|
|
Post by solidstate on Mar 25, 2013 6:44:27 GMT -5
And...it is this very logic I was referring to. I value hearing this type of thing. Ever company has to decide what features to include and not include. I really value Emotiva's candor in sharing their logic behind these choices. One can agree or not with their choices, but each time they share their logic - it makes me think harder about my choices. As I said, I do not disagree with their choice of a 24-bit DAC. But the "logic" that it is useless to use a 32-bit DAC because there is no 32-bit content misses the point. Of course there is no 32-bit content. That is not why 32-bit DACs were developed. Nor why the DSP used in the UMC-200 outputs 32-bit data, or has 72-bit accumulators. What he shared was not logic, but an illogical misunderstanding. Was this seriously his retort ? Didn't it occur to him that some people on here are engineers and would understand that's not exactly true !?! That makes me question his technical knowledge because it hints at a lack of understanding DACs and the way in-which they receive and process audio. Probably just a PR biz guy brainfart because it's hard to explain to consumers the maths et al and he's right in 90% of implementations/sources/transport. The reason you want 32bit is for volume attenuation. Lonny should understand that if you feed a UMC-200 a 24bit signal the digital volume attenuation on the converter will really create a lot of noise and because of the converter choice I wouldn't feed it 24bit in my transport chain. All of this depends on the design. If the DSP engine does 32bit attenuation then feeds 24bit attenuated audio to the converter vis I2S then it's not a problem with that DAC. If the DSP engine does volume attenuation calculations and feeds the results to the DAC so the DAC can attenuate you've got a problem. This is only with 24bit sources though. Usually with room correction the DSP engine does the PEQ and attenuation/channel trims then passes this signal via I2S to the converters. BUT in the designs I've looked at it the global master volume adjusts the DACs volume at the DAC not in the DSP. This is the case with 90% of designs. Perhaps this white paper might clarify for Lonny. www.esstech.com/PDF/digital-vs-analog-volume-control.pdfI don't know much about the UMC-200's design. Perhaps it has a separate analog volume control but I doubt it as it's a multi chan design and why would you have that when you're choosing the lower cost converter in the first place. |
|
|
|
Post by solidstate on Mar 25, 2013 7:03:03 GMT -5
There are two different "types" of 32-bit DAC chips, though. Some can accept 32-bit input, while others are still limited to 24-bit input but internally upconvert it to 32-bit (such as the ESS Technologies SABRE³² Reference ES9018). That's not true and I've never heard of a DAC that can process in 32bits but only receive 24bits on it's bus. Must have been short lived SKUs during transition. Hyperstream has I2S 32bit input as well as DSD and SP/DIF. It's sample rate can be as high as 1.536 Mhz. |
|
|
|
Post by solidstate on Mar 25, 2013 8:03:04 GMT -5
I would love to see a series of EXTERNAL 8P8C sockets on the back of gear with I2S so I can pump the DSP's audio into my OWN EXTERNAL CONVERTERS. Man I'd love to see that on the XMC-1 or one of it's future brethren. I've thought for sometime about lifting that buss off of a Denon AVR and placing my own headers/sockets on the back of the chassis. The tricky part would be master/global volume control as I said earlier this attenuation is done at the DAC usually and only PEQ/RQ/chan trim is done in DSP. Take an AVR 3313CI, yank the entire amp section, lift I2S on it's way to the converter and move it to the back of chassis on a 8P8C sockets. Or you simply make your AVR/pre/pro have modular DAC slash I/V on PCB plugin modules. That's a brilliant idea actually IMHO I've had for YEARS... You would have then after market converter I/V options for different audio tastes. At this point in the game if people haven't figured out the DAC and even more so the I/V is EVERYTHING. The difference is sound quality in most gear today is the bloody opamps driving the line level signal to the amplifiers. A lot of these opamps suck ass when compared to a discrete solution like a DEXA, Burson or Kingwa drop in units. DEXA www.newclassd.com/index.php?page=55Burson bursonaudio.com/DIY_HD_Opamp.htmKingwa www.audio-gd.com/Pro/diy/OPA/OPAEN.htmAs usual Kingwa kicks everyone's asses. |
|
KeithL
Administrator 
Posts: 10,261
|
Post by KeithL on Mar 25, 2013 9:50:25 GMT -5
