Onkyo TX-RZ50 Review (Home Theater AVR)

[SynthesisCinema]

@amirm have you been in touch with Onkyo yet sharing these results?

The power issue didn´t suprise me. I think we saw something similar with the Pioneer you measured and also Gene found similar issue with the Integra model he measured, that is basically identical to one model below this RZ-50 from Onkyo. This model was THX Select 2 Certified and was 4ohm rated. RZ-50 has larger power supply and larger caps. Would this circuit really activate pulling back power when watching movie with speakers that dip down to 4ohm? I remember Gene at some point thought that it wouldn`t.

The Integra DRX-4.3 had very touchy nanny circuits to protect it from overheating. I knew when I saw the big fan and smallish heatsink, this would be the case. Although I was able to measure some impressive 4-ohm power figures, the DRX-4.3 was unable to sustain those figures for more than a few seconds without current limiting kicking in along with the very quiet fan. This would limit output voltage to 13Vrms or 40 watts/ch (4-ohms) and the only way to restore nominal operation was to unplug and power-cycle the receiver, after you let it cool down for a bit.

Integra DRX-4.2 & DRX-4.3 9.2 Channel AV Receiver Review

The Integra DRX-4.3 AV receiver sports the latest in HD audio/video with 9 amplified channels on-board, expandable to a 7.1.4 speaker configuration. Our review discusses its performance and shortcomings.

www.audioholics.com

 

[symphara]

@amirm have you been in touch with Onkyo yet sharing these results?

Onkyo has ceased operations and filed for bankruptcy (check their main website), they apologise for the inconvenience to their customers and that's that.

 

[KMO]

Does this resampling affect bit depth? If I play a 44.1 KHz 24 bits song from Amazon dongle, am I getting 24 bits on these AVR products with bass management on? Is my understanding correct that bit depth is much more important for sound quality than sampling rates above 44.1 Khz ? If so then maybe truncating sample rate is a crime but not a felony?

The resampling shouldn't be affecting bit depth. If DSP is 24-bit, then they can't usually gain anything from using fewer bits. It's not usually "less work" to operate at 16-bit width on an 24-bit DSP. Whereas doing half the sample rate is.

(This might not remain true for some of the fanciest new floating-point DSPs, where conceivably you might be able to do more work in parallel using half-precision than single-precision, but we're looking at more traditional integer DSPs in these AVRs).

 

[KMO]

Onkyo has ceased operations and filed for bankruptcy (check their main website), they apologise for the inconvenience to their customers and that's that.

They took that review a bit badly, then.

 

[dlaloum]

I just went back and re-read the Voxx/Sharp announcement - they basically bought the business about 6 months ago.... so Onkyo Japan going bust may not matter.

The deal appears to have involved some sort of licencing fees back to Onkyo Jp.

Whoever ends up owning the remains of Onkyo, will be receiving those licencing fees - all of which could take a while to work through.

In the meantime, Sharp owns the factory that builds the gear, Voxx owns the marketing, distribution and support channel...

Onkyo JP only owns the right to annual licencing fees - it is not clear whether there are any other rights to its name etc....

 

[dlaloum]

I think Amir's main beef is the power amp 4 ohm nanny policy....

And he may be particularly sensitive to that, as he has been bitten before by a Pioneer LX303 and LX504 - with very similar issues (they are, of course, badge engineered Onkyo's)

The 48kHz discussion has been an ongoing one with various AVR's, and Amir has included testing to identify the conditions under which downsampling happens and/or doesn't.

The rest are discussions about design aspects that could potentially be done better.... (such as HD audio when running 2 channel... etc...)

We have been running with 48kHz DSP limitations since circa 2006... - around 15 years.

Needless to say, today's AVR's have a lot more processing power than even the best flagship units of 2006 - we should be able to do better. (is it needed? is it audible? will the mass market care? - are seperate issues!)

 

[BDWoody]

Not just up the sample rate, run the 'unused' D/A converters in parallel to improve SNR as well. Does any avr/processor out there do that for 2ch 'pure'?

It does make sense to give a little extra for the L and R channels. Krell seemed to think it was a good idea 20 years ago.


"The Showcase employs two different DACs for its various channels. For the front mains, Texas Instruments (formerly Burr Brown) PCM 1737s are used. These are a delta sigma DAC that samples at 192 kHz with a 24 bit word depth. For all other channels, they use T.I.’s 1605 DAC. It is also a 192 kHz/ 24 bit converter, but it is made to handle more channels. Krell has focused its best quality on the front mains, since two-channel is their passion."
https://www.audiogon.com/listings/l...:~:text=The Showcase employs,is their passion.

