I do really appreciate ALL the input - Thank you, this is actually very helpful - Johnny and Mark , Abraxalito, Everyone - thanks for great suggestions AND for trying to keep things on track-! I've been out reading checking the suggestions out. The Buffalo choice wasn't because of measurements over anything else, unfortunately I didn't have a chance to listen to one , but at the time it seemed that there were a lot of different company's using it, and quite a few people on this forum were taking part in TPA's design offering.... so it seemed like a good one to try - also it was almost turn key, just a few power supplies and case to add - easy for a newbie to try as well - 🙂
I will go to some kind of a computer- server-streamer-player setup down the road when I can -- just as soon as I afford it.... Keep the ideas coming!
I will go to some kind of a computer- server-streamer-player setup down the road when I can -- just as soon as I afford it.... Keep the ideas coming!
Here's the best idea for you if you want to improve the audible sound quality and you already have a DAC that's not broken, reroute the money you were going to spend on DAC to speakers upgrade and room acoustics. Those two are where the real audible gain per (your precious) dollar can be made, not DACs these days. If you aren't sure what I mean, look up the results of level matched double blind listening tests posted online.
This thread is about DACs and not about improving any other components. Please respect the OP and others like me who are interested in this specific topic. Thank you to those (as mentioned by the OP) who have already made posetive contributions so far.
You act like a moderator but it doesn't say "diyAudio Moderator" under your forum name. 
Oh, wait, you don't believe that it's a waste of money, right?
Posting money wasting options is positive contribution in your view?
Oh, wait, you don't believe that it's a waste of money, right?Fine. Everyone has seen it now.
That being the case, you guys can stop going OT and trolling.
You don't care about the financial situation the OP is in. All you care about is the audio business that you are in affiliation with. 🙄 No wonder your posts resemble so much of those who are in audio business.
Might look at the SMSL M300, see what the reviews say. Would not likely be the weak link in many/most systems I would think. Have seen them for $200-$250, might be better to wait until after the holiday, when it would also be a better time to sell other gear...
USB powered is always has its limitations on sound quality, IME. Also, AK4497 likely uses an internal CMOS opamp for internal I/V conversion. Distortion was much higher in AKM dacs before AK4499. At least they didn't have that characteristic Sabre sound that some people find hard to listen to for very long.
In the early version Allo Revolution I listened to, it sounded much better in that regard that other Sabre dacs I have heard. They are getting a lot better at optimizing the design. Probably would have happened a lot sooner if ESS was't so secretive about information sharing. Still though, Allo's earlier dac, Katana, sounded even better when all aspects of sound quality are considered, but only if all linear power supplies were used include for +-15v. Unfortunately, the trade off Allo has taken is to make Revolution USB input, but not able use external power supplies. The limitation of Katana is that it is RPi only.
There are no optimal solutions in low-ish cost dac world yet, IMHO.
In the early version Allo Revolution I listened to, it sounded much better in that regard that other Sabre dacs I have heard. They are getting a lot better at optimizing the design. Probably would have happened a lot sooner if ESS was't so secretive about information sharing. Still though, Allo's earlier dac, Katana, sounded even better when all aspects of sound quality are considered, but only if all linear power supplies were used include for +-15v. Unfortunately, the trade off Allo has taken is to make Revolution USB input, but not able use external power supplies. The limitation of Katana is that it is RPi only.
There are no optimal solutions in low-ish cost dac world yet, IMHO.
Last edited:
With no verification of your experience, it's only an anecdote. IOW, it doesn't hold water.
In a subjective auditioning? I believe that.
You are entitled to your own humble opinion. What's important is what that humble opinion is based on. As for the audible traits, still no supporting evidence for the claim, just hearsay, opinion, anecdote, audiophile myth... etc.
Mark..
I really do not want to join the mud slinging guys here.. I do not even read them any more.
But sometimes you are shooting from the hip.. like here, for example:
It happens to be that AKM 4490 had been chosen for several test instruments, as in : not an ESS dac... (RTX6001; RME ADI-2 Pro etc..)
I have measured just recently my favorite tweak subject, the Mirand dac.