Sure, it COULD be done, but it wouldn't make as much sense as you think..... 1) I2S is VERY fussy about cabling and lead length - adding twelve or eighteen inches of ribbon cable in between might ruin the performance altogether. 2) With multiple DAC units, you have to worry about channel matching (phase, delay, etc) between units - unless you have a single 8-channel unit. I haven't seen very many good audiophile quality 8 channel I2S input DACs lately. 3) A significant portion of what you listen to in surround sound is quite heavily processed - and very little of it is "worthy" of obsessing over minutae of sound quality. You'll be lucky to find a stereo audiophile recording that hasn't passed through at least fifty fifty-cent op amps on its way to the console. My guess is that you won't find ANY surround sound ones. You're much better spending your money on a good analog preamp with HT bypass and going that route. Trust us to make a good choice for what DACs and op-amps to use. 4) The difference in how "DACs" sound is mostly due to ancillary circuitry anyway - and very little of it is due to the DAC chips themselves. Therefore, while this would be cool for DIY types, it isn't going to be of much use to very many people. Also, you can forget about re-clocking and such things - since the delay they introduce would mess up the audio/video synch. 5) A "good" op amp should have no sound whatsoever (in the application for which it is chosen). It's sort of like spark plugs - if they work oddly, then there's a problem; if they work right, you shouldn't notice them. From what I've heard, SOME of the discrete versions are quite good, while some are actually rather bad. From an engineering point of view, if the op-amp changes the sound, then there's something wrong - either with the op-amp or your circuit. (Not to mention the fact that the added contact resistance and capacitance of the socket you would need to include so you could swap the op-amp will ALSO make changes for the worse.) The short answer is that we don't have enough customers who want a "modular pre/pro". If we wanted to sell them, we would have to spread the development costs to do so amongst the actual number of customers who bought them, and the end result would be a VERY expensive product - too expensive for anybody to buy. I would love to see a series of EXTERNAL 8P8C sockets on the back of gear with I2S so I can pump the DSP's audio into my OWN EXTERNAL CONVERTERS. Man I'd love to see that on the XMC-1 or one of it's future brethren. I've thought for sometime about lifting that buss off of a Denon AVR and placing my own headers/sockets on the back of the chassis. The tricky part would be master/global volume control as I said earlier this attenuation is done at the DAC usually and only PEQ/RQ/chan trim is done in DSP. Take an AVR 3313CI, yank the entire amp section, lift I2S on it's way to the converter and move it to the back of chassis on a 8P8C sockets. Or you simply make your AVR/pre/pro have modular DAC slash I/V on PCB plugin modules. That's a brilliant idea actually IMHO I've had for YEARS... You would have then after market converter I/V options for different audio tastes. At this point in the game if people haven't figured out the DAC and even more so the I/V is EVERYTHING. The difference is sound quality in most gear today is the bloody opamps driving the line level signal to the amplifiers. A lot of these opamps suck ass when compared to a discrete solution like a DEXA, Burson or Kingwa drop in units. DEXA www.newclassd.com/index.php?page=55Burson bursonaudio.com/DIY_HD_Opamp.htmKingwa www.audio-gd.com/Pro/diy/OPA/OPAEN.htmAs usual Kingwa kicks everyone's asses. |
|
KeithL
Administrator
Posts: 10,261
|
Post by KeithL on Mar 25, 2013 9:57:35 GMT -5
Somebody needs to stop being such a sourpuss... and do their homework ;D Actually, the UMC-200 DOES have a digitally controlled analog ladder network volume control... Which means that you DON'T lose resolution when you turn the volume down. (It also means that, if you use the Direct input setting, and have an analog input signal, your signal isn't even digitized.) But.... thanks for the great explanation about why you really need a 32 bit DAC IF you're going to use a digital volume control.  