• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Review and Measurements of Okto DAC8 8Ch DAC & Amp

Feyire

Active Member
Joined
Mar 9, 2019
Messages
272
Likes
314
Location
Netherlands

tensor9

Active Member
Forum Donor
Joined
Jun 26, 2018
Messages
149
Likes
90

From ESS’s own lead engineer, it has not been fully solved and it never will be because digital control does not attenuate the noise floor. He himself admitted analog is better. In fact, the very presentation you linked says just that. See Benchmark’s LA4 for a proper analog volume control.

Now, as a practical matter, it probably doesn’t really matter and digital is sufficient, but from and engineering and first principles point of view, there is a better way to do it.

For instance, the Matrix Element X has a hybrid digital/analog volume control, although Amir did not test its efficacy.
 

1234VICE

Member
Joined
Jul 25, 2019
Messages
27
Likes
27
Location
The Netherlands
From ESS’s own lead engineer, it has not been fully solved and it never will be because digital control does not attenuate the noise floor. He himself admitted analog is better. In fact, the very presentation you linked says just that. See Benchmark’s LA4 for a proper analog volume control.

Now, as a practical matter, it probably doesn’t really matter and digital is sufficient, but from and engineering and first principles point of view, there is a better way to do it.

For instance, the Matrix Element X has a hybrid digital/analog volume control, although Amir did not test its efficacy.

The analogue noise floor of the dac8 is so low though that quantization noise should be dominant till ~-23dB for 16 bit signals. So unless you want to resolve the quantization noise of 24 bits signal, the dac8 will act as a perfect volume control in most cases.
 

MWC

Active Member
Joined
Apr 15, 2019
Messages
153
Likes
64
Well of course we 'want to resolve the quantization noise of 24 bits signal' ! Why imagine anyone interested in this DAC8 Pro would want to limit themselves to 16 bit?
 

1234VICE

Member
Joined
Jul 25, 2019
Messages
27
Likes
27
Location
The Netherlands
Well of course we 'want to resolve the quantization noise of 24 bits signal' ! Why imagine anyone interested in this DAC8 Pro would want to limit themselves to 16 bit?

I am not fully sure whether or not your post is meant to be ironic, but if that is truly your goal, it is a much bigger challenge than replacing a digital volume control by an analogue one. It seems to me unpractical even to achieve the s/n of 16 bits.

Another thing to keep in mind is that the noise floor of the actual recording might be higher than the quantization noise. This will further extend the range in which the dac8 operates as an ideal volume control.

The people who are put off by a digital volume control have the opportunity to buy a separate analogue preamp. It makes perfect sense to me that the dac8 does not come with one, as most of us are not even remotely close to being bottlenecked by its analogue noise floor.
 

SpyB

Active Member
Joined
Sep 3, 2019
Messages
142
Likes
103
how low of a noise measurement is sufficient,before the law of diminishing returns kicks in.

A know there are some that will say,as low a figure as can be achieved,although all this

noise + distortion war, all this competition to achieve the the lowest level of distortion

makes one think,that if the figures are lowest, we have the best DAC.

While I think,that there must surely be a point where it does not matter.

What is that point, I ask
 

bigjacko

Addicted to Fun and Learning
Joined
Sep 18, 2019
Messages
721
Likes
359
I also have the same question. When we have headphones and speakers that can only do -60db at best, how much would the dac amp be so that it is not acceptable. I am also interested in manually making distortion to try but can't find a program to do it. I have seen that the distortion needs to be pretty high for people to start notice the distortion.
 

SpyB

Active Member
Joined
Sep 3, 2019
Messages
142
Likes
103
I also have the same question. When we have headphones and speakers that can only do -60db at best, how much would the dac amp be so that it is not acceptable. I am also interested in manually making distortion to try but can't find a program to do it. I have seen that the distortion needs to be pretty high for people to start notice the distortion.

Quote: "I have seen that the distortion needs to be pretty high for people to start notice the distortion."

My point exactly,the infinitesimal low distortion figures of a DAC can not be the guide,

if DAC (A) is superior to DAC (B),sure it tells us that it has lower distortion.

Although if it is beyond human hearing,then it is irrelevant.

And I think if you prefer one over the other, it is subjective anyway /I like this DAC,you like that DAC.

You buy DAC (A) and do not like it,at some point you sell it and buy DAC(B).

