timg
Minor Hero
Posts: 68
|
Post by timg on Feb 26, 2020 9:38:28 GMT -5
1. Amir didn't use HDMI - he couldn't get it to work. From the post above, it looks like AES, USB, and HDMI all show similar performance. The issues shown in the graphs are either from the digital domain/processing and/or the output DAC. If it's processing related, then hopefully it's something that can be fixed via a firmware update. Tim
|
|
|
Post by dimora on Feb 26, 2020 9:54:53 GMT -5
Hi kids, Everybody calm down... Ray is finishing up measurements and we will post them shortly. We're not seeing the same issues here, but that does not mean the reviewer is wrong. I'm not attacking anyone, but I am defending the integrity of the RMC-1's baseline performance. We've been building high quality gear for many, many years now. Do you really think we are knowingly going to ship a P.O.S. as our flagship AVP? EVERYONE that has heard the RMC-1 knows it a a great sounding unit and our internal measurements support that. Relax a minute and allow us to get to the bottom of things. Clearly something is out of whack somewhere. BTW, I just heard about this in the morning. Geez. Big Dan Big Dan! Welcome to the show. I own an RMC-1. I like it a lot. It sounds great. But...I bought it because I want DIRAC and the expansion module for DIRAC subwoofer control (I have 3X 18's). When are these well-past-due features coming? A lot of us are voting and pulling for your company with our wallets...but you need to be an honest broker with your customers. I've listened to your podcasts, yet nothing shows up in a timely manner. How about an honest update with some real plans and timelines?
|
|
|
Post by mgbpuff on Feb 26, 2020 10:06:22 GMT -5
1. Amir didn't use HDMI - he couldn't get it to work. From the post above, it looks like AES, USB, and HDMI all show similar performance. The issues shown in the graphs are either from the digital domain/processing and/or the output DAC. If it's processing related, then hopefully it's something that can be fixed via a firmware update. Tim I believe that's what I implied also. So what if the problem is due to the effing switching power supply?
|
|
|
Post by dimora on Feb 26, 2020 10:11:11 GMT -5
From the post above, it looks like AES, USB, and HDMI all show similar performance. The issues shown in the graphs are either from the digital domain/processing and/or the output DAC. If it's processing related, then hopefully it's something that can be fixed via a firmware update. Tim I believe that's what I implied also. So what if the problem is due to the effing switching power supply? Then we would need an external power supply in another box that feeds the RMC (pretty common on a lot of high-end gear), or a replacement / fix of the current internal power supply - if that's the case.
|
|
|
Post by Hair Nick on Feb 26, 2020 10:23:05 GMT -5
RMC-1 on audiosciencereview.comRay Dennison - Sr. Design Engineer, Emotiva Audio Corp.
Apples and Oranges Before I comment on the test results of the RMC-1 recently reviewed in your lab, I’d like to mount a defense of not just Emotiva, but of all the processors reviewed and compared unfavorably to 2-channel DACs. They are simply not the same thing. We’ve got to compare apples to apples, so to speak.
The analog path in a typical 2-channel DAC consists of a DAC chip and some sort of analog filter and buffer. With this structure you can get very close to the raw DAC chip specifications with just minor degradation due to the power supply, op-amps, resistors, etc. Performance under your test conditions looks great. However, if you were to retest with more typical listening levels you would find that SNR, THD, etc. may be degraded.
One can design a home theater processor the same way and get the same excellent results under your test conditions. However, a processor needs to account for listening levels, surround processing, downmixing, bass management, EQs, varying amplifier gains and widely varying speaker efficiencies. If our hypothetical processor does not account for these signal processing needs, then it will perform just as well as the 2-channel DACs. However, if I add EQ, require some downmixing or speaker balancing, etc. then our hypothetical processor clips and sounds bad.
Some inexpensive processors try to get around this by adding headroom in the DSP. You figure out the worst-case gain that might need to be applied due to signal processing requirements, attenuate all the signals by that amount and bring the level back in the analog output buffers. Unfortunately, the headroom ends up being in the neighborhood of 30 - 40dB. So now my 24-bit 120dB DAC is now a 17-bit, 80dB DAC. Not ideal.