As I said, I do not disagree with their choice of a 24-bit DAC. But the "logic" that it is useless to use a 32-bit DAC because there is no 32-bit content misses the point. Of course there is no 32-bit content. That is not why 32-bit DACs were developed. Nor why the DSP used in the UMC-200 outputs 32-bit data, or has 72-bit accumulators. What he shared was not logic, but an illogical misunderstanding. Was this seriously his retort ? Didn't it occur to him that some people on here are engineers and would understand that's not exactly true !?! That makes me question his technical knowledge because it hints at a lack of understanding DACs and the way in-which they receive and process audio. Probably just a PR biz guy brainfart because it's hard to explain to consumers the maths et al and he's right in 90% of implementations/sources/transport. The reason you want 32bit is for volume attenuation. Lonny should understand that if you feed a UMC-200 a 24bit signal the digital volume attenuation on the converter will really create a lot of noise and because of the converter choice I wouldn't feed it 24bit in my transport chain. All of this depends on the design. If the DSP engine does 32bit attenuation then feeds 24bit attenuated audio to the converter vis I2S then it's not a problem with that DAC. If the DSP engine does volume attenuation calculations and feeds the results to the DAC so the DAC can attenuate you've got a problem. This is only with 24bit sources though. Usually with room correction the DSP engine does the PEQ and attenuation/channel trims then passes this signal via I2S to the converters. BUT in the designs I've looked at it the global master volume adjusts the DACs volume at the DAC not in the DSP. This is the case with 90% of designs. Perhaps this white paper might clarify for Lonny. www.esstech.com/PDF/digital-vs-analog-volume-control.pdfI don't know much about the UMC-200's design. Perhaps it has a separate analog volume control but I doubt it as it's a multi chan design and why would you have that when you're choosing the lower cost converter in the first place. |
|
|
|
Post by mjrtoo on Mar 25, 2013 10:01:47 GMT -5
Somebody needs to stop being such a sourpuss... and do their homework ;D Actually, the UMC-200 DOES have a digitally controlled analog ladder network volume control... Which means that you DON'T lose resolution when you turn the volume down. (It also means that, if you use the Direct input setting, and have an analog input signal, your signal isn't even digitized.) But.... thanks for the great explanation about why you really need a 32 bit DAC IF you're going to use a digital volume control. Sourpuss, haven't heard that in a long time, but it fits.  I'm new here, but I swear I've read all this in a different thread before as well. Anyway, the analog ladder VC in this unit is a fairly high-end feature, pretty cool! |
|
|
|
Post by solidstate on Mar 25, 2013 15:32:43 GMT -5
Sure, it COULD be done, but it wouldn't make as much sense as you think..... 1) I2S is VERY fussy about cabling and lead length - adding twelve or eighteen inches of ribbon cable in between might ruin the performance altogether. 2) With multiple DAC units, you have to worry about channel matching (phase, delay, etc) between units - unless you have a single 8-channel unit. I haven't seen very many good audiophile quality 8 channel I2S input DACs lately. 3) A significant portion of what you listen to in surround sound is quite heavily processed - and very little of it is "worthy" of obsessing over minutae of sound quality. You'll be lucky to find a stereo audiophile recording that hasn't passed through at least fifty fifty-cent op amps on its way to the console. My guess is that you won't find ANY surround sound ones. You're much better spending your money on a good analog preamp with HT bypass and going that route. Trust us to make a good choice for what DACs and op-amps to use. 4) The difference in how "DACs" sound is mostly due to ancillary circuitry anyway - and very little of it is due to the DAC chips themselves. Therefore, while this would be cool for DIY types, it isn't going to be of much use to very many people. Also, you can forget about re-clocking and such things - since the delay they introduce would mess up the audio/video synch. 5) A "good" op amp should have no sound whatsoever (in the application for which it is chosen). It's sort of like spark plugs - if they work oddly, then there's a problem; if they work right, you shouldn't notice them. From what I've heard, SOME of the discrete versions are quite good, while some are actually rather bad. From an engineering point of view, if the op-amp changes the sound, then there's something wrong - either with the op-amp or your circuit. (Not to mention the fact that the added contact resistance and capacitance of the socket you would need to include so you could swap the op-amp will ALSO make changes for the worse.) The short answer is that we don't have enough customers who want a "modular pre/pro". If we wanted to sell them, we would have to spread the development costs to do so amongst the actual number of customers who bought them, and the end result would be a VERY expensive