It is all trial and error until you find, you have finally found

a synergistic match of components that simply give you what you want.

Again a subjective decision,put someone else in the room and it is quiet possible it will not be to there liking.

Just my thoughts anyway
 
Last edited:

MWC

Active Member
Joined
Apr 15, 2019
Messages
153
Likes
64
@1234VICE I'm not sure I understand your explanations it seems to me that you are completely disregarding any music recorded at 24bit, as if that is alien. Maybe 24bit does not feature in your world, but it does in the real world.
 

somebodyelse

Major Contributor
Joined
Dec 5, 2018
Messages
3,683
Likes
2,963
@1234VICE I'm not sure I understand your explanations it seems to me that you are completely disregarding any music recorded at 24bit, as if that is alien. Maybe 24bit does not feature in your world, but it does in the real world.
I think you don't understand the explanation. 24 bit files exist in the real world. So do noise floors in recordings, preamps, headphone amps, power amps, and rooms. Then there are limitations on human hearing. 16 bit (96dB) is the difference between a quiet room and hearing damage. 20 bit (120dB) is the commonly quoted range of human perception, from the threshold of hearing to hearing damage. Once the noise is quieter than the quietest thing you can hear it doesn't make any difference.
 

MWC

Active Member
Joined
Apr 15, 2019
Messages
153
Likes
64
@somebodyelse Thanks for shedding some clarity on this. One gets confused when comments are not specific and the commentator assumes prior knowledge of the readers, or one writes from only one's own perspective. I'm into music and the technology discussed here is only a means to my enjoyment and production of music, so engineers writing from their unique skill set are not quite speaking the same language as me. I respect that it is tough for those who's native language is not English to explain themselves here (in English) rather than in their native tongue. Sometimes one needs to find other/extra words, if one wants to consider another's understanding of what one is attempting to say. Despite the potential to be accused of being verbose. :)
 
Last edited:

1234VICE

Member
Joined
Jul 25, 2019
Messages
27
Likes
27
Location
The Netherlands
@1234VICE I'm not sure I understand your explanations it seems to me that you are completely disregarding any music recorded at 24bit, as if that is alien. Maybe 24bit does not feature in your world, but it does in the real world.

I can elaborate, please try to keep the conversation civilized though. If a preamp and 24 bit recordings makes you happy go for it, I am not plotting some scheme to sabotage your listening experience.

I few reasons why I think that resolving 24 bit quantization noise is unpractical:
-As mentioned by oktoresearch in this thread, the environmental noise in a typical house dwarfs the analogue noise by ~70-80dB.
-24 bit recordings might have noise much higher than the quatization noise of the medium, so it is not sufficient merely to have 24 bit recordings.
-If the noise floor of the amp is higher than the analogue noise floor of the dac8 + the amp gain, the amp will be the bottleneck in all cases. If is not, it can still be the bottleneck at normal volume levels.
-The THD+N of your speakers will be the bottleneck of your equipment ultimately.
-Once you have tackled all these problems, the final blocking factor is human hearing.

You could plug in the dac8 to your amp and try to listen if it introduces noise by putting your ear against the speaker drivers. If there is noise, see if it lowers by unplugging the dac8. If there is nothing to observe, you can be pretty confident that you wont be bothered by the analogue noise of this unit at the listening position. After you go through these steps, you can safely consider that the dac8 volume control only affects output level, as the analogue noise is not an audible factor.
 

1234VICE

Member
Joined
Jul 25, 2019
Messages
27
Likes
27
Location
The Netherlands
I
You could plug in the dac8 to your amp and try to listen if it introduces noise by putting your ear against the speaker drivers. If there is noise, see if it lowers by unplugging the dac8. If there is nothing to observe, you can be pretty confident that you wont be bothered by the analogue noise of this unit at the listening position. After you go through these steps, you can safely consider that the dac8 volume control only affects output level, as the analogue noise is not an audible factor.

To clarify; without playing music. So just to check the noise going through the silent system.
 