THD 0.000062% at -10dB FS I would not call "much high" in no circumstances...
https://www.diyaudio.com/forums/ven...4490-usb-dac-dsd-support-191.html#post5790821
Ciao, George
I really do not want to join the mud slinging guys here.. I do not even read them any more.
But sometimes you are shooting from the hip.. like here, for example:
It happens to be that AKM 4490 had been chosen for several test instruments, as in : not an ESS dac... (RTX6001; RME ADI-2 Pro etc..)
I have measured just recently my favorite tweak subject, the Mirand dac.
THD 0.000062% at -10dB FS I would not call "much high" in no circumstances...
https://www.diyaudio.com/forums/ven...4490-usb-dac-dsd-support-191.html#post5790821
Ciao, George
Attachments
Last edited:
Joseph,
Thank you for taking the time to write. I appreciate serious commentary by good people. Especially since my memory isn't always perfect when it comes to things I looked at awhile ago.
Let's take a look at the numbers readily available if that would be okay...
0.000062% expressed in dB is -124dB, and you say it was measured at a peak level of -10dB (down about 32% from full scale).
The equivalent dac chip to AK4497 from ESS is ES9038PRO, which is specified by ESS at -122THD+N
AK4499 is specified by AKM at -117THD+N
The mobile ESS part, ES9038Q2M is specified at THD+N of -120dB
If one really wanted to use one of the ESS parts for measurement, I have seen people get the mobile ES9038Q2M down to -127dB distortion at 0dBFS. The PRO part I haven't seen numbers for but one might get it down to -129dBFS, both measured at full scale.
Regarding the -10 level you specify, IME all the dacs seem to have less audible distortion down at -8dB to -10dB. Doubtful it all has to do with intersample overs. Don't have numbers for distortion optimized ESS part at -10db, and have not had much interest in such measurements for reasons which I will attempt to explain below.
The problem with designing a dac to have lower distortion than the manufacturer states on a data sheet is that the dac typically starts to sound worse for playing music, again this in IME. Therefore, I tend to go by manufacturers stated THD numbers when doing comparisons.
By that metric ES9038PRO is 5dB lower in THD+N than AK4497. It may not sound like much when we are talking about numbers well down below -100dB, but I find it matters. I hear the distortion myself. (Also can hear the very different kind of distortion when dacs a designed to give extra low distortion measurement numbers.)
Another thing that concerned me about AK4497 distortion was the table of more detailed distortion numbers on page 11 of the data sheet. https://www.akm.com/content/dam/doc.../audio-dac/ak4497eq/ak4497eq-en-datasheet.pdf
Of course, ESS doesn't publish the corresponding numbers so no fair way to use those for comparison, but AK4497 numbers are concerning to me since they look to get significantly worse at higher sample rates. On the other hand, the only data point we have on ESS are published measurements of Benchmark DAC-3, which runs ES9028PRO at 211kHz for all playback modes due to upsampling. They are able to hit pretty good numbers even up at 211kHz.
https://cdn.shopify.com/s/files/1/0321/7609/files/DAC3_Series_Manual_Rev_B.pdf
Page 38 shows THD at -126dB, and THD+N at -115dB. Somehow noise ends up kind of high, but THD is very low even at the high 211kHz sample rate.
On the other question you mentioned, as far as why manufacturers of low cost test equipment choose to use AK4497 there could be many reasons. It's harder to get the best out ES9038PRO IME and therefore more costly for what looks on paper like only a few dB distortion improvement. Also, up until fairly recently there was the hump thing which might not be so good for measurement applications.
My preferred dac chip, AK4499, is more complex at least in some ways as compared to ESS parts, IMO. From what I understand from a professional audio consultant in Japan AK4499 sounds much better than AK4497. Still, it sounds best to me if I run it at least down at -6dBFS to -8dBFS. Thus, I think it is important to compare all the numbers between parts under the same measurement conditions. If we are going to say -10dB is the best level for measurement for all dacs then we should measure consistently for all of them.
Anyway, I will concede a slip of the fingers in using a bit stronger adjective than might have been most appropriate: According to manufacturers specifications ESS is somewhat lower distortion than AK4497.