product - too expensive for anybody to buy. I would love to see a series of EXTERNAL 8P8C sockets on the back of gear with I2S so I can pump the DSP's audio into my OWN EXTERNAL CONVERTERS. Man I'd love to see that on the XMC-1 or one of it's future brethren. I've thought for sometime about lifting that buss off of a Denon AVR and placing my own headers/sockets on the back of the chassis. The tricky part would be master/global volume control as I said earlier this attenuation is done at the DAC usually and only PEQ/RQ/chan trim is done in DSP. Take an AVR 3313CI, yank the entire amp section, lift I2S on it's way to the converter and move it to the back of chassis on a 8P8C sockets. Or you simply make your AVR/pre/pro have modular DAC slash I/V on PCB plugin modules. That's a brilliant idea actually IMHO I've had for YEARS... You would have then after market converter I/V options for different audio tastes. At this point in the game if people haven't figured out the DAC and even more so the I/V is EVERYTHING. The difference is sound quality in most gear today is the bloody opamps driving the line level signal to the amplifiers. A lot of these opamps suck ass when compared to a discrete solution like a DEXA, Burson or Kingwa drop in units. DEXA www.newclassd.com/index.php?page=55Burson bursonaudio.com/DIY_HD_Opamp.htmKingwa www.audio-gd.com/Pro/diy/OPA/OPAEN.htmAs usual Kingwa kicks everyone's asses. I2S over an HDMI cable is being used by various outfits right now.  PS Audio and W4S have balanced I2S I/O with a quality buffer. It also helps to avoid DAC SRC. The volume and the way the XDA-1 DAC was programmed... no comment. Personally I don't see the analogy between an opamp and a sparkplug but hey the luddites will buy it. |
|
|
|
Post by solidstate on Mar 25, 2013 15:35:02 GMT -5
Somebody needs to stop being such a sourpuss... and do their homework ;D Actually, the UMC-200 DOES have a digitally controlled analog ladder network volume control... Which means that you DON'T lose resolution when you turn the volume down. (It also means that, if you use the Direct input setting, and have an analog input signal, your signal isn't even digitized.) But.... thanks for the great explanation about why you really need a 32 bit DAC IF you're going to use a digital volume control. I take it the resistor volume is only for L/R right ? Can you please share what volume chip you guys used and why do I hear no mention of it on the product specs nor from any reviews ? |
|
|
|
Post by solidstate on Mar 25, 2013 15:45:56 GMT -5
The difference in how "DACs" sound is mostly due to ancillary circuitry anyway - and very little of it is due to the DAC chips themselves. Therefore, while this would be cool for DIY types, it isn't going to be of much use to very many people. Also, you can forget about re-clocking and such things - since the delay they introduce would mess up the audio/video synch. A "good" op amp should have no sound whatsoever (in the application for which it is chosen). It's sort of like spark plugs - if they work oddly, then there's a problem; if they work right, you shouldn't notice them. From what I've heard, SOME of the discrete versions are quite good, while some are actually rather bad. From an engineering point of view, if the op-amp changes the sound, then there's something wrong - either with the op-amp or your circuit. (Not to mention the fact that the added contact resistance and capacitance of the socket you would need to include so you could swap the op-amp will ALSO make changes for the worse.) Huh... I don't believe that for a second. You start by saying that the DAC isn't that big a deal and that the "ancillary circuitry" is what makes the difference only to go on and basically say all opamps should sound that same. You contradict yourself !!! If opamps are all the same and discrete doesn't make any dif then why do you guys use discrete on some of your pieces ? Hey I wanna ask something else of you. I've heard recently, though I don't no if true, that some XDA-2s have different analog section and some are opamp based and others are discrete. Is this true ? |
|
|
|
Post by monkumonku on Mar 25, 2013 15:59:35 GMT -5
Well just a couple of my two cents observations about this thread... (1) I could care less what DAC or what spec or whatever is in my Emo gear. I care about what it sounds like. It seems to me that some people, when given a choice between specs and the actual sound, will choose specs every time because they won't be able to sleep at night, incredulous at how Emo made the mistake of selecting part number 456HYDDmm893300a instead of 456HYDDmm893300b because everyone knows the latter has a much higher pain tolerance and decibel overshoot. (2) It kinda boils down to do we trust Lonnie and Keith? I do. |
