filtering. Maybe someday later on that.
Another thing I did was finally ask ESS for a data sheet. I figured out a way to tap into I2C bus and configure the dac registers. ESS doesn't say much about some things, but to be thorough I tried adjusting DPLL bandwidth and found it helped sound quality. I found that upsampling to DSD rather than PCM sounded better by trying it, etc.
So, it comes down to meticulously following manufacturer recommendations, following well-known good engineering practices, and basically leaving no stone unturned in trying to address every variable I could that might affect sound quality, excepting external interpolation filtering. Maybe someday later on that.
Last edited: