What I mean is ultra-low distortion, lots of ENOB, accurate reproduction in every way (including for reverb tails 🙂, etc. For a dac there are actually a multiple things.
For other people it may mean something different. For the some 'high-end' folks it can mean the only distortion is some H2 and H3, and they are in a particular ratio. Also, all the little low level details are good to not have go missing. Don't know their position on noise, but they probably prefer it low.
There may be other views as well. Don't think there is any official definition of 'sounds good.'
More accurately, "WHY the DAC-3 gives an impression of better sound than OPPO 105?" because all we know so far is that RNMarsh came up with the verdict from a subjective casual listening comparison.
Are you sure they didn't say it's quite good for your age group? At least that's what audiologists typically tell the patients if their hearing checks out acceptable.
Say one listens to music at 90dB SPL average level, -100dB would be -10dB SPL. I would like to see a simple test, any gear would do, on the ability to hear even at full BW music at -10dB SPL in an average listening environment. Pretty east test to do.
Oppo was built to a price point as you know. They didn't want to implement even the few small design changes you suggested. The design appears to closely modeled after ESS evaluation boards.
DAC-3 was designed in an entirely different way. A lot of measurements and very open-minded analysis was done. They figured out what was good about ESS chips and what wasn't. The dac chips are designed so that some functions can be performed externally, and Benchmark took advantage of that to improve the sound quality of some functions. Mostly it was done digitally, and also through analog in the form of RF for clocking. Power supplies are also very well designed, and are good example of well-implemented SMPS. The analog output stages are also well-engineered, but mostly pretty unremarkable opamp design with a few thoughtfully implemented details here and there (near as I can tell). It seems that opamps can sound pretty good if done right and with very good low-impedance-across-the-audio-band power supplies (low-z up to maybe 100kHz is needed, IMHO, including fast transient response -- ladder network equivalent caps are not so good in that application, again IMHO). The reason they used LME49860 is so they could run +-18v rails to get professional output levels for studio and mastering use. That particular family of opamps LME49720/LME49860/LM4562(? IIRC) can sound very good except they are very sensitive to RF from DECT phone base stations, and similar. Thus, DAC-3 is in a nice well-shielded steel case with ferrite cylinders around every I/O wire going to back panel connectors. The PCB is mulitlayer (don't know how many), but it does a good job of keeping all the RF and digital from getting into the analog stuff. Good design there too.
Last edited:
The OPPO is Chinese and has the build quality of a midrange AV receiver.
At 90dB SPL across a room one will not perceive all that much in terms of little details. There will be enough room reflections and SPL-related masking to hide a lot of small differences. Allo Katana with good power supplies will sound about the same as DAC-3 in that type of situation, but Katana is still pretty darn good. Under other listening conditions such as in the speaker near field at lower SPL, or if using headphones, most people could probably hear the difference.
Perhaps I missed something, but I was hired by OPPO to evaluate their circuitry. I did not see any serious compromise except one which involved resistor values. My other suggestions to them were ignored, but if you met their chief engineer, you would find that he was much like many of my critics here. Why, should they change a part? What MEASUREMENT would it effect? What if we lost 1/2db of S/N for better listening quality, is that worth it? etc.
I dont Think there is an IC opamp which can drive a low Z headphone very well. As long as the load is not too low Z, they are good. Maybe the OPPO 105D had too low loads on their IC opamps. The ML-9600 uses 5532/4 which do a better job into lower Z's. I would have to find and open that highly "recommended" HPAmp to see what they used.
Higher supply voltage make a difference as well. I would say most of the trouble is from the IC's output stage in particular.
I think it is more than the opamps though.
The design I did that was better, I was able to tweek the values and select the transistors to reach a level of distortion that was below my AP2722 generator. Drives low Z loads with extremely low distortion and only uses 8 transistors.
Sounds a lot more accurate than other IC Opamp based HPAmps. Another better than IC Opamp based HPA is the one designed by Jam for Pass Labs... also discrete.
THx-RNMarsh
Higher supply voltage make a difference as well. I would say most of the trouble is from the IC's output stage in particular.
I think it is more than the opamps though.
The design I did that was better, I was able to tweek the values and select the transistors to reach a level of distortion that was below my AP2722 generator. Drives low Z loads with extremely low distortion and only uses 8 transistors.
Sounds a lot more accurate than other IC Opamp based HPAmps. Another better than IC Opamp based HPA is the one designed by Jam for Pass Labs... also discrete.
THx-RNMarsh
Last edited:
IC op-amps don't drive 8 Ohm loads what kind of obvious inappropriate criticism is that? You can't fry a steak with a cigarette lighter either. Actually Walt Jung did an all IC driver with the AD813/AD815 that was essentially distortion-less.
Last edited:
OPA1622 does fine, TPA6120. Same with anything using a buffer in the loop. This isn’t 1980. Doug Self made a power amp with NE5534s... not that I’d recommend that but you can easily drive headphones with ICs.
Last edited:
I wish the OPPO 105 sounded better! But it is probably better than most here could not find anything wrong about it.
There is one interesting difference between the 105 and the DAC-3 apparently, and that is 12V vs 18V in the I-V section power supply voltages. It would have helped the 105 to have higher voltage rails, because the output resistive loading could have been reduced.
- Status
- Not open for further replies.