|
hemster
Global Moderator
Particle Manufacturer
...still listening... still watching
Posts: 51,951
|
Post by hemster on Mar 25, 2013 16:15:35 GMT -5
The difference in how "DACs" sound is mostly due to ancillary circuitry anyway - and very little of it is due to the DAC chips themselves. Therefore, while this would be cool for DIY types, it isn't going to be of much use to very many people. Also, you can forget about re-clocking and such things - since the delay they introduce would mess up the audio/video synch. A "good" op amp should have no sound whatsoever (in the application for which it is chosen). It's sort of like spark plugs - if they work oddly, then there's a problem; if they work right, you shouldn't notice them. From what I've heard, SOME of the discrete versions are quite good, while some are actually rather bad. From an engineering point of view, if the op-amp changes the sound, then there's something wrong - either with the op-amp or your circuit. (Not to mention the fact that the added contact resistance and capacitance of the socket you would need to include so you could swap the op-amp will ALSO make changes for the worse.) Huh... I don't believe that for a second. You start by saying that the DAC isn't that big a deal and that the "ancillary circuitry" is what makes the difference only to go on and basically say all opamps should sound that same. You contradict yourself !!! That's not quite how I read it: The difference in how "DACs" sound is mostly due to ancillary circuitry anyway - and very little of it is due to the DAC chips themselves... Keith said the DAC chip itself is not the be all and end all. It's the implementation that matters. |
|
|
|
Post by solidstate on Mar 25, 2013 16:30:45 GMT -5
Huh... I don't believe that for a second. You start by saying that the DAC isn't that big a deal and that the "ancillary circuitry" is what makes the difference only to go on and basically say all opamps should sound that same. You contradict yourself !!! That's not quite how I read it: The difference in how "DACs" sound is mostly due to ancillary circuitry anyway - and very little of it is due to the DAC chips themselves... Keith said the DAC chip itself is not the be all and end all. It's the implementation that matters. AND Didn't I say the EXACT SAME THING !!! |
|
|
|
Post by solidstate on Mar 25, 2013 16:41:31 GMT -5
Somebody needs to stop being such a sourpuss... and do their homework ;D Actually, the UMC-200 DOES have a digitally controlled analog ladder network volume control... Which means that you DON'T lose resolution when you turn the volume down. (It also means that, if you use the Direct input setting, and have an analog input signal, your signal isn't even digitized.) But.... thanks for the great explanation about why you really need a 32 bit DAC IF you're going to use a digital volume control. Was this seriously his retort ? Didn't it occur to him that some people on here are engineers and would understand that's not exactly true !?! That makes me question his technical knowledge because it hints at a lack of understanding DACs and the way in-which they receive and process audio. Probably just a PR biz guy brainfart because it's hard to explain to consumers the maths et al and he's right in 90% of implementations/sources/transport. The reason you want 32bit is for volume attenuation. Lonny should understand that if you feed a UMC-200 a 24bit signal the digital volume attenuation on the converter will really create a lot of noise and because of the converter choice I wouldn't feed it 24bit in my transport chain. All of this depends on the design. If the DSP engine does 32bit attenuation then feeds 24bit attenuated audio to the converter vis I2S then it's not a problem with that DAC. If the DSP engine does volume attenuation calculations and feeds the results to the DAC so the DAC can attenuate you've got a problem. This is only with 24bit sources though. Usually with room correction the DSP engine does the PEQ and attenuation/channel trims then passes this signal via I2S to the converters. BUT in the designs I've looked at it the global master volume adjusts the DACs volume at the DAC not in the DSP. This is the case with 90% of designs. Perhaps this white paper might clarify for Lonny. www.esstech.com/PDF/digital-vs-analog-volume-control.pdfI don't know much about the UMC-200's design. Perhaps it has a separate analog volume control but I doubt it as it's a multi chan design and why would you have that when you're choosing the lower cost converter in the first place. You guys should bring a Kingwa piece to hear his non feedback diamond output stage with zero coupling caps and four groups class A PSUs with a I/V conversion current differential volume. This circuit blows the doors off of ANY OPAMP PERIOD.  |