The high-end way to do this is by using an analog volume control after the DAC. This allows you to adjust the headroom only as needed for the current processing requirements. I’m confident that all of those “bad” processors reviewed use some sort of analog volume control. But we must be aware; analog volume controls are not perfect. At one end you have the noise floor; if your incoming signal is relatively low, then the overall SNR is dominated by the volume controls noise. If you use a higher-level input signal, then the volume control will distort; not good! The designer must choose a sweet spot between SNR and distortion. This analog volume control in the signal path is the main reason you can’t compare 2-channel DACs to HT processors. There are many other, more subtle issues, but this is the main one.
Emotiva designs tend to be biased somewhat towards highest practical SNR, over lowest distortion. Our distortion levels at 0dBFS may be a bit higher than some other brands, but our SNR may be lower. This is a design choice based on real-world conditions. Signals are seldom at 0dBFS and when they are the speaker distortion will be orders of magnitude higher. But this is just my opinion, man. Your mileage may vary. Another designer may feel that optimizing for distortion is the better choice. In either case, the distortion will be better at more reasonable levels. Most reviewers (and spec. writers for manufacturers) use a more reasonable -20dBFS for distortion measurements.
The main point of this discussion is that comparing a home theater processor against a DAC or dongle is simply not valid. Conflating the distortion measurement with SNR measurement simply adds to the confusion. IMO, the SINAD measurement just adds more confusion. SINAD is exactly the same as THD + Noise. People are used to seeing THD + Noise measurements. No one will complain about .005% at 0dBFs, full volume. But converting this to dB makes it look bad to the uninformed. None of the processors you have reviewed are bad compared to 2-channel DACs. They simply have different functions, technical requirements, and circuit topology’s. Respectfully, I suggest that they should be tested differently.
RMC-1 measurementsOn to the specifics of the RMC-1 review. I do not know the exact test conditions used with the Audio Precision or the RMC-1. I am using PCM 48/24 from my lab AP 585 over HDMI. I am using a factory reset RMC-1, and have all speakers set to large and am in “Direct” mode. It is using the recently released v1.8 code. Here is our test result using the view similar to first part of the review.
I do not see anything other than expected small 2nd, 3rd, and 5th harmonic distortion. I can’t comment on the spurious signals the reviewer saw because I don’t see them. Perhaps there is some setup difference on the AP or RMC-1 that might account for this. Note that the THD+N (and SINAD) measurements are essentially identical to the review even without the spurious signal. So how useful is this measurement?
Yes, the THD+N is higher than a 2-channel DAC. Please see discussion above regarding analog volume controls and our design choice to optimize for better SNR.
Here is the same test with -20dbFS signal. This is an extremely clean and quiet signal.
Not too shabby.
I don’t have time right now to re-run all the tests in the review, but here are a few highlights. Jitter test (48kHz). Unclear why our result is so different from the reviewer’s.
Note, no weirdness. Nice and clean.
THD vs frequency. Testing at 0dBFs I get similar results as in the review. The .1% @ 20Hz seems high, but your sub will likely be at 10% THD!! Again, design choices, based on experience and reality. The benefits in overall performance using analog attenuators overwhelmingly outweigh the slight THD penalty at low frequencies and high signal levels.
THD (not THD+N) vs frequency at -20dBFS:
Rising THD at low frequency is characteristic of the analog volume control. Not bad power supply design.
I measure slightly better SNR (>119dB) than the reviewer (116dB) but that is a small difference and excellent either way,
It’s also unclear why the reviewer observed differences between Direct and Reference Stereo modes.
With 2 PCM stereo coming in and large front speakers, there should be no difference at all. The power of ‘Reference Stereo” really comes in play for analog signals. Reference Stereo is a true analog bypass of the analog signal to the volume controls via relay. This raises the performance to that of a high-end analog pre-amp. Note the flat response to the limits of the Audio Precision and very low distortion, even at the 4V levels used by the reviewer. SNR was >123dB under these conditions.
So, as you can see, things are sometimes more complex that they first appear. The RMC-1 and its offspring are well designed AV processors with excellent measured performance and outstanding sound quality.
|
|
|
Post by Mikomill on Feb 26, 2020 10:38:29 GMT -5
Here we go.
|
|
richb
Sensei
Oppo Beta Group - Audioholics Reviewer
Posts: 859
|
Post by richb on Feb 26, 2020 10:49:44 GMT -5
Emotiva,
Thank you for the explanation and measurements.
