So you'd prefer AK4493/97 over ESS9038/39?
(I want to try them all but only had 9038q2m so far)
Had other things on my mind but slowly I got the feeling that the numbers game is not the complete picture. Not willing anymore to go into great depths only to discover my DIY endeavours will at least be 4 times more expensive as ready built ones I just buy affordable stuff, modify or correct errors when necessary and listen.
All DACs I have or can use at the moment perform subjectively better than dual ES9038 in my Eversolo DMP-A6, ES9028 devices and especially ES9018 I compared them with. Even my own ES9023 DAC outperforms some recent ESS chips based DACs. The emotion is that I start to dislike ESS, their silly NDA policy, their game and the herd of blind followers.
Then I used transformers connected to the Eversolos outputs. Bam, good sound. I see no difference whatsoever but it can easily be heard. RF, leakage current?!?
Sorry eclipsevl, missed your post. Yes AK4493SEQ and AK4497 anytime over ES9038. Maybe the 9039 will convince, I don’t know yet. I don’t run as fast anymore for audio and try to actually use the stuff 😉
All DACs I have or can use at the moment perform subjectively better than dual ES9038 in my Eversolo DMP-A6, ES9028 devices and especially ES9018 I compared them with. Even my own ES9023 DAC outperforms some recent ESS chips based DACs. The emotion is that I start to dislike ESS, their silly NDA policy, their game and the herd of blind followers.
Then I used transformers connected to the Eversolos outputs. Bam, good sound. I see no difference whatsoever but it can easily be heard. RF, leakage current?!?
Sorry eclipsevl, missed your post. Yes AK4493SEQ and AK4497 anytime over ES9038. Maybe the 9039 will convince, I don’t know yet. I don’t run as fast anymore for audio and try to actually use the stuff 😉
Last edited:
Numbers game is not the complete picture but it is an indicator of implementation quality. IME the reason for not reaching datasheet levels is most often flaws in design or implementation.
And some prominent designers seem to believe in numbers game. Here is what Bruno Putzeys said in an interview:
"It is my experience -- confirmed by every new thing I do -- that when you get into really high measured performance, really low distortion, superlow noise, then the ultimate subjective sound quality starts improving and continues to improve in step with the measurements. At some point you will find that a product that measures absolutely perfectly under an extensive battery of tests will sound a lot better than a product with more typical high-end audio performance that has been tuned by ear for years."
(from https://www.soundstageultra.com/ind...s-of-mola-mola-hypex-and-grimm-audio-part-one)
And some prominent designers seem to believe in numbers game. Here is what Bruno Putzeys said in an interview:
"It is my experience -- confirmed by every new thing I do -- that when you get into really high measured performance, really low distortion, superlow noise, then the ultimate subjective sound quality starts improving and continues to improve in step with the measurements. At some point you will find that a product that measures absolutely perfectly under an extensive battery of tests will sound a lot better than a product with more typical high-end audio performance that has been tuned by ear for years."
(from https://www.soundstageultra.com/ind...s-of-mola-mola-hypex-and-grimm-audio-part-one)
Many persons that sell stuff declare their stuff indirectly or directly to be the best. Nothing new there.
Many persons that sell stuff declare the above mentioned stuff to have imperfections and enhancements that were done on the type they introduced today.
Many users with deep pockets of the follower type buy such products that measure perfectly to connect them to excellent eh…..tube amplifiers 😂 It is all relative.
If flaws in implementation are that relevant manufacturers could choose to have all things necessary embedded in just 1 IC on a tiny PCB with just a 5V connection and local regulation. Guaranteed always the same performance from the ESS Watermelon DAC 1, 2 and so on. Maybe the tiny implementation details keep the industry going.
Many persons that sell stuff declare the above mentioned stuff to have imperfections and enhancements that were done on the type they introduced today.
Many users with deep pockets of the follower type buy such products that measure perfectly to connect them to excellent eh…..tube amplifiers 😂 It is all relative.
If flaws in implementation are that relevant manufacturers could choose to have all things necessary embedded in just 1 IC on a tiny PCB with just a 5V connection and local regulation. Guaranteed always the same performance from the ESS Watermelon DAC 1, 2 and so on. Maybe the tiny implementation details keep the industry going.
Last edited:
Don’t underestimate designers with commercial talent.
Given the already extreme numbers: how much will we actually be able to hear from yet another few dB improvement? DAC technology seems ripe. So ripe that some walk back to the past to not miss out on that good old jittery 16 bit technology.
Given the already extreme numbers: how much will we actually be able to hear from yet another few dB improvement? DAC technology seems ripe. So ripe that some walk back to the past to not miss out on that good old jittery 16 bit technology.
Last edited:
I agree that DAC technology seems ripe in the sense that sound differences in well designed DACs are close to non-audible.
But why settle for less? Do you make your voltage regulators perform worse than stated in the datasheet because all that performance is not needed?
But why settle for less? Do you make your voltage regulators perform worse than stated in the datasheet because all that performance is not needed?
Not settle for less but what sense does it make for me as an enthusiast to invest again in extreme performance under the hood if the maximum speed allowed is 80 kmh? The current one already can do 220 kmh which it never runs. The latest one does 250 kmh. Probably not good analogy but imagine comparing flagship stuff and everyone prefers the cheapest most basic device. Happens with DACs. SMSL SU-1 is great with slight modifications.
Probably the market will be owned by the Chinese companies that have really improved DAC products (I mean that, kudos to them) and made them affordable as well. Can’t say that of Mola Mola. No middle field but either cheap and good or very expensive show off stuff.