Beyond that, looking at information deeper in AK4497 data sheet and at DAC-3 specifications makes me a bit leery about likelihood of AK4497 distortion at higher sample rates as verses ESS PRO parts at similar high sample rates.
As far as ease of design, AK4497 looks like probably the easiest of the aforementioned dacs to work with.
Thank you for taking the time to write. I appreciate serious commentary by good people. Especially since my memory isn't always perfect when it comes to things I looked at awhile ago.
Let's take a look at the numbers readily available if that would be okay...
0.000062% expressed in dB is -124dB, and you say it was measured at a peak level of -10dB (down about 32% from full scale).
The equivalent dac chip to AK4497 from ESS is ES9038PRO, which is specified by ESS at -122THD+N
AK4499 is specified by AKM at -117THD+N
The mobile ESS part, ES9038Q2M is specified at THD+N of -120dB
If one really wanted to use one of the ESS parts for measurement, I have seen people get the mobile ES9038Q2M down to -127dB distortion at 0dBFS. The PRO part I haven't seen numbers for but one might get it down to -129dBFS, both measured at full scale.
Regarding the -10 level you specify, IME all the dacs seem to have less audible distortion down at -8dB to -10dB. Doubtful it all has to do with intersample overs. Don't have numbers for distortion optimized ESS part at -10db, and have not had much interest in such measurements for reasons which I will attempt to explain below.
The problem with designing a dac to have lower distortion than the manufacturer states on a data sheet is that the dac typically starts to sound worse for playing music, again this in IME. Therefore, I tend to go by manufacturers stated THD numbers when doing comparisons.
By that metric ES9038PRO is 5dB lower in THD+N than AK4497. It may not sound like much when we are talking about numbers well down below -100dB, but I find it matters. I hear the distortion myself. (Also can hear the very different kind of distortion when dacs a designed to give extra low distortion measurement numbers.)
Another thing that concerned me about AK4497 distortion was the table of more detailed distortion numbers on page 11 of the data sheet. https://www.akm.com/content/dam/doc.../audio-dac/ak4497eq/ak4497eq-en-datasheet.pdf
Of course, ESS doesn't publish the corresponding numbers so no fair way to use those for comparison, but AK4497 numbers are concerning to me since they look to get significantly worse at higher sample rates. On the other hand, the only data point we have on ESS are published measurements of Benchmark DAC-3, which runs ES9028PRO at 211kHz for all playback modes due to upsampling. They are able to hit pretty good numbers even up at 211kHz.
https://cdn.shopify.com/s/files/1/0321/7609/files/DAC3_Series_Manual_Rev_B.pdf
Page 38 shows THD at -126dB, and THD+N at -115dB. Somehow noise ends up kind of high, but THD is very low even at the high 211kHz sample rate.
On the other question you mentioned, as far as why manufacturers of low cost test equipment choose to use AK4497 there could be many reasons. It's harder to get the best out ES9038PRO IME and therefore more costly for what looks on paper like only a few dB distortion improvement. Also, up until fairly recently there was the hump thing which might not be so good for measurement applications.
My preferred dac chip, AK4499, is more complex at least in some ways as compared to ESS parts, IMO. From what I understand from a professional audio consultant in Japan AK4499 sounds much better than AK4497. Still, it sounds best to me if I run it at least down at -6dBFS to -8dBFS. Thus, I think it is important to compare all the numbers between parts under the same measurement conditions. If we are going to say -10dB is the best level for measurement for all dacs then we should measure consistently for all of them.
Anyway, I will concede a slip of the fingers in using a bit stronger adjective than might have been most appropriate: According to manufacturers specifications ESS is somewhat lower distortion than AK4497.
Beyond that, looking at information deeper in AK4497 data sheet and at DAC-3 specifications makes me a bit leery about likelihood of AK4497 distortion at higher sample rates as verses ESS PRO parts at similar high sample rates.
As far as ease of design, AK4497 looks like probably the easiest of the aforementioned dacs to work with.
Last edited:
A very good proposal. I agree. I also find that quite a few DAC (and ADC) measure and sound a lot better when the peak is a few dB down from their max level.