 

[Martin_320]

Does this resampling affect bit depth? If I play a 44.1 KHz 24 bits song from Amazon dongle, am I getting 24 bits on these AVR products with bass management on? Is my understanding correct that bit depth is much more important for sound quality than sampling rates above 44.1 Khz ? If so then maybe truncating sample rate is a crime but not a felony?

From what I read, it would seem that Yamaha's AV receivers and preamp-processors don't suffer the problem of DSP downsampling.

Here is a discussion on Audioholics, which would suggest that Yamaha's choice of DSPs keeps 192kHz sample rate intact in their products:

Yamaha CX-A5200 and MX-A5200

Also, I would respectfully ask that you be more careful with your edits when quoting others. You edited my statement to read: There is something more happening that is not reflected on the FR...using AutoEQ with an App like D&M has to control how the AutoEQ is applied combined with PEQ is the...

forums.audioholics.com

Extract:
"All Yamaha YPAO (even cheapest AVR) can do 192kHz YPAO. I confirmed that with the Yamaha people.
So Yamaha = 192kHz YPAO, Anthem = 96kHz ARC, and Audyssey/Dirac = 48kHz"

Hi @peng , I see that you were in that above referenced thread too. Did you hear anyone subsequently confirm that the DSPs in Yamaha's pre-pros don't downsample?

 

[peng]

Amir, thank you for the detailed measurements on a newest Onkyo model AVR. Can you explain a little how to read the blue/red curves that is for SINAD vs output voltage with (volume = Max), do you mean +18, that is 99 on the absolute scale. Was the input still set to 0 dBFS, I would think in order to match it with the curve for volume = 90 you must have reduced the input voltage right? So if I read this correctly, the AVR would perform with much lower SINAD at lower input voltage?

I would be great if you include the analog input voltage vs pre out SINAD in every AVR/AVP tests.

Thanks again.

 

[hmt]

From what I read, it would seem that Yamaha's AV receivers and preamp-processors don't suffer the problem of DSP downsampling.

Here is a discussion on Audioholics, which would suggest that Yamaha's choice of DSPs keeps 192kHz sample rate intact in their products:

Yamaha CX-A5200 and MX-A5200

Also, I would respectfully ask that you be more careful with your edits when quoting others. You edited my statement to read: There is something more happening that is not reflected on the FR...using AutoEQ with an App like D&M has to control how the AutoEQ is applied combined with PEQ is the...

forums.audioholics.com

Extract:
"All Yamaha YPAO (even cheapest AVR) can do 192kHz YPAO. I confirmed that with the Yamaha people.
So Yamaha = 192kHz YPAO, Anthem = 96kHz ARC, and Audyssey/Dirac = 48kHz"

Hi @peng , I see that you were in that above referenced thread too. Did you hear anyone subsequently confirm that the DSPs in Yamaha's pre-pros don't downsample?

Well the flipside is that YPAO is basically crap.

 

[PeteL]

Well the flipside is that YPAO is basically crap.

What's wrong with it?

 

[hmt]

Doesn't do anything meaningful to correct bass / room modes.

 

[capt.s]

Applying a Bass Management crossover (I use a 5.1 speaker setup) on its own did not appear to downsample the audio (I would need to re-check);

Please do..... I'd be very interested to know this for sure!

 

[AndreaT]

This is a review and detailed measurements of the Onkyo TX-RZ50 9.2 8K THX Certified AV Receiver. It was purchased new by a member and drop shipped to me for testing. It costs US $1,399.
View attachment 186165
Looks like all AVR companies are trying to break the mold of just dual rotary controls to something different. While I appreciate the additional buttons and knobs here, they are impossible to see in the dark as far as labels. Here is the back panel:

View attachment 186167
I appreciated the quick start-up and the high resolution of the on-screen display. It was also nice of it to tell me at the end of setup that there was a new firmware. What I didn't appreciate was it not recognizing my hardwired Ethernet cable requiring a second try even though I had it plugged it a while back. Update was slow which I could tolerate but not finishing with "Update Complete" (or something like it) and then sitting there non-functional! You had to power cycle it to get it to work.

The distinguishing aspect of this AVR is adoption of Dirac by a mass market company like Onkyo in their AVR. Dirac room EQ should perform far better than simpler scheme in the AVR (although I have not tested either in this instance). There is a nice app to control it all as well so you don't need a computer if you don't want to have one (and can use AVR's own mic).

Note: the follow set of measurements are expanded from my normal set which is already quite extensive for AVRs. If you are new to these reviews, don't be taken back. Most of them are for the diehard measurement folks in the forum. I will summarize everything in the conclusion section.

Onkyo TX-RZ50 DAC Measurements
As usual I start with the important bit which is how well the AVR converts digital audio samples to analog. Pure mode was selected for all the measurements (see later in the review on effect of this setting). Let's start with S/PDIF:

View attachment 186172

As usual I set the volume control such that we get 2 volt output which required setting it to 84.5. Unit displays "Reference" at 82 by the way but at that level, you won't be able to drive all amplifiers to their full power rating. Let's now switch to HDMI:


View attachment 186173

General performance is the same but now we have symmetrical spikes around our main tone which indicates low frequency jitter component. This is unfortunate. Putting SINAD which is the relative sum of noise and distortion in context we get slightly above average rating:

View attachment 186174

Since the unit can output more than 2 volts, let's see the full range of output vs performance:

View attachment 186177
Good to see you have headroom up to 3.5 volt if you need it to drive a low gain amplifier.

Measuring dynamic range we get:
View attachment 186175

Multitone test shows jitter components we saw in the dashboard which I ignored for my distortion-free range computation:
View attachment 186178

Oddly when I ran my specific jitter test, it was S/PDIF which fell behind HDMI:

View attachment 186179

Objectively neither performance is proper but from audibility point of view, spikes are low enough as to not matter.

Linearity was "good enough" for the class:

View attachment 186180

I was pleasantly surprised to see a choice of filters although you only have a choice of two (and "auto" which seems to set to Sharp):

View attachment 186181

Notice the typical roll off in the audible band with slow filter so my choice would be the sharp one. The filter choice naturally makes a difference in our THD+N versus frequency:

View attachment 186183

As you see the slow filter does better but I think that is because it is attenuating the in-band response, reducing the noise and power of harmonics. Either way though, this is not anything to write home about.

While we are here, let's see the effect of Pure mode:

View attachment 186184

Odd to see that low frequency rise in distortion.

Base Management and Pure Mode Investigation
These AVRs are opaque when it comes to what they do when they activate their digital processing and accompanying resampling. So I decided to investigate with frequency response measurement using 192 kHz sampling. This rate is usually outside of unit's internal processing rate so if there is some, we will see its effect. Let's see what happens when we turn Pure mode on and off:

View attachment 186185

It is clear that turning ON the Pure mode disables all internal resampling giving us almost full bandwidth of 96 kHz which we would expect. Turn Pure mode off and response gets truncated sharply. My guess is that internal sample rate is 44.1 or likely 48 kHz.

I then configured an 80 Hz high-pass filter for the fronts and measured its effect:

View attachment 186186

As soon as you enable bass management, resampling is forced on you (blue curve). This is a shame as bass processing should be doable at higher sample rate. I guess the pipeline is all or nothing when it comes to processing. Fortunately you have the option of using Pure mode but that defeats bass management so no bookshelf speaker may apply. You have to use full range speakers.

Notice that this run in Pure mode (red and green) has a spike. This would come and go depending on settings of the frequency response sweep (green spike). It may be a bug or ringing of the resampling filter. Onkyo should investigate and fix this. Fortunately it is at 60 kHz so not an audible concern but maybe it pushes some external amp into oscillation.

Onkyo TX-RZ50 Amplifier Measurements
My goal here is to figure out how good the internal amplifier is relative to an external one. This is hard to do as we have to tease out the transparency level of various ways of feeding the amplifier. Let's start with analog input by adjusting our gain to be 29 dB:

View attachment 186188

If you don't use the Pure mode -- which is probably how most of you use it -- you will lose performance:

View attachment 186189

Sadly that is in the form of rising noise floor. Going with the better of the two, the ranking is not great and well below competition:


View attachment 186190

Some of you have been looking at using external DACs as a way to improve performance of these AV Products. There, you would be feeding the amplifier up to 2 volts in unbalanced mode so I decided to test that condition. To get to the same 5 watt output, I cranked the volume down until I achieved that output:


View attachment 186191

Noise floor goes down enough to buy us 3 dB more SINAD. Let's do the same with HDMI input, i.e. using internal DAC:


View attachment 186192

So you could do better with external DAC but it is likely not worth the hassle as the amp itself doesn't have much more to give.