Can you provide measurement of comments on the behavior of the DAC filter. A linear phase fast filer should not include a rise after attenuation.
- Rich
|
|
|
Post by steelman1991 on Feb 26, 2020 10:51:10 GMT -5
To quote Tej in 2 Fast 2 Furious - "It's gonna be an all timer tonight"
|
|
|
Post by hsamwel on Feb 26, 2020 10:54:52 GMT -5
What he is saying is that if you are using HDMI 3 for example and you want to make a change in HDMI 5 in the setup -> input -> HDMI 3....hit the up arrow to go to HDMI 5, the unit changes the input instead of just letting you go to that HDMI 5 input configuration to make a change while still watching/listening whatever in HDMI 3. It’s just how the menu structure is and it’s probably not trivial to make a user interface change like that. Especially with the effort they had to put in getting Dirac done which meant fixes / other enhancements had to suffer a bit. I’m still baffled at how so many other users experience issues such as a simple firmware update. I haven’t had to reset or unplug HDMI or anything- just insert my USB stick and confirm firmware update. Not sure why there is such variety among the various units. I'm in a similar boat Lsc - any upgrades I've done have been hassle free (but I've also been one of the lucky ones to have hassle free usage as well). Clearly some owners need to use the menu after initial set-up, but I'm curious what for?. Once I've set-up the system, I could count on the fingers of one hand the times I've been back into it and when I have, for testing after 1.8 it's been very reactive. I just don't see the whole "menu" as an issue - appreciate others do though. Just curious what the usage scenarios are that require constant change. I’m also mostly trouble free. That’s why I can complain about UI quircks instead of stability Why? Well it’s probably just me but when I try something I like to set it back to default after testing. Also like to have unused stuff set to off. Probably a tick I have I’m not in the menus that often but when I go in there I like it to fast and smooth.. Similar to when a relative asks me to fix their computer which have Windows XP and is so slow in reacting that you want smash it to pieces and tell them to buy a new one.. The automatic input also changes my amp because I have trigger settings to only start on approriate inputs. In the settings all inputs are available so if I happen to stop at a not used input by mistake the trigger settings shuts down my amp. One time I even changed a setting in an analog input while on hdmi, still had video going. Couldn’t figure out why I didn’t have audio for several minutes until I saw the amp was shutdown.
|
|
richb
Sensei
Oppo Beta Group - Audioholics Reviewer
Posts: 859
|
Post by richb on Feb 26, 2020 11:02:29 GMT -5
|
|
|
Post by steelman1991 on Feb 26, 2020 11:26:23 GMT -5
I'm in a similar boat Lsc - any upgrades I've done have been hassle free (but I've also been one of the lucky ones to have hassle free usage as well). Clearly some owners need to use the menu after initial set-up, but I'm curious what for?. Once I've set-up the system, I could count on the fingers of one hand the times I've been back into it and when I have, for testing after 1.8 it's been very reactive. I just don't see the whole "menu" as an issue - appreciate others do though. Just curious what the usage scenarios are that require constant change. I’m also mostly trouble free. That’s why I can complain about UI quircks instead of stability Why? Well it’s probably just me but when I try something I like to set it back to default after testing. Also like to have unused stuff set to off. Probably a tick I have I’m not in the menus that often but when I go in there I like it to fast and smooth.. Similar to when a relative asks me to fix their computer which have Windows XP and is so slow in reacting that you want smash it to pieces and tell them to buy a new one.. The automatic input also changes my amp because I have trigger settings to only start on approriate inputs. In the settings all inputs are available so if I happen to stop at a not used input by mistake the trigger settings shuts down my amp. One time I even changed a setting in an analog input while on hdmi, still had video going. Couldn’t figure out why I didn’t have audio for several minutes until I saw the amp was shutdown. Thanks mate. As I thought. I'm not much of a tinkerer - more a set and stick guy, hence why I don't get sticky menu syndrome (not sure if you have to "self-quarantine" for that ) . The odd occasion I do have to, I find it really snappy, the occasional "please wait" isn't an issue for me - still each to their own.
|
|
timg
Minor Hero
Posts: 68
|
Post by timg on Feb 26, 2020 11:27:39 GMT -5
Hair Nick. Thanks for posting. It's good to see factory measurements and some explanation for why they are what they are. It's also reassuring to see that some are very similar to ASRs.