Kees Huizinga

Member
Joined
Aug 21, 2016
Messages
45
Likes
33
Are there plans to develop an own ASIO-driver for the Okto-dac8? I ask this because the Asio4All driver doesn't work in my set-up. I am using dePhonica-software (dephonica.com ) to do the crossover and dsp. This program is running into JRiver with it's own ASIO-Sink driver. With an Asus multi-channel soundcard or the DIYINHK Multichannel dac, with their own ASIO-drivers, it's running fine, but not with Asio4All.
I think it is because only one application can play audio at a time when ASIO4ALL is in use, but correct me when I'm wrong.
 

dean70

Member
Joined
Oct 29, 2019
Messages
8
Likes
3
Location
Planet Earth
I plan to use the DAC8 for 7.1 channel home theater, but also want to use for 2 channel duties with HQPlayer. Does the DAC support higher than DSD128/PCM 192K with 2 channel audio? The Sabre 9038pro? should support DSD512/PCM 768k - is there a reason for the limited rate?
 

Okto Research

Member
Joined
Dec 4, 2018
Messages
55
Likes
588
Location
Prague
Are there plans to develop an own ASIO-driver for the Okto-dac8? I ask this because the Asio4All driver doesn't work in my set-up. I am using dePhonica-software (dephonica.com ) to do the crossover and dsp. This program is running into JRiver with it's own ASIO-Sink driver. With an Asus multi-channel soundcard or the DIYINHK Multichannel dac, with their own ASIO-drivers, it's running fine, but not with Asio4All.
I think it is because only one application can play audio at a time when ASIO4ALL is in use, but correct me when I'm wrong.
The audio manufacturers mostly do not develop their own ASIO driver, but they licence it from a German company called Thesycon. We are deciding on the following steps to take regarding a Windows ASIO driver. But we have tested ASIO4ALL as an output of JRiver, combined with JRiver's "virtual soundcard" WDM driver and its DSP studio and this combination works well. We didn't test with dePhonica though.

16 bit (96dB) is the difference between a quiet room and hearing damage. 20 bit (120dB) is the commonly quoted range of human perception, from the threshold of hearing to hearing damage. Once the noise is quieter than the quietest thing you can hear it doesn't make any difference.
Exactly. But it is also good to have some DNR headroom for DSP purposes, since to make a "+20dB" correction in bass region, actually the rest of the audio band needs to be attenuated by that amount to avoid clipping, so except for the corrected region, 20dB of DNR is lost (regardless of whether an analog or a digital attenuation is applied later on).

I am also interested in manually making distortion to try but can't find a program to do it. I have seen that the distortion needs to be pretty high for people to start notice the distortion.
A single-tone THD measurement is a useful indicator of a nonlinearity of a transfer function, but the harmonic distortion components themselves (as per their name) are in harmony with the original tone, being at whole-number multiple of its frequency (according to the music theory a whole number of octaves higher), so although they do produce coloration, they are not offensive. When multiple tones are introduced, their interaction on such a nonlinearity creates tones at differences and sums of the original frequencies (intermodulation) - not in harmony with the originals. That is addressed by the 32-tone test in ASR reviews, which is close to a real music data.

A synthetic test to determine distortion audibility should include a base tone and a non-integral "offender" tone (maybe a swept one).

I plan to use the DAC8 for 7.1 channel home theater, but also want to use for 2 channel duties with HQPlayer. Does the DAC support higher than DSD128/PCM 192K with 2 channel audio? The Sabre 9038pro? should support DSD512/PCM 768k - is there a reason for the limited rate?
No, 192kHz PCM and DSD128 are the maximum data rates for the DAC8 PRO regardless the number of active channels. The limitation is due to XMOS firmware for 16 channels in total (8 in, 8 out), but honestly, we don't see much reason to go further. Performance of DAC chips gets worse at high sample rates (that applies to both ESS as well as AKM ones) so you gain bandwidth in ultrasonic (inaudible) region in expense of making things worse within the audible region.

Pavel, Okto Research
 

mlknez

New Member
Joined
Oct 27, 2019
Messages
1
Likes
1
...

No, 192kHz PCM and DSD128 are the maximum data rates for the DAC8 PRO regardless the number of active channels. The limitation is due to XMOS firmware for 16 channels in total (8 in, 8 out), but honestly, we don't see much reason to go further. Performance of DAC chips gets worse at high sample rates (that applies to both ESS as well as AKM ones) so you gain bandwidth in ultrasonic (inaudible) region in expense of making things worse within the audible region.

Pavel, Okto Research

The reason that I would want support is to play ALL of the files that I currently have and not have to worry about decimation. I realize there may be no perceptible difference, but is there a reason why you wouldn't support all possible rates and also issue a warning?
 
  • Like
Reactions: MWC
Top Bottom