Crosstalk was below average:
View attachment 186193

In Pure mode, frequency response is extended and nice:
View attachment 186194

And here is our multitone test:

View attachment 186195

Disappointing.

Onkyo TX-RZ50 Power Amplifier Testing
This testing cost me anther hour of sleep (went to bed near 4:00 am!). I started with measuring 4 ohm power as I usually do and got ridiculously low level of output before clipping. I checked all the settings, set the amp to 6 vs 4 ohm and back, power save mode off, etc. but nothing would fix it. On a hunch, I disconnected the power cable, waited a bit and powered the unit back on. I got full power twice but then it limited:


View attachment 186197

Clearly there is some monitoring going on internally causing the amplifier to go into ECO/power limiting mode. With 8 ohm load, I didn't see this problem:
View attachment 186198

Given that vast majority of speakers today are 4 ohm or even lower, this is a serious problem. I could understand it if the amp was cooking but it was quite cool running with the upper fan off. Traditionally this mode would be triggered if you set the amplifier to 4 ohm mode but per above, this did not have an effect here.

Given this problem I could not run my max and peak power. But did run my frequency sweeps post a power cable removal:
View attachment 186199

Notice how it pulled power back half way through the test. I am hoping this is microprocessor controlled so can be fixed with a firmware upgrade. This is a showstopper bug as far as I am concerned.

Conclusions
The inclusion of proper Room EQ in mid-priced AVRs is major news indeed. Alas, my job here is to characterize the performance of the platform itself and the news there is not good. The amplifier has serious power limiting "feature" and even without it, it is not competitive with other products out there. Company can and should do better than this. I am confident a clean up pass with measurements can produce a better product. For too long review sites have done such a cursory job here as to allow these practices to continue. Hopefully that era is behind us. I know a number of companies are taking this domain seriously like the companies did in 1970s. So if Onkyo doesn't get with the program, it will fall behind.

The morass of pipeline processing also needs documentation and this is a message for the entire industry. Don't leave the customer in the dark and lead to dissatisfaction when they read an analysis like mine. Up the processing capability and make it a feature to perform all the functions at the native sample rate of the content. I think people will pay a few hundred dollars more to get this.

Anyway, I can't recommend the Onkyo TX-RZ50. If they fix the power handling issue, they may then get a marginal recommendation from me.

So…we basically have a full range 30 W amplifier (and maybe less than 30 W with all channels engaged) for $ 1,400 + tax…

I think you are making quite a jump/conclusion, that something would sound "poor" based on some certain measurements.

If you were to say, "Not measure great", I might agree, but having actually heard some of the AVRs tested here, and actually having extensive listening also, I can tell you with certainty, no AVR I have heard, has sounded "Poor" in any way.

You are making the assumption that a high Sinad guarantees great sound and a mediocre Sinad means poor sound. Simply not true.
In actual listening, I would wager, you would probably not be able to tell one from the other under almost all circumstances.

I say all of this NOT to defend AVRs or yamaha AVRs in particular, but due to me thinking the same way. "They must sound mediocre and horrible". When after actually hearing one, and then getting two used ones, found it just not even remotely true.

So…we basically have a full range 30 W amplifier (and maybe less than 30 W with all channels engaged) for $ 1,400 + tax…Thank you Amir for steering us away from flawed and expensive products!

 

[peng]

Hi @peng , I see that you were in that above referenced thread too. Did you hear anyone subsequently confirm that the DSPs in Yamaha's pre-pros don't downsample?

As as I know, based on commentaries I read on Audioholics and Sound and Vision reviews, Yamaha's 64-bit YPAO version does not need to down sample from 192 kHz like their older version(s) did.

Dirac Live and Audyssey don't require down sampling to 48 kHz either but the AVP/AVR manufacturers, such as Denon/Marantz chose to (or have to) sample, likely to preserve processing power. As far as I know, all Audyssey equipped AVRs/AVPs require down sampling to 48 kHz if the source's sampling rate is higher. For Dirac Live, going by memory the PC beta version might have been limited to 96 kHz but I think the latest version can do 192 kHz. For the non PC versions, such as those running on Onkyo, NAD etc., I am not sure if some of them would get limited to 48 kHz or 96 kHz, may be @Kal Rubinson or @Flak can help clarity this.