Questions to be answered: Why does the ASR review have a significant difference between Reference Stereo and Direct? Why does the ASR review show high levels of jitter and Emotive sees none? What are the Emotiva Intermodulation results? Are these similar? Is the rising distortion at higher levels part of a design decision to optimize around -20 dbFS? What's causing instability in these readings over time and potentially also unit to unit? How can we get to the point where the performance after months of operation is identical to immediately after a factory reset without changing anything else? Where do most people actually run their volume controls during a movie? Are the results at 0, -20 dbFS, or something else most representative of the typical operating conditions?
Thanks.
|
|
|
Post by megash0n on Feb 26, 2020 11:30:09 GMT -5
Hair Nick. Thanks for posting. It's good to see factory measurements and some explanation for why they are what they are. Questions to be answered: Why does the ASR review have a significant difference between Reference Stereo and Direct? What are the Emotiva Intermodulation results? Are these similar? Is the rising distortion at higher levels part of a design decision to optimize around -20 dbFS? What's causing instability in these readings over time and potentially also unit to unit? How can we get to the point where the performance after months of operation is identical to immediately after a factory reset without changing anything else? Where do most people actually run their volume controls during a movie? Are the results at 0, -20 dbFS, or something else most representative of the typical operating conditions? Thanks. I'm typically around -18, but will shift 2 or 3 depending on the encoding.
|
|
|
Post by aswiss on Feb 26, 2020 11:34:33 GMT -5
I finished Installation of FW 1.8 (coming from 1.5) and I'm mostly happy with it. One minor thing I'm still not happy is: Stopping a DSD Direct Stream (from SACD, played on oppo203) is still causing a loud knack noise (Like a click) in the speakers. Could also be an oppo behaviour. Due to the "please Wait" function, first 1-2 secs are not hearable from whatever source I use. Maybe this will go away once this function is taken out of the FW. But it's bit faster than before.
Overall it works fine for me (at least for the major things I've tested so far in my setup) as I wished it to have it working from the beginning. Since 10 months of waiting have ended to get it like this I'm not jumping around beeing happy but satisfied for the moment.
|
|
cawgijoe
Emo VIPs
"We made too many of the wrong mistakes." - Yogi Berra
Posts: 4,919
|
Post by cawgijoe on Feb 26, 2020 11:35:48 GMT -5
Hair Nick. Thanks for posting. It's good to see factory measurements and some explanation for why they are what they are. Questions to be answered: Why does the ASR review have a significant difference between Reference Stereo and Direct? What are the Emotiva Intermodulation results? Are these similar? Is the rising distortion at higher levels part of a design decision to optimize around -20 dbFS? What's causing instability in these readings over time and potentially also unit to unit? How can we get to the point where the performance after months of operation is identical to immediately after a factory reset without changing anything else? Where do most people actually run their volume controls during a movie? Are the results at 0, -20 dbFS, or something else most representative of the typical operating conditions? Thanks. I'm typically around -18, but will shift 2 or 3 depending on the encoding. My system is in a family room. Depending on the film and how it's encoded, I will listen between -25 to -30. That's plenty loud enough for me.
|
|
richb
Sensei
Oppo Beta Group - Audioholics Reviewer
Posts: 859
|
Post by richb on Feb 26, 2020 11:44:31 GMT -5
Where do most people actually run their volume controls during a movie? Are the results at 0, -20 dbFS, or something else most representative of the typical operating conditions? The RMC-1 seems to attenuate some signals more than others. For example, listing to DD via TiVo, the typical level is -30. Using the ATV4K, the same levels require at 6 to 12 dB additional gain. Typical is -20 dB. Typical amplifiers require between 1.6 and 2.0 volts to achieve full output so measurements at 2 volts are a good choice for most applications. 4 volts should also be supplied as this is a defacto standard. The XMC-1 seems to have improved the THD+N at 4 volts. - Rich Attachments:
|
|
|
Post by megash0n on Feb 26, 2020 11:46:35 GMT -5
Where do most people actually run their volume controls during a movie? Are the results at 0, -20 dbFS, or something else most representative of the typical operating conditions? The RMC-1 seems to attenuate some signals more than others. For example, listing to DD via TiVo, the typical level is -30. Using the ATV4K, the same levels require at 6 to 12 dB additional gain. Typical is -20 dB. Typical amplifiers require between 1.6 and 2.0 volts to achieve full output so measurements at 2 volts are a good choice for most applications. 4 volts should also be supplied as this is a defacto standard. - Rich I agree that it seems to be all like over the map depending on many factors.
|
|
Lonnie
Emo Staff
admin
Pay no attention to the man behind the curtain
Posts: 6,999
|
Post by Lonnie on Feb 26, 2020 11:51:07 GMT -5
Hair Nick. Thanks for posting. It's good to see factory measurements and some explanation for why they are what they are. It's also reassuring to see that some are very similar to ASRs. Questions to be answered: Why does the ASR review have a significant difference between Reference Stereo and Direct? Why does the ASR review show high levels of jitter and Emotive sees none? What are the Emotiva Intermodulation results? Are these similar? Is the rising distortion at higher levels part of a design decision to optimize around -20 dbFS? What's causing instability in these readings over time and potentially also unit to unit? How can we get to the point where the performance after months of operation is identical to immediately after a factory reset without changing anything else? Where do most people actually run their volume controls during a movie? Are the results at 0, -20 dbFS, or something else most representative of the typical operating conditions? Thanks. Just as a point of reference. Our production AP test are run on every processor when they are first built and again after burn in. The same production AP test is run on any processor that comes back to the company for any reason. Unit to unit and over time, the results are always constant. Why ASR measures differently is not a function of time or variances in production but in test procedures. As I have stated before, we use standardized test that virtually every other audio company in the world uses, so to get to the bottom of why they are different would be asking Amir how he test. One last thing, the setting on the volume control is not the level that is being referenced in the test. That level is a function of input signal level, not output. Best regards, Lonnie
|
|
|
Post by thxultra on Feb 26, 2020 11:55:34 GMT -5
RMC-1 on audiosciencereview.comRay Dennison - Sr. Design Engineer, Emotiva Audio Corp. Apples and Oranges Before I comment on the test results of the RMC-1 recently reviewed in your lab, I’d like to mount a defense of not just Emotiva, but of all the processors reviewed and compared unfavorably to 2-channel DACs. They are simply not the same thing. We’ve got to compare apples to apples, so to speak.
The analog path in a typical 2-channel DAC consists of a DAC chip and some sort of analog filter and buffer. With this structure you can get very close to the raw DAC chip specifications with just minor degradation due to the power supply, op-amps, resistors, etc. Performance under your test conditions looks great. However, if you were to retest with more typical listening levels you would find that SNR, THD, etc. may be degraded. One can design a home theater processor the same way and get the same excellent results under your test conditions. However, a processor needs to account for listening levels, surround processing, downmixing, bass management, EQs, varying amplifier gains and widely varying speaker efficiencies. If our hypothetical processor does not account for these signal processing needs, then it will perform just as well as the 2-channel DACs. However, if I add EQ, require some downmixing or speaker balancing, etc. then our hypothetical processor clips and sounds bad. Some inexpensive processors try to get around this by adding headroom in the DSP. You figure out the worst-case gain that might need to be applied due to signal processing requirements, attenuate all the signals by that amount and bring the level back in the analog output buffers. Unfortunately, the headroom ends up being in the neighborhood of 30 - 40dB. So now my 24-bit 120dB DAC is now a 17-bit, 80dB DAC. Not ideal. The high-end way to do this is by using an analog volume control after the DAC. This allows you to adjust the headroom only as needed for the current processing requirements. I’m confident that all of those “bad” processors reviewed use some sort of analog volume control. But we must be aware; analog volume controls are not perfect. At one end you have the noise floor; if your incoming signal is relatively low, then the overall SNR is dominated by the volume controls noise. If you use a higher-level input signal, then the volume control will distort; not good! The designer must choose a sweet spot between SNR and distortion. This analog volume control in the signal path is the main reason you can’t compare 2-channel DACs to HT processors. There are many other, more subtle issues, but this is the main one. Emotiva designs tend to be biased somewhat towards highest practical SNR, over lowest distortion. Our distortion levels at 0dBFS may be a bit higher than some other brands, but our SNR may be lower. This is a design choice based on real-world conditions. Signals are seldom at 0dBFS and when they are the speaker distortion will be orders of magnitude higher. But this is just my opinion, man. Your mileage may vary. Another designer may feel that optimizing for distortion is the better choice. In either case, the distortion will be better at more reasonable levels. Most reviewers (and spec. writers for manufacturers) use a more reasonable -20dBFS for distortion measurements.
