I am interested in measuring audio latency of an HDMI sink device (DUT). I have an 8K Murideo Seven Generator, but I believe that its audio latency measurements may be off by 15 or 20 ms, so I would like to verify these.

I can easily generate a low-bandwidth HDMI signal with large back porch and embed a short audio pattern using a desktop computer. I expect that, without the TMDS encoding, I should be able to see the audio pattern within an oscilloscope capture of this large back porch. In fact, I've used an old analog oscilloscope and found that the voltage pattern on TMDS 2 and 3 pins visibly change on the oscilloscope display when audio is played... But I do not currently have a storage scope and I expect the TMDS encoding will make it too difficult to know exactly when the change is happening and ending.

To address this, I would like to know if there is a simple TMDS decoder IC that would take in a twisted pair TMDS and output a non-TMDS digital signal that I can record with an oscilloscope. Or, maybe there is some sort of TMDS decoder device that exists. I am having a hard time finding the correct internet search terms for what I am looking for.

I understand that it is more common to use an HDMI receiver for this purpose, but an HDMI receiver does quite a bit more than I would like and is much more complicated (?) to wire up than I imagine a simple decoder IC for a single set of TMDS pins. In fact, if I used an HDMI receiver, it would not give me access to the data islands that contain the audio packets, but instead would regenerate the audio clock and audio signal, introducing audio latency... which is the exact thing that I would like to measure, so avoiding this audio extraction behaviour is necessary.

The closest thing I have found so far is the "capture" mode of tools like the Teledyne LeCroy quantumdata 980 Protocol Analyzer. I believe this would give me access to time stamped audio packets inside of the data islands. Unfortunately, I don't believe that this tool has an additional analog/trigger input that I can use to record or trigger on the DUT audio output like an oscilloscope would.

  • \$\begingroup\$ What sink device it is? A consumer TV would have significant latency, as they have to delay audio due to video processing. Audio packets also don't have timestamps, and when audio stream is converted to packets in transmitter and from packets to stream in the receiver, if you have a 67.5 kHz Hsync rate, youll quickly realize that the FIFOs used won't hold that many samples to have significant effect even at 32kHz rate. 32 samples is 1 millisecond, and audio packets hold 1 to 4 samples per packet. A DVI chip might give you all packets on the data bus, but it might require manufacturer help. \$\endgroup\$
    – Justme
    Jan 15, 2022 at 21:54
  • \$\begingroup\$ Thanks Justme, that’s a good point about using a receiver chip, such as a DVI chip. The types of sink devices I am interested in are HDMI audio extractors, home theatre receivers, sound bars, gaming monitors, TVs, etc. Some of these have very low latency, as they are meant to be paired with gaming monitors or TVs that have less than 5 ms of video latency. I am specifically interested in determining the offset between analog and HDMI audio output from this HMDI audio extractor and re-embedder: github.com/AVLatency/Latency-Measurement/wiki/HDV-MB01 \$\endgroup\$ Jan 16, 2022 at 2:04
  • \$\begingroup\$ That's a never-ending job. Each TV/monitor will most likely make sure the delay audio to match video processing delay. And AVRs&soundbars can comnunicate the delay between the display to keep lipsync. And it will work for compressed audio as well, not just for PCM. \$\endgroup\$
    – Justme
    Jan 16, 2022 at 9:07
  • \$\begingroup\$ Oh, of course. I would personally say that knowing PCM audio latency is better than nothing. Also, my intent is to do this sort of measurement only once: after I have accurately determined the offset of the $35 HDMI audio extractor for a few different PCM formats, anyone will be able to use my free open source software tool with that audio extractor to determine the audio latency of their consumer electronics. Or reviewers could do the same without needing a $7,000 tool that appears(?) to be inaccurate. \$\endgroup\$ Jan 16, 2022 at 14:06


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.