|
hemster
Global Moderator
Particle Manufacturer
...still listening... still watching
Posts: 51,951
|
Post by hemster on Mar 25, 2013 16:41:40 GMT -5
^Perhaps you'd like to commission Emotiva to build you a custom modular pre/pro. I for one am glad they're making units that are affordable and sound pretty bloody good. BTW, I didn't stay at a Holiday Inn Express but I am an electronics engineer, albeit without OCD. |
|
hemster
Global Moderator
Particle Manufacturer
...still listening... still watching
Posts: 51,951
|
Post by hemster on Mar 25, 2013 16:51:41 GMT -5
Qinghua (“Kingwa”) units could be found in some of the esoteric gear such as WFS, Audio-GD, PSAudio etc.. but they're hardly affordable gear for most people. True story below: Ok, just for kicks, I was listening to some tunes and suddenly pressed mute and exclaimed "Oh, how I wish I had a Kingwa!" to which my wife said "I can make you some Quinoa if you'd like!" ;D ;D |
|
|
|
Post by solidstate on Mar 25, 2013 17:03:46 GMT -5
Somebody needs to stop being such a sourpuss... and do their homework ;D Actually, the UMC-200 DOES have a digitally controlled analog ladder network volume control... Which means that you DON'T lose resolution when you turn the volume down. (It also means that, if you use the Direct input setting, and have an analog input signal, your signal isn't even digitized.) But.... thanks for the great explanation about why you really need a 32 bit DAC IF you're going to use a digital volume control. e first place. Does UMC-200 use the same Hitachi M61532FP that's in the UMC-1 ? So you guys seem to really like the Ti JFET TL072C opamp as you used everywhere in the UMC-1. |
|
|
|
Post by geebo on Mar 25, 2013 17:04:37 GMT -5
Well just a couple of my two cents observations about this thread... (1) I could care less what DAC or what spec or whatever is in my Emo gear. I care about what it sounds like. It seems to me that some people, when given a choice between specs and the actual sound, will choose specs every time because they won't be able to sleep at night, incredulous at how Emo made the mistake of selecting part number 456HYDDmm893300a instead of 456HYDDmm893300b because everyone knows the latter has a much higher pain tolerance and decibel overshoot. (2) It kinda boils down to do we trust Lonnie and Keith? I do. +1. And a lot of people just won't trust their ears. |
|
klinemj
Emo VIPs 
Official Emofest Scribe
Posts: 15,089
|
Post by klinemj on Mar 25, 2013 17:28:31 GMT -5
I personally trust my ears...don't worry too much about what chips and such are in a unit if it doesn't sound great. And if it does sound great, I don't really care what it's made of.
Mark
|
|
|
|
Post by solidstate on Mar 25, 2013 17:31:25 GMT -5
The difference in how "DACs" sound is mostly due to ancillary circuitry anyway - and very little of it is due to the DAC chips themselves... Keith said the DAC chip itself is not the be all and end all. It's the implementation that matters. And I agree with Keith 100% on the analog section being more important. I just don't agree that opamps are equal to a discrete buffer and I/V. |
|
cgolf
Emo VIPs
Posts: 4,615
|
Post by cgolf on Mar 25, 2013 18:50:17 GMT -5
Qinghua (“Kingwa”) units could be found in some of the esoteric gear such as WFS, Audio-GD, PSAudio etc.. but they're hardly affordable gear for most people. True story below: Ok, just for kicks, I was listening to some tunes and suddenly pressed mute and exclaimed "Oh, how I wish I had a Kingwa!" to which my wife said "I can make you some Quinoa if you'd like!" ;D ;D Quinoa is one of the best foods/grains there is to eat. Love it!!! My wife will cook a pot for breakfast along with an over easy egg on top. Oh man, how good it tastes!!! ;D But I can tell you that I have no idea what all it took to get the egg or the Quinoa. Just that it tastes wonderful. And that's all I really care about. Same holds true for me with the DAC. I care about the sound because telling me what DAC I have doesn't do much for me. If it sounds good to my ears, then I'm good. Everyone is different and different people want more details than others. Nothing wrong with that. Again, what is wrong is when one person tells another person they are wrong, stupid, ignorant, etc. just because they think differently!!!  |
|
I'm new here, but I swear I've read all this in a different thread before as well. Anyway, the analog ladder VC in this unit is a fairly high-end feature, pretty cool!