The main point of this discussion is that comparing a home theater processor against a DAC or dongle is simply not valid. Conflating the distortion measurement with SNR measurement simply adds to the confusion. IMO, the SINAD measurement just adds more confusion. SINAD is exactly the same as THD + Noise. People are used to seeing THD + Noise measurements. No one will complain about .005% at 0dBFs, full volume. But converting this to dB makes it look bad to the uninformed. None of the processors you have reviewed are bad compared to 2-channel DACs. They simply have different functions, technical requirements, and circuit topology’s. Respectfully, I suggest that they should be tested differently. RMC-1 measurementsOn to the specifics of the RMC-1 review. I do not know the exact test conditions used with the Audio Precision or the RMC-1. I am using PCM 48/24 from my lab AP 585 over HDMI. I am using a factory reset RMC-1, and have all speakers set to large and am in “Direct” mode. It is using the recently released v1.8 code. Here is our test result using the view similar to first part of the review. I do not see anything other than expected small 2nd, 3rd, and 5th harmonic distortion. I can’t comment on the spurious signals the reviewer saw because I don’t see them. Perhaps there is some setup difference on the AP or RMC-1 that might account for this. Note that the THD+N (and SINAD) measurements are essentially identical to the review even without the spurious signal. So how useful is this measurement?
Yes, the THD+N is higher than a 2-channel DAC. Please see discussion above regarding analog volume controls and our design choice to optimize for better SNR. Here is the same test with -20dbFS signal. This is an extremely clean and quiet signal. Not too shabby. I don’t have time right now to re-run all the tests in the review, but here are a few highlights. Jitter test (48kHz). Unclear why our result is so different from the reviewer’s.
Note, no weirdness. Nice and clean. THD vs frequency. Testing at 0dBFs I get similar results as in the review. The .1% @ 20Hz seems high, but your sub will likely be at 10% THD!! Again, design choices, based on experience and reality. The benefits in overall performance using analog attenuators overwhelmingly outweigh the slight THD penalty at low frequencies and high signal levels. THD (not THD+N) vs frequency at -20dBFS: Rising THD at low frequency is characteristic of the analog volume control. Not bad power supply design.
I measure slightly better SNR (>119dB) than the reviewer (116dB) but that is a small difference and excellent either way, It’s also unclear why the reviewer observed differences between Direct and Reference Stereo modes. With 2 PCM stereo coming in and large front speakers, there should be no difference at all. The power of ‘Reference Stereo” really comes in play for analog signals. Reference Stereo is a true analog bypass of the analog signal to the volume controls via relay. This raises the performance to that of a high-end analog pre-amp. Note the flat response to the limits of the Audio Precision and very low distortion, even at the 4V levels used by the reviewer. SNR was >123dB under these conditions. So, as you can see, things are sometimes more complex that they first appear. The RMC-1 and its offspring are well designed AV processors with excellent measured performance and outstanding sound quality. This is what I have been saying about comparing a DAC to a AV processor glad you have the same conclusion. I did just did the update on my XMC-2 to 1.8 last night (posted this in the XMC-2 thread also). Had some issues with the usb input not working with my Ropiee box (when playing 192khz24bit tracks it would cut out and skip). I was able to turn UAC2 off and back on and reset my xmc-2 a few times and got it to work. I will have to see tonight if it still works. My question is after doing the firmware update do you recommend resetting to factory settings and redoing the settings? I noticed you did this for your review and want to know if it is recommended.
|
|
richb
Sensei
Oppo Beta Group - Audioholics Reviewer
Posts: 859
|
Post by richb on Feb 26, 2020 12:03:28 GMT -5
One last thing, the setting on the volume control is not the level that is being referenced in the test. That level is a function of input signal level, not output. There has been a great deal of confusion generated by the HTP-1 amplifier voltage setting (I think that is what is called). It is of interest to those with amplifier-gain settings to match for best performance. A THD+N sweep would provide this information. Also, can you comment on the digital filter behavior? This is certainly something that should be software addressable. - Rich
|
|