There are many articles on why 48 kHz sampling rate is high enough for humans, that is, no loss of audible information but people just won't let that go. Me neither but that's only because it is better to do less if you don't have to so why down sample if you have the processing power. It is the same idea that I want FR to be 10 to 30,000 Hz +/- 0.1 dB even though I know it wouldn't make any audible difference to me if I only get 20 to 20,000 Hz +/- 0.5 dB.

 

[beagleman]

You are jumping to conclusions. Please refrain from explaining to me what I do or don’t know. I read the reviews and understand full well that sinad is just one Measure of many. They definitely don’t all sound the same and it’s not subtle. I have a NAD T778 and and Anthem mrx1140 in my home theater right now and the Anthem is superior at driving mid bass and woofer response at much higher levels than the NAD at the same 70 to 75db avg listening range.

Then explain your initial comment...............without going off on a tangent about something else as you did in your first reply....

"It's great to see this graph fill out over time and show an ever growing group of "green" AVRs to show we don't have to settle for poor audio quality in favor of codecs, channels and features."

Your initial comment fully does sound like you are saying SINAD is all that matters, and if it is low it will sound poor......Unless you are just backtracking after realizing it was not what you meant to say.

 

[beagleman]

As as I know, based on commentaries I read on Audioholics and Sound and Vision reviews, Yamaha's 64-bit YPAO version does not need to down sample from 192 kHz like their older version(s) did.

Dirac Live and Audyssey don't require down sampling to 48 kHz either but the AVP/AVR manufacturers, such as Denon/Marantz chose to (or have to) sample, likely to preserve processing power. As far as I know, all Audyssey equipped AVRs/AVPs require down sampling to 48 kHz if the source's sampling rate is higher. For Dirac Live, going by memory the PC beta version might have been limited to 96 kHz but I think the latest version can do 192 kHz. For the non PC versions, such as those running on Onkyo, NAD etc., I am not sure if some of them would get limited to 48 kHz or 96 kHz, may be @Kal Rubinson or @Flak can help clarity this.

There are many articles on why 48 kHz sampling rate is high enough for humans, that is, no loss of audible information but people just won't let that go. Me neither but that's only because it is better to do less if you don't have to so why down sample if you have the processing power. It is the same idea that I want FR to be 10 to 30,000 Hz +/- 0.1 dB even though I know it wouldn't make any audible difference to me if I only get 20 to 20,000 Hz +/- 0.5 dB.

I have completely failed to be convinced that any "Hi Res" audio, truly sounds better. Not against the idea, in fact really liked it way back when it first became a thing, but having done a few Blind tests and also read results of a few others, not sure of how much was based in better sound from Hi res itself or simply different masterings and expectations.

 

[hmt]

You are jumping to conclusions. Please refrain from explaining to me what I do or don’t know. I read the reviews and understand full well that sinad is just one Measure of many. They definitely don’t all sound the same and it’s not subtle. I have a NAD T778 and and Anthem mrx1140 in my home theater right now and the Anthem is superior at driving mid bass and woofer response at much higher levels than the NAD at the same 70 to 75db avg listening range.

Can you provide measurements to back that off?

 

[Martin_320]

people just won't let that go. Me neither but that's only because it is better to do less if you don't have to so why down sample if you have the processing power.

This is exactly the point.
And it's baffling to me that a AV receiver's DSPs have no problem rendering 11 channels at 48kHz 24bit, and apply bass-management plus realtime room-correction EQ etc; yet when there are much fewer active channels (e.g 5.1 in my case), and on one of which I want to apply only some simple EQ eg. the center channel, then it still truncates a 96kHz flac down to 48kHz.
There is lots of horsepower in any modern AV receiver, but clearly most of it isn't being used -- unless you have a Yamaha-built unit it seems.

 

[capt.s]

What is more important ist that the Onkyo only uses a 12 db oct slope for the bass management of the mains. That will cause problems when using portet speakers or any speaker that has an in room response digging deeper than the crossover.

That a good point and it would be nice to have a choice of x-over slope or at least the option of 24dB/Oct. I'm not aware of any AVR that offers that. In addition to your point with ported speakers, I find with only one sub, when listening to music, that it's location is given away likely due to higher frequencies that the 12db/oct slope allows to pass.... Just a thought.