Probably the market will be owned by the Chinese companies that have really improved DAC products (I mean that, kudos to them) and made them affordable as well. Can’t say that of Mola Mola. No middle field but either cheap and good or very expensive show off stuff.
Last edited:
did you publish more than the graph, e.g. a description of your configuration, a picture of the board or even a schematic? thanks
Hi I thought about this when listening to various devices in the background.
A clearly imperfect transformer output DAC called Quasimodo with slight hum performed best subjectively. A visitor by brand X famous from ASR did not attract attention. Mmmmm.
The output stage I'm using consists of just an I/V op amp stage but the op amp is a composite. There are many ways to implement a composite op amp. Mine is similar to what Samuel Groner and Martin Polak published in 2017 (https://www.diyaudio.com/community/threads/samuel-groners-super-opamp.337866/). That composite op amp provides the differential output and for single-ended output there is a summing op amp so in total 5 op amps per channel.
Attached is a picture of the ES9039Q2M dac and another with my dual AK4493 dac. The board below is my USB-I2S board. All my diy dacs have the same format so they can be fitted in the same chassis (shown here with RTZ dac).
Attachments
Basically one offs but there are other members using my stuff as well. I'm not too keen on spending time supporting other people tackling with my designs. This hobby is already far too time-consuming.
Same over there. Built DACs as a habit since my youth when finding parts was a real challenge. All TH slow technology and no education available in the field. Many errors and assumptions, many mistakes but some were better than the often mediocre stuff produced then.
Today I feel that the 2 box approach is silly and 1 device too many. I demand my gear to have decent DACs internally (no one listens 🙂) so I only test/listen external DACs but use the internal DACs mainly even though they are not ultra non plus quality. It is just technology reproducing music and should not be a too large factor in a life and also not have a larger importance over the content itself. Content of which about 10% is listenable and then often the music that does not appeal 😀
Today I feel that the 2 box approach is silly and 1 device too many. I demand my gear to have decent DACs internally (no one listens 🙂) so I only test/listen external DACs but use the internal DACs mainly even though they are not ultra non plus quality. It is just technology reproducing music and should not be a too large factor in a life and also not have a larger importance over the content itself. Content of which about 10% is listenable and then often the music that does not appeal 😀
Last edited:
Thanks for the link to the super opamp link. I was thinking along the same lines after seeing the output stages proposed by ESS for the 9039q2m and especially the PRO. Those OPA1612 see an awfully low load, so a composite will help just by removing the output stage from the input stage, even before the extra loop gain kicks in.
I am not too keen on using a resistive divider in composite op amps because it throws away gain, but I guess if you need to be unity gain stable, that may be what you have to do.
Do you think old 16 bit dac chips are more prone to jitter than modern delta sigmas?
Well I am sure many old devices containing 16 bit DACs as a whole had more jitter in general. External older DACs were often worse than the internal ones because of SPDIF and jittery receivers. When PCM1710 was introduced mine at first only produced noise with music in the background.
Maybe/possibly modern DAC technology is more sensitive to it intrinsically but at the same time methods to reduce it were developed and embedded in the ICs and also clock and interface technology improved.
So no hard statement possible.
Maybe/possibly modern DAC technology is more sensitive to it intrinsically but at the same time methods to reduce it were developed and embedded in the ICs and also clock and interface technology improved.
So no hard statement possible.
Last edited:
Well, really old DACs with R2R design certainly were more sensitive to jitter. You get some uncertainty window on when the MSB is latched and you have big time distortion. Sigma-Delta designs are much more robust in this regard because their MSBs are smaller. Then, DIR and ASRCs (asynchronous sample rate coverters) can attenuate jitter, and ESS DACs even have a built-in ASRC.
In modern delta sigmas MCK only is critical and it is fairly easy to provide low jitter MCK straight from clock. Whereas in older DACs either word clock or bit clock is critical. Those are in most cases generated upstream and can have more jitter than MCK.
When digital audio was introduced the average hobby person could not influence anything regarding the chips and their technology. Technology was so new that even professionals were coping to understand. Many just copied.
Habit then was to use predefined SPDIF receiver-digital filter-DAC chipsets made by only a few companies according their hard to grasp datasheets. Obtaining the chips was the beg, steal or borrow method. It did not occur to many to be busy with inchip or technology/protocol pecularities. No internet, no books, no oscilloscope, no simulation, only prototyping in real life and all autodidact. The power supplies and output stage was where most action was. Analog knowledge.
Only the external factors were to be manipulated certainly by hobby persons. The first DACs I built were based on disassembled expensive devices photographed 1:1 with very large analog cameras. The film was manipulated and own improvement was implemented, new PCB tracks drawn with black ink pens and the film then was used for PCB production.
Habit then was to use predefined SPDIF receiver-digital filter-DAC chipsets made by only a few companies according their hard to grasp datasheets. Obtaining the chips was the beg, steal or borrow method. It did not occur to many to be busy with inchip or technology/protocol pecularities. No internet, no books, no oscilloscope, no simulation, only prototyping in real life and all autodidact. The power supplies and output stage was where most action was. Analog knowledge.
Only the external factors were to be manipulated certainly by hobby persons. The first DACs I built were based on disassembled expensive devices photographed 1:1 with very large analog cameras. The film was manipulated and own improvement was implemented, new PCB tracks drawn with black ink pens and the film then was used for PCB production.
Last edited: